ML20198H338

Evaluation of Possible Water-Hammer Loads in Prairie Island Svc Water Sys for DBA Conditions

ML20198H338
Person / Time
Site:	Prairie Island
Issue date:	10/31/1996
From:	FAUSKE & ASSOCIATES, INC.
To:
Shared Package
ML20198H322	List: ML20198H317 ML20198H326 ML20198H338 ML20198H377 ML20210U319 ML20210U326 ML20266A493 ML20266A497 ML20266A501 ML20266A505 ML20266A507 ML20266A511 ML20266A515 ML20266A523 ML20266A527 ML20266A531 ML20266A538 ML20266A546 ML20266A549 ML20266A555 ML20266A558 ML20266A561 ML20266A566 ML20266A570 ML20266A575 ML20266A578 ML20266A585 ML20266A588 ML20266A590 ML20266A594 ML20266A597 ML20266A600 ML20266A604 ML20266A607 ML20266A610
References
FAI-96-89, NUDOCS 9709180013
Download: ML20198H338 (89)

Text

{{#Wiki_filter:_ _ _ _ _ _ _ _ _ _ _ _ FAI/96-89 Evaluation of Possible Water-Hammer Loads in the Prairie Island Service Water Systemfor DBA Conditions Submitted To: Northern States Power Prepared By: Fauske & Associates, Inc. 16WO70 West 83rd Street Burr Ridge, Illinois _60521 October,1996 9709180013 9iO915 DR ADOCK 050002 2: l ._________U

7 1 L TABLE OF CONTEN115 l 1.0 INTROD UCTION..................................... 1 - 1 2.0 DIFFERENCES BETWEEN OPEN AND CLOSED SYSTEMS.......... 2-1 3.0 PRAIRIE ISLAND EXPERIMENTAL INVESTIGATIONS............ 3-1 3.1 Scaling Considerations.............................. 3-1 3.2 Experimental Apparatus............................. 3-13 3.3 Experimental Results ..............................3-21 4.0 A DISCUSSION OF THE PRAIRIE ISLAND EXPERIMENTAL RESULTS.............................. 4-1 5.0 OTHER PERTINENT EXPERIMENTAL INFORMATION............ 51 5.1 Configurations Investigated............................ 5-1 5.2 S ummary of Results................................ 5-6 5.2.1 Without Drain-Down of the Supply Riser and Without An Extended Horizontal Loop Seal.................. 5-7 5.2.2 Experience With an Extended Length of Discharge Piping... 5-10 5.3 Considerations of the Two-Phase Flow Pattern During Voiding..... 5-24 6.0 ANALYTICAL CONSIDERATIONS.......................... 6-1 7.0 CONDITIONS WHICH AFFECT THE LOADINGS................ 7-1 7.1 The Influence of Dissolved Air in the Water................. 7-1 7.2 Evidence of Noncondensible Gases Exiting From Solution......... 7-2 7.3 Importance of Small Steam or Gas Voids on the Mixture Sonic Velocity.................................. 7-10 7.4 Noncondensible Gases Exiting Solution................... 7-15 L\\ pal \\9H9. TOC

11 8.0 WATERHAMMER LOADS FOR COLUMN REJOINING DURING REFILL WITH RESIDUAL OAS............................ 8 1 8.1 Description of the Approach........................... 8 1 9.0 CONSIDERATIONS OF DISSOLVED OAS IN THE VOIDED REG I ON........................................... 9 1 10.0 CONC LU S IONS...................................... 10 1

11.0 REFERENCES

1 1 1 APPENDIX A: An Interpretation of the Westinghouse Steam Generator Waterhammer Experiments.................. A 1 APPENDIX B: An Interpretation of the MIT Steam Column Collapse Ex peri men ts................................. B 1 APPENDIX C: Comparison of the Residun! Vold Analy:.0 With the Water Cannon Experiments........................ C-1 APPENDIX D: Comparison of the Residual Void Model With the Creare Steam Generator Model Data................... D 1 APPENDIX E: Comparison of Bubble Collapse Model with Available Test Data............................. E 1 L\\FAPMat. TOC l l \\\\ l

I ill - 1 i i l LIST OF MGURES Figure 31 Experiment to assess water removal by hydrodynamic forces....... 3 3 Figure 3 2 Filmwise condensation on a vertical surface film growth, i temperature distribudon, and velocity profile (taken from j Kreith, 1960).................................... 3 8 l Figure 3 3 Representation of tM bubble rise for a steam void during l re fill........................................ 3 1 1 [ Figure 3-4 Two-phase flow patterns during refill of the discharge piping...... 312 Figure 3 5 Water experiment for Prairie Island fan cooler service water configuration................................... 3 14 l Figure 3-6 Water experiment for fan cooler service water configurations with drain-down of the supply riser with 2" pip.ng in the i [ fan cooler discharge............................... 3 15 Figure 4-1 Measured waterhammer pressures for Test #49 in the 2 inch i mock up for the Prairie Island service water system............,, 4 2 Figure 4-2 Waterhammer Test #47 in a 2 inch scaled representation of the Prairie Island service water system with drain-down included for the supply piping ...............................48 4 Figure 4-3 ' Approximate boundary layer th!vaess for the water flow during a refill transient with a Froude number approximately ] equal to unity................................... 4 13 Figure 51 1" pipe diameter waterhammer experiment.................. 5 2 j Figure 5 2 2" pipe diameter waterhammer experiment.................. 5 3 i Figure 5-3 Experimental configuration with an extended horizontal loop i seal for investigating possible waterhammer conditions in the service water system................................ 5 5 Figure 5-4 Pressure history for the voiding and refill phases in a - configuration without drain down of the supply piping and Without an extended horizontal loop seal...........,........ 5 8 i IVAF#49. TOC - 5 ~. - -. - -, - - - - - -, - - - -.. - - -.. _ _,.. _. - ,-,..--,--._,---.--_.---n

• lY =

l Figure 5 5 Measured pressure transients in the 2 inch piping configuration l with extended length of discharge piping and no drain down in the supply riser.................................. 5 11 Figure 5-6 Dermocouple response in the vertical downcomer during the voiding phase................................... 5 16 Figure 5-7 Dermocouple response in the horizontal loop seal during the vold phase ..................................517 Figure 5 8 Comparison of waterhammer incidents and the thermocouple response during volding of the horizontal loop seal............ 5 20 Figure 5 9 Expanded plot of a waterhammer event showing the depressurization event preceding the rapid pressure increase....... 5 21 Figure 510 Simultaneous pressure traces from three slug impact experiments.... 5 22 Figure 61 Influence of water attempting to form a separated configuration y in a horizontal piping segment during the voiding phase.......... 6-2 Figure 6-2 Flooding and entrainment supplying saturated water to the surface where local condensation is occurring....... 6-5 Figure 7-1 The formation of bubbles of vapor over cavities in a heated surface (taken from Rohnenow, 1973)..................... 7-4 Figure 7-2 Experimental apparatus for two phase critical flow due to flashing (taken from Henry, Fauske and McComas,1970a)........ 7 6 Figure 7 3 Pressure and void fraction profiles for Run 3 of TS R7, taken from Henry, Fauske and McComas (1970s).................. 7 8 Figure 7-4 Experimental data of Klingeblel (1964) and Cruver (1%3) indicating dissolved gases are coming out of solution.......... ,, 7 9 Figure 7 5 Comparison of the homogeneous adiabatic model and Eq, 7 5 with the steam water data of Karplus.................. 711 Figure 7 6 Comparison of the proposed correlation and the steam-water da ta......................................... 7 12 Figure 7-7 Comparison of Eq. 82 and the steam water data of DeJong and Firey..................................... 7-13 t\\PAf409. TOC 5

.v-i Figure 7 8 Test apparatus for observing noncondensible gases exiting solution at subatmospheric pressures..................... 7 16 l Figure 7 9 Measured void fractions for room temperature saturated with air at one atmosphere, at subatmospheric pressures......... 717 Figure 81 Schematic of the test apparatus showing the void measuring section in place. For the pressure-time readings that section was not there. Different lengths were accomnxxisted by changing the lengths of the straight sections of pipe............. 8 2 I Figure 8 2 Comparison of the peak pressure reported by Sweeney and Griffith for the 24 and 40 ft (7.3 and 12 m) sections of pipe and calculated values for the residual void model. The bounding lines result from the measured vold fraction boundaries..... 8-4 Figure 8 3 Measured void fractions reported by Sweeney and Griffith and the boundary values used in these calculations................ 8 5 Figure 91 Calculated pressure increase in the voided region as the two water columns approach.............................. 9-4 Figure 9-2 Calculated water velocity of the water column from the fan cooler (U1) and that of the stagnant water column (U2) ..........95 Figure 9 ' The displacement of the two water columns following the on set of re fill.................................... 9-6 t 1.\\ FAIN 89. TOC

. vi. l l l LIST OF TAM 25 Table 3-1 Experimental data for Water Removal Due to Hydrodynamic Forces Caused by Rapid Depressurization................... 3-4 Table 3 2 Drainage Calculation for a Vertical Pipe.................... 3 9 Tsble 3 3 Approximate Representation of the Transient Water Film Development....................................39 Table 3-4 Prairie Island Waterhammer Test Data (1" Dia. Test Apparatus)..... 3 17 Table 3 5 Prairie Island Waterhammer Test Data (2" Dia. Test Apparatus).... 3 20 Table 3-6 Waterhammer Data (1 in. Diameter Piping)................. 3 22 Table 3-7 Waterhammer Data (2 in. Diameter Piping).................. 3 24 Table 51 Summary of Observations During the Voiding Phase........... 518 Table 71 Solubilities of Nitrogen in Water........................ 7 2 Table 7 2 Example Calculation of Noncondensible Gas (Assumed to be N ) Dissolved in Water.............................. 7 3 2 Table 9-1 Assigned Parameter Values for a Generic Evaluation of Column Rejoining in an 8' (0.2 m) Pipe......................... 9 7 L\\ pal \\4M.10C

l1

1.0 INTRODUCTION

l l 'Ihe purpose of this experimental program is to investigate the potential severity of postulated water hammer events on the Containment Fan Coll Units (CFCUs) and associated piping during mitigation of design basis events (DBE). The specific DBE under consideration is a large break loss of coolant accident (LOCA) coincident with a loss of off site power (LOOP). The LOCA results in a high steam partial pressure and heat input on the containment side of the CFCU cooling coils. The LOOP results in a loss of power to the cooling water pumps and the CFCU fans. Cooling water flow is restored when the diesel driven pumps start and/or the motor driven pump is loaded on the emergency diesel generator (EDG). The fans restart when they are loaded on the EDGs. Due to coast down times and subsequent sequencing on the EDGs, the possibility exists that the CFCU coils could be volded prior to the reinitiation of flow. When flow is reintroduced, under these volded conditions, water hammer pulse (s) could be generated. The objective of these tests is to model and, thns, predict the severity of these events for the Prairie Island Nuclear Generating Plant. This report presents the modeling considerations (scaling), a description of the test apparatus, the experimental results and conclusions. IAFAl%49.1

21 2.0 DIFFERENCES BETWEEN OPEN AND CLOSED SYSTEMS When considering the issues of column separation and rejoining in tall systems, as well as the innuence of two phase now through a circuit, it is necessary to consider the role of noncondensible gases which are dissolved in the water and which may exit from solution. For such evaluations, the differences between an open and closed system, with respect to dissolved gases, can he substantial. The Davis Besse service water system is an open connguration. As a result, the water drawn into the service water pumps is saturated with air at atmospheric conditions and remains as such for all situations considered in these DBA conditions. Certainly there are numerous experiments reported in the literature (Rohsenow,1973) where extensive (long term) boiling of heated surfaces has resulted in the degassing of the nucleation sites that result in an increased wall superheat typical of a reduced gas concentration in the water. Similarly, waterhammer experiments for feedwater systems (Roldt,1975), which are closed systems, took precautions to degas the water used in the tests. However, for an open system which typically operates under single phase conditions, there would be no long term degassing and one would expect the performance under these DBA situations to be substantially innuenced by gu exiting from solution consquently affecting both column separation and incipient nucleation on heated surfaces. Moreover, as the respective coolant systems return towards an all water state after the pumps are restaned, noncondensible gas in the steam bubbles and cavities would act to substantially cushion the resulting response due to both compression of the gas space and a reduction of the sonic velocity in the mixture due to the entrained small gas bubbles. A similar situation may not be achieved for a closed cooling circuit. In particular, the long term operation of a closed system could result in degassing parts of the system if there is substantial heat input in one location, or a component which tends to degas the system as a result of very low pressure and actions taken to remove gas from solu'.lon such as a gas ejector, Given this substantial difference between a closed and open system, the response of those systems which are always open should be assessed as if the coolant is saturated with gas at I atm and the gas exiting from solution under very low pressures or substantial heat addition (or both)

1. war #89.2

32 l should be part of the assessment for the dynamic response of the coolant, including the influence of the dynamic response on the piping supports and hangers, l l L\\PAAM-49,2

31 3.0 PRAIRIE ISLAND EXPERIMENTAL INVESTIGATIONS 3.1 Sc=Ila Considerations For an experiment to provide meaningful results, the apparatus (unless it is full scale) needs to be scaled with respect to the controlling physical processes. A format for developing the appropriate scaling was presented by the Nuclear Regulatory Commission (NRC,1991) and the first major step is to develop a Phenomena Importance Rt.nking Table (PIRT). For waterhammer evaluations in this part of the service water system, the set of physical processes of interest are: 1) column separation, 2) net steam generation in the fan coolers, 3) the configuration of the service water piping including loop seals, 4) steam condensation due to the discharge pipe heat sink and 5) the cold water refill rate when the service water flow is established. Cela== Senaration Since the fan coolers are located at different elevations, the potential for column separation will be different for each elevation. Nevertheless, an experiment should have the capability to represent the potential for significant column separation. Net Steam Generation As discussed above, the high steam partial pressure in the containment resulting from the LOCA causes substantial energy transfer to the fan cooler coils and the service water inside. However, if the steam is generated too rapidly, the steam outflow would " flood" the s.ater in the tubes and force the service water out of the coils. Therefore, consideration should be given to the maximum steam generation rate that would enable water to remain within the fan cooler L\\PAf%49.3

32 l tubes and support further vaporintion if the vaporintion rate becomes sufficient to displace the water from the fan cooler tubes, the vaporization would cease and the final location of the water in the discharge piping would essentially be determined by column separation. l Conversely, if the vaporintion process is sustained, the net steam generation could wntinue to displace the water inventory in the discharge piping well beyond that location representing the equilibrium limit for column separation. To provide a simple characteriution of the possibility for fluid removal from the fan cooler tubes as a result of a depressuriution like that induced by column separation, an experiment was performed for a single fan cooler tube. Figure 31 illustrates the apparatus in which a 5 inch diameter copper tube was evacuated by opening the ball valve to a large evacuated receiver vessel. Next, this ball valve was closed and the tube was filled with water from an atmospheric water supply and the temperature of the fan cooler tube was recorded. Once the tube was completely filled with water, the water supply ball valve was closed such that the water was captured at a pressure of 1 atm, with the temperature being greater than room temperature but substantially less than the saturation temperature (212'F/100'C). Next, the ball valve connecting the test apparatus with the evacuated receiver vessel was cycled full open and full closed as fast as it could be manually moved; by stop watch, this cycle was timed to be at least two seconds, no longer than three. After the test tube was exposed to this low pressure environment, the remaining liquid in the tube was drained and measured. Table 31 illustrates the results for the three experiments. As shown, they all have essentially the same results, i.e. two thirds of the water was ejected as a result of the exposure to a depressurized environment. Certainly, the addition of energy from the containment would increase the potential for sustained removal of liquid by hydrodynamic forces, thereby minimizing the water inventory available to provide sustained vaporintion in the fan cooler during the interval over which the service water flow would be stagnated. IAFAM449J

1 g 4 5 Evacuated Chamber y g / V -0.15 m Ball Valve Thermocouple Ball Valve m l CopperTubing ag_ 6.4 m 9' (21 ft.) gi ; %%Y?: Fill / '" SteenSECOR S-748 Reservoir Figure 3-1 Experiment to assess water removal by hydrodynamic forces.

34 Table 31 Experimental Data for Water Removal Due to Hydrodynamic Forces Caused by Rapid Depressurization Initial Initial Water Initial Water Remaining Water Temp Pressure Inventory Water Inventory Fraction l Duration 'F/'C psig/MPa (ml) (m() Remaining 3 secs 122/50 14.5/0.1 1300 - 450 0.35 i l 3 secs 140/60 14.5/0.1 1300 ~ 450 0.35 3 secs 131/55 14.5/0,1 1300 - 450 0.35 Typically, the velocity necessary to " flood" liquid in a horizontal configuration is taken to be the same velocity for " flooding" vertically downward flow. This velocity can be expressed as (Kutateladze,1972 and Wallis,1%9) { l U' = 3 8"(E* ~ E'} (31) {p; 1 where g is the acceleration of gravity, a is the steam-water surface tension, p, is the water density and p, is the steam density. For a saturated steam water mixture at I atm, this is a ve1xity of 59 ft/sec (18 m/sec). Translating this into an effective heat load that would be-removed in the fan coolers corresponds to a removal rate of approximately 1 MW, or about 1/10th of the design heat load for the fan cooler units. This is important when considering the extent of steam generation; the heat removal that the fan coolers could produce in terms of net steam generation within each unit is only about 1/10th of the design load without disrupting 'he capabilities for sustaining the steam generation. 'Ihis steam production rate of 1 MW net steam generation at 1 atm corresponds to a steam velocity (without considering condensation losses) of 79 ft/sec (24 m/sec) in the 8 inch discharge piping. Since the Froude number (discussed-later) characterizes the potential to develop a stratified two phase flow pattern, it is important that the experiment has a similar steam supply velocity. A: an example, for a pressure of I atm this corresponds to a flow rate of 0.016 lbm/sec (7 x 10 3 kg/sec) in a 1 in, pipe. In this experimental program both 1 in and 2 in, scaled tests were fabricated and tested. LWAf%M.3

35 The steam flow was determined in the experiment by the single phase critical flow through an orifice plate, which was sized to deliver approximately 79 ft/sec (24 m/sec) in the 1-inch pipe test apparatus; this represents the upper limit of steam generation within the fan coolers that would enable liquid to remain in the fan cooler coils. For an upstream pressure of 15 psig (2 bars absolute) an orifice phte of 1/4" would be sufficient to deliver this steam flow. Obviously, higher pressures in the steam generator would cause higher flows into the test apparatus. Orifice sizes as large as 5/8" were also used to investigate the influence of small vaporization rates even though such rates would " flood" the water in the fan cooler coils. Bplae Continuration Many of the issues related to the formation of significant waterhammer loads are related to the specific piping layout configuration. In particular, loop seals are places where stratified steam-water configurations could possibly develop, thereby leading to the formation of water slugs that could be accelerated through the piping. For the Prairie Island fan coolers, this includes the fan cooler coils which could provide a loop seal between the supply and discharge lines. Another aspect specific to a given configuration is the boundary condition on the supply piping side, i.e. whether t..e supply piping would experience column separation. These geometrical features were included and will be discussed in the section on the experimental configuration. Sta== Condeaution on the Discharme Platan Condensation on the discharge piping is a function of the piping wall heat sink, the wall thermal conductivity, and the nature in which energy is added to the wall. For similar levels of design prescure for the piping, we would expect the ratio of the pipe radius to the wall thickness to remain constant. However, since the steam flow rate depends on the cross sectional area, one would expect larger pipes to have somewhat less of a heat sink per unit of flow. Furthermore, in the postulated plant situation, the pipe wall could experience energy addition from steam flowing through the pipe as well as from the high steam partial pressure in the containment atmosphere. In this regard, one would expect the service water piping to have more IAPAl\\El9.3

3-6 energy input from the postulated accident than would be the case in the experiment. Thus, voiding could be faster in the case of the service water piping than in the experiment. To address this, the experimental test matrix includes various steam addition rates to examine the result of different steam voiding rates for the discharge piping. Moreover, by fabricating experimental models using both 1 in, and 2 in, pipes, potential influences of experimental scale related to condensation were also examined. it is also noted that the thermal conductivity governs the rate at which energy is conducted into the pipe wall. in the experiments we have used carbon steel pipe which is similar to that used in the service water system. - Hence, the development of the thermal boundary layer in the pipe wall is similar for the experiments and the service water system. Cold Water Refill Rate The final scaling issue relates to the water refill rate since this dictates the manner whereby the steam void would be condensed (or pushed out of the piping) and fluid transients (including water hammer) experienced during the return to steady state service water flow. An important aspect of this refill rate is the Froude number given by (Wallis et al.,1977) U* Fr = (32) (s0 where U, is the refill velocity and D is the pipe inner diameter, if this Froude number is 0.5 the horizontal pipe will run filled with water (Wallis et al.,1977). For conservatism, some waterhammer characterizations (Bjorge and Griffith,1984) have used a Froude number of unity as a convenient reference of when significant condensation induced waterhammer events would not occur. *lhis means the pipe would fill in a plug flow manner as opposed to a stratified configuration in the horizontal piping segments and significant condensation induced waterhammer would not occur during the refilling transient (Bjorge and Griffith,1984; Izenson et al.,1988). If this is the case, the dynamic loads on the piping system and the piping supports would be those associated with the water refilling velocity, i.e. the pressure associated with this is given by the waterhammer equation UFAF#993

37 AP = p, a, U, (33) where a,is the speed of sound and water. (Note that thl is the expression for a wall stagnating a moving water column. If the moving column meets a stagnant water eclumn, the pressure increase is one half this value.) As will be discussed, the Froude number for the refilling of the I fan cooler service water discharge piping is approximately 1. Derefore, this is an important l scaling element to be represented in the experimental investigation. l Anothu aspect of the refilling rate is the behavior of the vertical piping with water being l-added at the top of the piping configuration. In this regard, drainage within the piping segment can be related to the refilling rate to determine if the vertical piping can run full (plug flow) in the downward direction, or whether it is determined by the film drainage and/or rivulets as well as water falling through the central region of the pipe. For this we need to consider film drainage and the bubble rise velocity for slug flow. Figure 3 2 illustrates the laminar film draining process represented in Nusselt's analysis for a condensing film (Kreith,1960). For this particular assessment, we are only concerned with the drainage rate and not the temperature distribution through the film. Table 3-2 outlines a calculation that examines the steady-state film thickness needed _to drain the imposed water flow rate for the 8 inch discharge piping at an assumed refilling rate of %5 gpm. As indicated by this calculation, the 8 inch pipe would not run full of water. However, the calculation shown in Table 3-2 represents only the steady state drainage rate; before this can occur, the water must accelerate, by gravity, to this condition. Table 3 3 approximates this acceleration and shows that 'after 3 ft ( 9 m) the average film drainage rate would be approximately 10 ft/sec (3 m/sec) and the film thickness would be somewhat larger than one inch, nl approximate representation for the water drainage illustrates two important aspects which control the two-phase flow pattern: First, once a liquid film is formed and accelerates, the drainage rate for modest film thicknesses is more than sufficient to drain the incoming flow rate. Second and r 'nversely, the accumulated water at the top of the vertical segment has no significant downward velocity, hence, water would accumulate and tend to be pushed downward 13PAIM89.3

E. ._L.au, p-#e.*. _A e ,. sam __.,a subam a a ._wes-Wm,4 m 4 4.a h A4-dem e se be._h4Ws_4mae-_A__---JmMs. A4M 38 E 1 ~ m me N a R s$ .+ T 'g i 1 i r g i TiWhNV5Whi;,;;f.:.it.G.'4.:awgh 4 g .o } o r N2{ i 1 r i i / c -Q u 1 s a %%%MEGM%WEwah .!!e e a E i 1 cd C j Q 4 q i I 8 i m 6 'B' 5 k 4y-l 1 .g I h N 'O g:2 3 4 g 1 t i EE i 55 I e l % % % \\\\ % V M 6 $ W M Q W Mi M h N t A 4 a H g M =u C I l.\\PAf489.3

3-9 Table 3 2 Drainage Calculation for a Vertical Pipe Assume Nusselt's analysis for laminar film flow Flow Rate per Unit of Wall Length I' (kg/sec/m) = p* g 85 3 6 = f34 (p' g s 61 ka/sec = 97 kg/sec/m I P = w (0.2) m i 3 100 kg/m = p 5 x 10 Nosec/m2 d = f.8 m/sec2 g = 2.5 mm (0.1 in) 6 = This is much less than a thickness that would fill a pipe, i.e. the 8 inch discharge would not nm full. Flow Rate Per Unis Length for the 21/2" Discharge Pipe P = 12.1 kg/sec/m 6 = 1.2 mm (0.05 in) The 21/2" outlet piping from each coil unit will not run full. Table 3 3 Approximate Representation of the Transient Water Film Development u, = gt m _u = 2/3 u = 2/3 gt t~ U = 2/3 gx x 5 6 (m)[ft) (m/sec)[ft/sec] (mm)[in] 1[3,3] 3[9.8) 3 3 [1.31 4 [13.1) 6 [19.7] 16 [0.63] 8 [26.2] 8.4 [27.5] 12 [0.47] IVAh9649.3

3010 at the refill velocity, his would form a steam bubble (Figure 3 3) which would tend to rise against the incoming water flow rate. Wallis (1969) characterizes the rise velocity of such inertially dominated steam bubbles as 8 ~E 8 U, 0.345 (3-4) P. - where the constant of 0.345 has been deduced experimentally but is quite close to that derived analytically by Davies and Taylor (1950) and others. Since the water density is far greater than that of steam, this essentially reduces to U,, = 0.345 /g5 (35) For the 8 inch discharge piping, this rise velocity is about 1.6 ft/sec (0.5 m/sec), which is much less than the refill velocity Consequently, while the film drainage rate is sufficient to drain away the water after the film has accelerated, the steam bubble rise velocity is too small compared to the water refill rate to enable a steam void to remain in place. Thus, from hydrodynamic considerations alone, the steam void would be " pushed" ahead of the water refill region. As will be discussed later in this report, those thermocouples inserted into the' central region of the vertical and horizontal pipes indicate that the regions are " quenched" sequentially at a rate that agrees with the refill velocity. With these considerations for the two-phase flow behavior in the horizontal and vertical components of the piping configuration, the general two-phase flow behavior in the discharge piping during refill is illustrated in Figure 3-4. Since there is some potential for either film drainage or free fall of water droplets through the pipe central region, there is a strong likelihood that a two-phase front would be developed at the leading edge, particularly in the horizontal'part of the piping, nis two-phase region would certainly tend to " cushion" any impacts associated with the refill at sharp turns in the piping configuration or upon impacting a partially open valve. The influence of such behavior is discussed further with respect to the. - experiments in subsequent sections. L\\PAIMM.3

3 11 Water Refill vhhu ) f u $g Pipe Wall fN Bubble Rise Velocity, Um / / / '/ / s / / / / / / u / Steam Vold / (Steam Bubble) l / / / / / / / l / / 1 / / t / / / / / / / / '/ / s / / / / / / / / / / /J / / / / / / / / / / ,/ / / / / / / J D RH96S033 CDR 9 5-96 Figure 3-3 Representation of the bubble rise for a steam void during refill, l.TAMl9.3

3 13 a Water Initially Discharged from the Fan Cooler Coils i'" 3 Water Drainage Down the Discharge Piping:locity Because the Refill Ve r [ Exceeds the Bubble Rise Velocity, the Vertical Segment Essentially Runs Full of Water. f Water Drainage Reaches the Horizontal Piping: Horizontal Pipe Does Run Full. Likely has a Two-phase Mixture at the Leading Edge. = Water Fills the Horizontal Pipe and Begins . to Fill Vertical Piping. Likely has a Two-phase Mixtura at the Leading Edge. AMOS5031 CDR 44 04 Figure 3-4 Two-phase flow patterns during refill of the discharge piping, 1.TAf%89.3

3 13 3.2 Erner4-hl Angarttus To address the phenomena identified abo" experimental apparatus was constructed to create a situatior in which: 1 a) column separation could occur once the imposed flow rate was removed j (decayed), b) steam get.cration would be added to the piping configuration, c) a significant loop seal exists in the simulated fan cooler, l d) a significant size loop seal exists at the bottom of the piping e) the supply piping can experience column separation as well as some drainage and l l f) water refill rates would be imposed on the system for Froude numbers typical of those of interest in the Prairie Island service water system. Figure 3 5 illustrates the general configuration used for these experiments and also indicates the relative elevation differences in the piping configuration. This experimental configuration was assembled with 1-inch piping and 2 inch piping (Figures 3 5 and 3-6, respectively) in the region ofinterest from the mixing zone immediately downstream of the two solenoid driven ball valves, through the 19.4 ft (5.9 m) vertical down leg and the remainder of the piping into the large evacuated receiver vessel. As illustrated in Figure 3-5, the instrumentation consisted of pressure transducers located at the major changes in flow direction and thermocouples (TCs) along the pipe with TCs 1 through 11 installed such that they penetrate to the pipe centerline. With these type K thermocouples in the flow stream, the extent of void ingression can be monitored as can the refill once it is re-established. Also, the comparison of the measured temperature in a volded region provides a check on the pressure measurements. The response of the pressure transducers can be checked from measured behavior at the end of the experiment when the flow is suddenly stopped by closing the manual ball valve at the discharge into the receiver vessel. twe m u

,A -m AB a m-. u*. 4.d a 44.___ m .._-"mm_ .h. __-a_ ima a 4-4. J..-.4-aJw., m .&..----,_a-4.2u--.ha-.i em.m. Ad ,.r-4wJ.

• • A m.

m -a .a. n--.--...- 3 14 'l o3 3 N b# } 1 je la gj. b y 9 i 8 e e-E N i I g l a 3 k-Iq-3 = l E- ] to ID U 9 ~, c Q l og I 53 Ej 2 &) - 3 w y o-d .I s, m g ~ ll h-N N g-a a J i. ~~ m ( o gi e m e 1.\\ pal \\9689.3

3 15 e E N '1 ) -l ( ]. 1 1 j ( e ) } 5 h 'IA e .y = 1 a 'st t' o .,r ql4 .g = 8 v M\\ a e a s !\\ !E" -l m t I e \\ bm a e e' 8 D g i "I g :1 9 y!5.L............)A }]'L l l_ ,q g j; \\...j...'.............',' ej'i ,g 3 T _1 's s g l u. h_ i L, t a e3 } e2 l I o ma u .i i-p i es .. III I E n. 5.g e' s '9 m 8 LW ! 9649.3 inum' i

3o16 To set up the test conditions, isothermal experiments were performed to obtain the settings for the gate valves used to control the flow through the test apparatus. Typically these were set up to achieve the desired Froude number in the experiment, in particular, considering that the experiment was performed in 1 inch piping (ID = 1.049 in/0.0266 m) and the service water system has 8 inch piping (8.071 in/0.205 m), the refill velocity decreases as the square root of the pipe diameter. Hence, the water velocity in the experiment should be about 36% of the velocity in the service water system. For 900 gpm flow rate, which approximates the DBA conditions, the experimental flow rate should be about 2.1 ft/sec (0.65 m/sec). A test was penormed by establishing cold water flow through the entire flow circuit for five to ten seconds, simultaneously closing the ball valve for the water sepply and opening both the ball valve for the steam flow into the piping configuration and the ball valve controlling draining in the water supply line. The steam flow was maintained for a pre-determined interval then the steam supply and water drainage ball valves were closed and the water ball valve opened again to introduce cold water to refill the coolant circuit. One of the uncertainties to be addressed for the service water system is whether the evaluation is sensitive to a) the time at which off site AC power is lost to the service water system and b) the interval between when the AC power is lost and when the flow is re-established. To address this in the experimental investigations, the interval for steam addition was varied from approximately twenty seconds to one minute to examine the influence of the magnitude of the initial steam bubble. Table 3 4 shows the variations in the parameters considered in the 1" pipe test program and Table 3 5 lists the test conditions for the 2" pipe experiments. Note also that the steam addition rate was varied by using different orifices to control the steam line flow. Since most of the attention was to be directed at formation of a steam void and subsequently refilling of the horizontal loop seal, the largest steam orifice was used for most of the experiments. One of the important parameters in setting up the condition for a given experiment is the pressure in the test section during the water flow conditions prior to initiating flow decay. To provide a situation similar to that in the Prairie Island fan cooler discharge, the control valve at the discharge and at the test apparatus was adjusted such that the pressure in the test apparatus lAPAfttH9.3

1 g taw M ? Prairie Island Waterbasenser Test Data (l' Die. Test Appentus) u 1 Steam Initial Water Initial Steam Receiver Vessel Initial Water Steam Flow Orifice Size Water Temp Reservoir Tempduce Pressure Flow Interval Interval Test (in) (*F) Pressure (psig) (*F) (in Hg) (secs) (sec) 1 7/16 68 72 287 -263 5 30 2 7/16 68 76 268 -25.9 5 20 3 7/16 68 75 311 -28.1 5 40 4 7/16 68 ,68 286 -27.9 5 50 5 1/4 68 73 278 -27.8 ~5 60 6 1/4 68 72 268 -27.6 5 20 y 7 1/4 68 67 268 -27.6 5 30 8 1/4 68 70 266 -27.6 5 40 9 1/4 68 70 287 -27.4 5 50 10 1/4 68 70 287 -27.2 5 70 11 5/8 68 72 284 -29.0 15 30 12 5/8 68 72 272 -28.4 5 20 13 5/8 68 70 262 -28.3 5 40 14 5/8 68 72 264 -28 3 5 50 15 5/8 68 70 276 -28.0 5 10 16 5/8 68 72 294 -28.0 5 60 !E

Table 3-4 r! Pmirie Isised Waterham-cr Test Data (1* Dia. Test Apparatus) Steam Initial Water Initial Steam Receiver Vessel Initial Water Steam Flow Orifice Size Water Temp Reservoir Temperature Fwe Flow Interval Interval Test (in) (*F) Pressure (psig) (*F) (in Hg) (secs) (sec) 17 5/8 68 72 287 -27.8 5 25 18 5/8 68 72 288 -27.6 10 20 19 5/8 68 72 280 -27.4 10 30 20 5/8 68 72 262 -27.0 10 40 21 5/8 63 72 253 -26.0 10 50 22 5/8 68 72 262 -26.3 10 60 23* 5/8 68 72 288 -26.6 10 40 E 24* 5/8 68 72 295 -26.5 10 60 25* 5/8 68 72 285 -26.4 10 90 26* 5/8 68 72 275 -26.4 10 120 27** 5/8 68 72 N/A -28.0 5 10 28** 5/8 68 72 N/A -28.0 5 5 29'* 5/8 68 72 N/A -28.0 5 15 30 5/8 68 70 271 - 7.1 10 30 31 5/8 68 70 262 -27.0 10 20 32 5/8 68 70 269 -27.0 10 40

E C 8 Tam M Prairie Isised Wateda==m-Test Data (1" Dia. Test Apparatus) Steam Initial Water Initial Steam Receiver Vessel Initial Water Steam Flow Orifice Size Water Temp Reservoir Tewg4we Pressure Flow Interval Interval Test (in) (*F) Pressure (psig) (*F) (in Hg) (secs) (sec) t 33 5/8 68 72 280 -27.0 10 30 t 34 5/8 68 70 277 -27.0 10 40 t 35 5/8 68 72 267 -27.0 10 60 36* None 68 71 277 -28.1 10 20 37' None 68 68 273 -28.0 10 30 Y 38' None 68 68 268 -28.0 10 55 39' None 68 68 273 -28.0 10 60 40* None 68 70 296 -27.2 10 20 41' None 68 68 288 -27.2 5 30 42' None 68 70 278 -26.9 5 45 43' None 68 69 281 -26.7 5 50 44' None 68 70 296 -27.7 5 15

• No reverse flow through the inlet piping.
  • No steam flow, i.e. experiment observed column separation and rejoining only.
• Bypass valve closed at refill.

1 m

3 Table 3-5 5 Prairie Island Waterbasneer Test Data 3 (2* Dia. Test Apparatus) Steam Initial Water Initial Steam Rex:civer Vessel I.itial Water Steam Flow Orifice Size Water Temp Reservoir Temperature Pressure Flow Interval Intend Test (in) (*F) Pressure (psig) (*F) (in Hg) (secs) (sec) 45 None 68 68 303 -27.7 10 15 46 None 68 68 293 -27.5 10 20 47 None 68 65 285 -27.1 10 30 48' None 68 69 287 -26.8 10 20 49' None 68 76 305 -26.8 10 40 50' None 68 71 296 -26.6 10 50 "g, o 51 None 68 74 284 -26.3 10 60 52 None 68 72 274 -26.2 10 60 53 None 68 60 294 -28.2 10 20 54 None 68 60 283 -28.2 10 40 55* None 68 60 276 -28.0 10 20 56* None 68 60 297 -27.8 10 40 57' None 68 60 279 -27.1 5 20 58' None 68 61 272 -26.0 5 20

• No reverse flow through the inlet piping.
  • No steam flow, i.e. experiment observed column separation and rejoining only.
• Bypass valve closed at refill.

3 21 was approximately 30 psig (a total of approximately 3 bers). This provides a reasonable representation of the pre"ure within the test assembly under nominal flow conditions, therefore provides a good characterization of the steam pressure that could be created by vaporization within the fan cooler coils. In several cases the run was initiated by first opening the ball valve at the entrance to the evacuated receiver vessel then the solenoid operated ball valve, initiating water flow into the test apparatus. Through this set of conditions, the processes of column separation and column rejoining could be observed in the experiment. Once water flow was initiated, it was sustained for a few seconds to establish the imtial condif. ions. The water supply solenoid driven ball valve was then closed and the ball valve for the steam flow opersed. This enabled steam to enter into the upper horizontal leg of the test apparatus and, depending upon the orifice size and the duration of the steam flow, the void propagated into the supply and discharge legs and loop seal components of the test apparatus. After the steam had flowed into the test apparatus for a pre-determined interval, the steam valve was closed and the water valve re-opened to permit water flow to the test apparatus as dictated by the driving pressure and the specific experimental configuration. 3.3 Experimental Results While the experiment was capable of recording the fluid transients associated with column separation, rejoining, steam addition, water refill and closing of the manual ball valve immediately upstream of the evacuated received vessel, the elements of primary interest were the pressures recorded as the water refilled the steam space and re-established the nominal flow rate from the water supply vessel to the evacuated received vessel. This was the configuration of interest for waterhammer in the test configuration including the loop seal formed at the bottom of the test apparatus. Thus, the element of principle interest here was the dynamic pressure associated with this part of the complete pressure history recorded by the various transducers. Tables 3-7 and 3-8 show the peak pressures measured for the various experiments, for the 1 in. and 2 in. diameter piping configurations respectively, as well as the estimated refill velocity as indicated by the time interval necessary to fill the void within the test apparatus. L\\PAh9649.3

3-22 Table 3-6 Waterhammer Data (1 in. Diameter Piping) Estimated Water Peak Pressure Peak Pressure Refill Velocity Experienced During at Other Times Test (ft/sec) Refill (psig) (psig) 1 7.5 20 200 at flow stop 2 7.5 20 150 at flow stop 3 7.5 50 35 when the system runs full 4 7.5 20 170 at flow stop 5 7.5 15 220 at flow stop 6 7.5 20 230 at flow stop 7 7.5 15 170 at flow stop l 8 7.5 15 230 at flow stop I 9 7.5 30 45 at flow stop 10 7.5 30 195 at flow stop l l 11 7.5 15 155 at flow stop 12 7.5 15 85 at flow stop 13 7.5 15 210 at flow stop j 14 7.5 20 50 at Sow stop 15 7.5 30 125 at flow stop 16 7.5 20 200 at flow stop 17 7.5 25 200 at flow stop 18 0.85 2 0 19 0.85 2 62 at flow stop 20 0.85 2 76 at steam addition 21 0.85 2 64 at flow stop 22 0.85 8 63 at flow stop 23 0.85 5 62 at flow stop (20 curing voiding) 24 0.85 15 73 at steam addition (10 during voiding) 1:\\PAB96 49.3

3-23 -- Table 34 - Waterhammer Data (1 in. Diameter Piping) Estimated Water Peak Pressure Peak Pressure Refill Velocity Experienced During at Other Times Tent (ft/sec) Refill (psig) (psig) 25 0.85 35 64 at flow stop 26 0.85 25 63 at flow stop 27' l.7 2 60 at flow stop 28 1.7-2 60 at flow stop (40 psi during initial column separation) 29 1.7 2 60 at flow stop (40 psi at initial l_ column separation) 30 1.7 10 30 at flow stop 31 1.7 5 60 at flow stop 32 1.7 10 28 at flow stop 33 1.7 5 67_ at flow stop -34 1.7 10 66 at flow stop 35 1.7 15 57 at flow stop 36 1.7 5 68 at flow stop 37 - 1.7 20 65 at flow stop -38 1.7 25 70 at flow stop 39 1.7 10 74 at flow stop (30 during voiding) 40 6.2 100 200 at flow stop 41-6.2 100 at refill 290 a: flow stop 42 6.2 140 at refill 260 at flow stop 43-6.2 30 at refill 240 at flow stop 44 6.2 30 120 at flow stop LWAM4493

3-24 Table 3 7 Waterhammer Data (2 in. Diameter Piping) Estimated Water Peak Pressure Peak Pressure Refill Velocity Experienced During at Other Times Test (ft/sec) Refill (psig) (psig) 45 2.3 25 50 at flow stop i 46 2.3 8 50 at flow stop 47 2.3 2 50 at flow stop. 48 2.3 24 54 at flow stop 49 1.8 30 100 at flow stop. 50 2.1 40 80 at flow stop 51 2.1 40 80 at flow stop 52 2.1 10 72 at flow stop 53 1.1 5 62 at flow stop (60 at steam initiation) 54 1.1 12 63 at flow stop 55 1.1 15 66 at flow stop 56 1.1 15 64 at flow stop 57 6.0 38 130 at flow stop (50 psi st initiation of refill) 58 6.0 20 100 at flow stop (30 psi at steam - initiation) t\\PAA9649.3 ..__ _j

4-1 4.0' A DISCUSSION OF THE PRAIRIE ISLAND EXPERIMENTAL RESULTS ) As indicated by the test matrix, the experimental results explored a range of possible transient conditions. These included different steam voiding rates, different voiding intervals, dra!n-down of the supply piping, no drain-down of the supply piping, and 1-inch and 2-inch scaled mock ups of the service water configuration. Of particular interest in these experiments were the peak pressures generated as a result of waterhammer transients, during both the voiding and refill intervals. This spectrum of conditions is used to examine the possibility that certain conditions could lead to significant two-phase instabilities that would cause substantial waterhammer loads, i l As indicated in the table describing the test conditions, the peak pressures measured during the waterhammer transients are quite mild. Figure 4-1 illustrates the measured behavior for test #49, which is for the 2-inch appantus and also is an experiment in which reversed flow (drainage of the supply piping) did not occur. As illustrated in this figure, waterhammer transients were recorded when the water was introduced into the test apparatus fifty-five seconds into the test. As illustrated, several small waterhammer events were recorded, each with a magnitude of approximately 30 psig, and these events were not observed by the downstream pressure transducers. This is to be expected since the compliance of the downstream two-phase mixture is very large. Also, measurable events were recorded by transducer P5 during the voiding transient, particularly when the vertical risers and the throttle valve were undergoing the voiding behavior. Here again, these transients have a magnitude of about 20 psig and are not observed upstream of pressure transducer 4 to any significant degree. The compliance of the upstream mixture during the voiding transient is very large and it is not surprising that these pressure signals are not transmitted. It is helpful for each experiment to have an inherent point of calibration to assure that waterhammer transients can be appropriately monitored if they occur. One means to accomplish this is to observe the pressurization when the test is completed, i.e. the entire flow is stopped. This is shown in Figure 4-1 as " Flow Stopped" and results in a pressurization of 70 to 100 psi L\\FAh9649.4 .A

l i

l l! i il i ?N me tsys r 0 e ta i

::l: m i 7 : _

i. 1 1 [ w 4 ec i v 4 0 r g 0 es 1 d n a ls 0 ^ I l 9 e i W i i r i a i P r P i t 0 he g S 8 t r 9 E o f 4 4 p f i u f j 0 k T i d 7 c ~ o e S i t 4 ) m r T t e C E 6 a h E c g S 0 n % li S l i 6 l R e i C 2 fe E i R l h i ~ e i t M i ~ n i 0 M g I 5 9 i E A 4 M 4 H I t e T t i s Ri i e Ep l d i 0 T I I e r 4 r i t T 9 i t r i s o a f A 4 i S ^ n i e Wt i o i 0 s ru s j i e i c s t 2 l 3 e I Pt i j e p e i r n e I i er i m f 0 m j d a l 2 e e m i t tS a r e a hr t e i S t i w a ) o I 0 w e F i d i l 1 er u i s a b_ y_:I_::_: F _ :. e i : } O M o oo O ob oW om OT on om o~ o o_' e a l =ObQ:

• ~a*

-4 erug iF .,. ?wfI i

4-3 E S mhm 0 ,i....,,,.;,..,,....,...i.,.;...i....i.,,,,,,,,,... g - o o Q ~ l l - o ~ c) .o .c g b S 5 m 7 y 4 ~ ' lo 'f I ~~~ ~ 6 M 8 m o E u -. o u u .c g ,5, m w z = w E p .s r c - o,'.a . a. c, pt g r m m Q A 4-z.s -. a I y" W9 T H (y) l w v S c 2s 7,- g w ui o a u e

a. -

m m [ N v . o e E m ~ .c y 9 g 1 = ,,,, i,,,,1,,, l... E... l.... l... I.. I,,,,1,,,, l.,,, l.... o 0 .4 01100106 08 0/_ 09 05 Ob OE 02 01 0 01- .o 91 S d,... E d.. ~ n O= .9.0 u. L\\ PAL 9449 4

,IlllI l\\i

l l

aL me tsys r 0 e t i 2;:^~ 7. . 7:7: j :R - : i M 5, a 1 w 1 ec i 0 v r g 0 es 1 dna i l i 0 s I g 9 e i 2 i r i i a r i P i e g 0 h 9 i 8 t ro f 4 p ff i 0 u k T i 7 c ~ o S 4 ) m E C h T i 0 n E c g I S i 6 R i ( 2 E i ~ e M i ht M i ~ n A j l 0 i E H a i 5 9 M 4# i I Ri i T t t s ~ e Ep i 0 T j T 9 i 4 ro A i f 4 Wt se s i ru e g 0 s I Pt i , 3 se r i p i re i , 0 m l I i , 2 m i a i hre e t i a j I 0 w 1 i der i u 6 i as __:~::&:~2-{:: L :: ?~_

i

_'__2, e 4 O M o~ oo oB ow o6 oD on oT om om o o_' ( n c I

2Of 6 a.

-4 erug iF e?,f L

45 ESm o a ....,....,....,...g...i....i....i....,,,,,,,,,q....i.,n g 8 o 1 ~ mm u 'C g 6 lg d w y ~ a m O .:d' g W 8

m E

W o 5 + W w _ o T m .E i v N W ~ t o 5 E = c E U [', z E T ~'= ~ W o.* ma .s - O .M "q-Q .cH2 .o v m 2g , _g e -e ,g,E m 0 g g S b# m ou 4 y - o E M. ~~ o S ~ 1 a ,,,l ...l.,,,1 I,,1,,,,i.,,,1.,,,l....l ...l.,,,l ,,,l. r j g OIICOI 06 08 O/ 09 OG Ob OE 02 O! O OI-w 9ISd .. b d., ~ n N mM$ L\\PAB%89.4 {

4-6 d3mW o b _...,,,,,,,,,,i....,,,,,;....;,,,,i,.,... ;, ;.. ;.,,, i,,, _, g _- o g-o ~. r g 2 a m r .o 1 .c 8 1 m _ o o o e m a u y n ~ o h n 8 g m u o e l H ~ ~ o' 5 w a o e P Z u o r 5 o= r o u .s c A.._ -.r x2 - o' D ua F g@ y m8 c z- _ a a@ m - ca m n 1-e e, _ o e o =m =- n e 155Eim=- E n ~ s 9 E 1 s a ,,, i,,,, l.,,1,, t t..,, i,,,, 6.,,, i,,,,1... l.,,, I,,,,1,,, r @2 01100I 06 09 OL 09 09 Ob OE 02 01 0 OI-o 915 d.... S d.. 7, Oa 60 C L\\FAl\\96 89.4 I -u

i 4-7 psi depending upon the specific transducer, This is comparable to that expected for water flowing at a velocity of approximately 3.3 ft/sec (1 m/s:c). Hence, the transducers in the experimental system are capable of monitoring significant waterhammer events. As mentioned above, the test data illustrated in Figure 4-1 is for an experiment without drain-down of the supply piping. Figure 4-2 shows a comparable run for 2-inch piping (test

1. 47) in which the inlet piping was drained during the voiding process. In this, it is evident that the dynamics of the voiding and refill transients are much less than those of test #49. This is typical of all the experiments performed, i.e. those conducted without drain-down of the supply piping experience stronger waterhammer events, even though the strongest events are in the range of a few tens of psi. A compilation of all test results is given in FAI report FAI/%107.

It is also instructive to investigate the thermal boundary layer developed in the water flow during the refill transient. This is particularly important since the water-steam interface is the location of the energy transfer that distinguishes between rapid condensation of the steam or slow condensation where the remaining steam is displaced during the refill. With the temperature measurements in the center of the pipe we can obtain a perspective of the thermal boundary layer near the interface. To achieve this, we transform the time dependent temperature behavior into a distance verse temperature behavior using a transformation of L = U (t - ty (4-1) p where Up is the velocity of refill and t is the time of quenching for a particular thermocouple. q This transformation results in a temperature profile similar to Figure 4-3 which shows that the boundary layer behind the quenching surface is at least a few tens of centimeters thick. This would suggest that rapid steam condensation would not exist in the refill transient and the steam would be displaced from the piping as a result of the refill. I:\\FAl\\M H.4

I i l; ,!ll1 IlI) y& n wo d-n ia 0 r d n'_ n: 1 I ^ -..

_:_T:li 1

ht i w i 0 m g l 0 e t 1 sy 6 s i re i 0 a t i 9 w j l ec i iv i r e j I 0 s i 8 dna i l 7 i Is 4 i e g I 0 t t i r T 7 a i r S ) P i C e h E i E t T l l 0 f S o 6 C n R i ~ o i E i ta M ' i 5 t M l 0 n E es M e A r p H T e Ep l u7, I 0 d r Ri 4 t e l T. 4 4 a c 7 i s A 4 i h c s l 0 Wt 4 t. n i - g Pt e v , 3 2 n e i I i ap nip i i 4 7 y l r , 0 4 p l i 4w i 2

1. p t u s

s e i Te i h r t i e j 0 mro 1 i mf i ad i h e rd e u a_ _ (- ._:b:- [:~: t 2 al _ ~ O c Win om oo ow on ow om 2 o O g7 R' i a ~@ C, h-2 L = 4 e rug iF E?f: ll

Il !llI1l'l!l

l!

1l w6 n l. wo d-n ia r 0 d i U:_:~~ 1 h 1 t i i w e m i 0 e g l 0 ts i 1 y i sr e te i 0 w a g I i 9 ec iv a r i es g 0 d 7 6 8 na 4 i ls e I e i i 0 r T g I 7 a i r S i ~ P E i ) 6 e C h T 4 m 0 o t E f g S R i 6 n ( E 4 t o ~ i M e ta M i ^ ~ n 0 e A l 5 e E s H 6 M r 4 p t i I e Ri 6 T r d Ep l r f 0 e T. l 7 6 4 ac A s 4 1 h Wt 4 a - c s 6 0 n i e j - g I 2 n Pt i _,l, 3 i ap i m,r i i a np i 7 y l w 4 p 0

1. p g

I i r 2 t u s s e i Teh 4 r t 6 e r l I 0 m o mf 1 6 ad 6 h e rd 4 te u 3 h _a- ~ ___ ~- alc ~ O Win oCD R og o* o7 oM om 9 O oio' b2 ,o mn: NQ= 4 erug i F =!1$E Il

l l C 4!B4-F igure 4 Qm: Q5 O: 2c ' Oio O No m

• o

,o $o ro no ( c s iW O n 2. c a r ;. 7 _7 : ~ - ~ - - [ b- - _ l t ~ i u e i dr i eh da i i fm l 1 i om 0 i r g e i t r h eT i i e i s s i ut 2 p # 0 p 4 g l i y 7 e i p n ip a i P in2 3 l i t g eI 0 g -i n i s c t h m. 4 W A s 1 7 T ca 4 ,1. g pE l e 0 i iR d ~ r T t 4 e p I H M re 5 s E A e 0 M g n t M a t E io n C R 6 i o S f 0 E g T th C e E ) 4 P a 7 S r ~ i T i l r 0 g ie 4 I ff s 4 l e a 7 n 8 i d 0 g s e r v i c e 9 w 0 a t e r s y 1 s 0 f te g m 0 w i i t 1

: J ~n n.

i h 1 d 0 ra in -do wn 5t j 'll;\\ Il1jl1j \\!(;j l 1

E h PI WATER HAMMER TEST #47 2 t e s t 4 7.pi t om ;iiiijiiiij 4 i.jiiiajiei4 l4 j6 i.ijiiiij.>>>l6 ii.jiiii . 4 i o o e / wr' o tn go ul a_ ? .h ? ~ = = / ~; v o i F" O ~ 7 ~ o i O 2, i e !,,,,!,ieil,, , I i , i I,,, i I,, i i l i i, i l i i,, I i i i e I,,,,- i g 'O 10 20 30 40 50 60 70 80 90 100 110 " TIME ~ "(SEC)" Figure 4-2d Waterhammer Test #47 in a 2-inch scaled representation of the Prairie Island service water system with drain-down included for the supply piping.

iilI lli!\\l! rb n wo d-n ia r 0 d -[ ___- [- 2, h 1 1 t i w m 0 e g I 0 ts 1 ys r te i a g l 0 w i i 9 ec i iv re 0 s g I d 8 na ls 7 I e 4 0 ir i 1 g l 7 r t a ~ P T ) e S C h t E f E g l 0 S o T i 6 n ( o i ~ i R i t ta E i 0 e n M g 5 e E s M i M rp A e I e H i T r i ~ d Ri g 0 e t l Ep i m 4 a cs a. h T. 7 A i nc 4 i i 0 Wt g I 3 i - g 2 n s i ap i e i I i np Pt i 7 y i i 1 4lp g I 0

1. p i

2 t u s s e a i Te i rht e i e r g l 0 mo mf 1 i i ad i e h e rd a i e u ~ t u b?

- ~ - b - _ ' _._: _:-

alc 0 Win ac ow o o* O o( o-O O O N t t e 2 0 5 n_ * :nO -4 t e rug iF E>!

i! ll l1j;jlll1I!l]f j eG re b mun 0 e ~ 5 d ~ e _. _ ~. _ ~ ~ ~ . ~ uorF + a h t i i w 6 tne l 0 s i n 4 ar t 4 i l l if 6 ,e er 1 a 5 g n f f e iru T l 0 d a I 3 S w E o S lf T R r i E e R ta T w E 4 M E e M 4 e M ht A 'r r 0 o H g I 2 f s R e s E i en t e k T Ai i c i p h W. 6 e r t 5 ey c yt I Pt 6 ai n l u l I 0 yo rat 1 dl 4 i n au uoq i e b ey el t e at 4 m m a i xi x t or o ~ ___~_ 0 pr ~ _ _ - ~ p p ~ p Aa or-oe on ov o O o~ t o: b0s: 3 4 erug iF E?! ll

51 l ' 5,0 : OTHER PERTINENT EXPERIMENTAL INFORMATION 5.1 Configurations Investigated In addition to those experiments which modeled the Prairie Island fan cooler configuration, FAI has performed numerous waterhammer experiments on different configurations related to the geometries of components, in particular fan coolers, for open service water systems. These include one-inch and two-inch diam:ter experimental configurations with elevated configurations which experience column separation. Moreover, the test programs have net steam generation with no reverse flow permitted in the supply riser. Figures 5-1 and 5-2 schematically represent two configurations examined which did not have l drain-down on the supply side. These figures show the test geometries for the two different pipe sizes investigated. Each experimental test matrix included conditions which would result in significant steam void penetration into the discharge piping. For such experiments, the focus is particularly on fan coolers since this is the only service water component that would receive substantial energy addition under the improved DBA conditions. Like the tests performed for the Prairie Island configuration, a typical procedure for these experiments was to set the water flow rate through the test apparatus (from the water supply vessel to the evacuated receiver vessel) to represent a desired steady-state condition. To accomplish this, the gate valve immediately upstream of the evacuated receiver vessel was used in the one-inch pipe diameter experiments, whereas in the two-inch diameter experiments, this same gate valve and that valve immediately downstream of the water supply vessel were used. The two-inch experimental configuration is more representative of the plant condition since the gate valve at the exit of the water supply vessel enabled the experiment to control the water addition rate during refillin a manner more representative of the service water pumps restarting. (In particular, the refill rate cannot be greater than the pump discharge flow.) Without this gate valve, the one-inch pipe diameter expriments could initially have quite large refill rates since only the pipe friction limited the flow between the pressurized water supply vessel and the low pressure steam void region. (For example, the 1" configuration included tests where the I:\\FAl\\96-89.5

5-2 Steam In.ection to Simulate Heat Trans"er in the Fan Cooler f..........s TC 1 p' 3.4 m I l*-11.0 ft. 'F ,? r,_ ~1g1) 3 Operated Ball Valves -TC 2 5.9 m 'S'" Evacuated i N Receiver 2 Vessel -TC 3 Manuel 3 p Bau vaNo e* i V/ orifice M 30 psig h P ~ -28.6 "Hg TC 8 M p, V*

• Steam TC -

g 1< 4_1 Gan**r }p w(.20 %' ) p, y1 C5 TC 7 w va~e 250*F Weter _ 1 ' y 2.3m. (~121 T) 7.6 ft f w W9984007.COFI 1 14-97 Note: All the TC's are surface mounted on the pipe. Configuration 1 1 Figure 5-1 1" pipe diameter waterha.imer experiment. I:\\FAl\\96-89.5

5-3 i Steam in ection to Simulate Heat Trans 'er in the Fan Cooler QP ,.........* TC 1 l 3.4 ml 1 H110 n:

l

?- ( Operated Ball Vab/es --TC 2 5.9 m '8'" Evacuated l Na Receiver Vessel -TC 3 Manuel A 3 p Ball Ve No % do orifice ~70 psig h P - 28.6 "Hg h TC 8 n t TC-9 mie 2.1 m Valve e.s n Steam TC 4-. h Generator 88 1 Y 8 (-20 *C) T4 ., _ _ ___ __, { Manual C5 TC 7 Of Ball Valve ~250 T Water ) ,2.3 m, (~121 *C) ae 7.6 n W994A117.COR 1 14 97 Note: TC's 1,8, and 9 are surface mounted on the pipe all others are sensing intemal to the pipe. Configuration 2 Figure 5-2 2" pipe diameter waterhammer experiment. IaFAh96-89.5

5-4 estimated refill velocity was 22 ft/sec or about 3 times the flow rate under steady-state conditions.) Once steady state conditions were established, the solenoid operated ball valve on the water supply line was ciced to simulata the loss of service water pumping capabilities. 1 Simultaneously, the solenoid operated ball valve on the line from the steam generator was opened to admit steem to the downstream piping configuration. This net steam addition represents the result of heat transfer from the containment atmosphere to the fan coolers for the DBA condition. In the experimental test matrix, the rate of steam generation was varied as was the steam addition interval. All of the experimental configurations investigated included tests in which sufficier.t steam was added to completely void the downstream piping to the gate valve on the receiver vessel. l l nermocouple measurements were recorded along the discharge pipt.4, which enabled one to observe the void progression into the cold piping configuration as well s the water refill transient. For the one-inch experiments, the thermocouples were mounted on the piping surface and for the two-inch experiments, thermocouples TC-2, TC-3, TC-4, TC-5, TC-6 and TC-7, were inserted into the flow stream (pipe center line), nis provided much greater fidelity on the single-phase /two-phase transients within the pipe and greaJy' aided the interpretation of the two-pham flow state during the voiding and refill portiens of the hydrodynamic transient. Once these configurations were investigated, fcz a range of net steam generation rates and duration of steaming, an additional consideration to be addressed was whether an extended length in the horizontal segment of the loop seal for the downstream piping would substantially alter the waterhammer pressures. Therefore, an additional configuration was investigated with extended downstream piping in the discharge piping loop seal. This is illustrated in Figure 5-3. As shown, the 2-inch piping at the bottom of the. loop seal has a length of 23 ft., which corresponds to a length-to-diameter ratio of approximately 138. This is more than sufficient to initiate condensation induced waterhammers from a stratified (steam over cold water) situation if this could occur for the parameters investigated. With the other 2-inch configuration, this extended downstream length also has thermocouples in the center of the flow stream to monitor I:\\FAl\\96-89.5

r:*Gg 4 (r . 7 <-rc a rce I r ~ ee - rc. Stea n injection to Sanulate Heat f. e f, p-Transfer in the Fan Cooler [ g e e m.w --q. rC t = h e 1 l te s l O i i l t -rc, U~g Y.), ru Arn wa I g '"""" Evacuated F Receiver O n--. - ~~ m vessen u, lii; 6. s:ae N - Tc-1 s: .na 2 } a e. m _-70 peg .ac., (-148 'C) se W e.s abe HHe = j e

t. sew 3

r.. --rc4 g% ( q,3 4 e a1% 2 cxm i 1 Figure 5-3 Experimental configuration with an extended horizontal loop seal for investigating possible waterhammer conditions in the service water system. o__ _ _ _ _.

56 the progression of the steam void during the voiding phase and the water front during the refill transient. 5.2 S.asnaarv of Results All of the above configurations were studied for a variety of net steaming rates as well as intervals for steam addition. As will be discussed, waterhammer events were observed, some during the voiding phase and some during the refill phase (discussed in the next section). However, as will be presented, these detected pressure increases of tens of psi and all were bounded by the column rejoining pressures for the refill velocity.- The first concern for such experiments is that the appropriate pressure increases could be monitored. In all of the experiments performed this was tested at the end of the experiment by rapidly closing the downstream manual ball valve and monitoring whether the measured pressures are sufficient to stataate the water flow by a single increase (waterhammer) given by i AP = p, c, U, (5-1) In this equation p, and c, are the water density and sonic velocity with U, being the steady-state waw velocity before the valve closure. If the velocity in this equation is replaced with the refill velocity, the calculated pressure is twice the column rejoining pressure assuming one water column impacts on ather. Specifically, the column rejoining pressure is one-half the stagnation pressure. As will be shown, the resulting experiments always demonstrated the capability to measure such pressure increases at the end of the experiment. This was a convenient way to assure that the experimental apparatus could monitor the waterhammer events of interest. Furthermore, the measured pressure increases are bounded by the column rejoining preisures. I:\\FAh96-s9.5

5-7 - 5.2.1 Wh" Drat =-Dm of the Su='v Blaas==d Whha==* am F='-+d HM==!al I_-:-r-g-Seal-The experiments performed as if %re was a check valve on the fan cooler supply piping were those in which there was no drain-down of the supply riser. Figure 5-5 illustrates one of the measured pressure histories for an experiment in the 2-inch configuration. As illustrated, this particular transient was initiated by opening the manual ball valve immediately upstream of the receiver vessel, thereby causing column separation, which was followed by opening of the solenoid operated ball valve to establish normal flow through the test apparatus. During this normal flow period, the pressure in the loop seal of the discharge piping is approximately 35 psig. After normal flow was experienced for approximately 5 seconds,-the solenoid valve for the water supply flow was closed and, simultaneously, the solenoid valve controlling steam l addition was opened. his caused a short term pressurization transient as a result of the steam L addition. He subsequent behavior resulted in a depressurization to approximately 15 psig as the steam " void" pushed the water column through the discharge piping into the receiver vessel, which was at a pressure of approximately -28 in. Hg. During the first 5 seconds of the steam addition transient, some small pressurization events were observed, which is typically the interval over which the horizontal run at the highest elevation was voided. After approximately - 5 seconds, the steam void was pushing the water column through the downcomer and there were essentially no waterhammer events recorded, which is expected since this is-a stable configuration. ~ After 22 seconds of steam addition, the thermocouples in the flow stream indicated that the steam void had penetrated through the downcomer to the bottom of the loop seal. At this stage, some waterhammer events of approximately 10 psi are observed and the thermocouples indicate that 5 seconds are required for the steam void to saturate the coolant in the 7.6 ft long . loop seal. Hence, this translates to an ingression " velocity" of about 1.5 ft/sec. At the end of this interval the steam void ingresses into the vertical riser and somewhat stronger waterhammer events are recorded with the largest being approximately 60 psig. This was the largest event observed in any of the experiments performed in this configuration during the voiding phase of the test sequence. After this time the entire test apparatus was voided as indicated by the IAFAl\\96-s9.5

58 a 8 O a i i i i i iie i e i i i i i 1 I .I -N .s ,e. o, b 2 ))g: 3 o g 1 o a N-o o a 35: 1-.s - 7 o r T c c a 1 o f t o '- o e r D, Z o 's c t Q m = m o o 2 e. = - 6 fo y 0-w v g - o' a - D g w ,l r l r 1 8c e a n z eQ --se. - o e o. r - r %8 .,s 3 w a - o 52 m >.c gs my E - o em r i. T m by A 80 E: 5a 9 !s e O as s's I i i i ..a OEE O t. I 01! OS Ol-T m _otsa. _ba_ E.e w l 1:\\FAh96-89.5 c,

5-9 increasing pressure in the system as the piping structural heat sink increases in temperature due l to sustained steam flow. During the interval of complete voiding, the thermocouples in the voided region also demonstrate an increasing temperature corresponding to the increasing system pressure. l After 60 seconds of steam addition, the refill transient was initiated as shown in Figure 5-4. With the water addition, the system pressure in the loop seal decreases from approximately 15 to O principally psig as a result of the termination of the steam addition but also due to atmospheric heat losses and condensation on the cold injection water. During this time, the pressure transducer at the top of the apparatus (P ) observes some waterhammer events that are 1 approximately 20 psi in magnitude. However, these are not observed at the measurement station represented in Figure 5-4 due to the large compliance of the steam void separating the two f pressure measurements. In fact, with the measured refill rate in this experiment, the thermocouples show a rate of 3 ft/sec, the Froude number is greater than unity and one would l expc. the refill process to be proceeding in essentially a " plug flow" manner. For these l conditions, the Froude number is defined as F, = (5-2) M where g is the acceleration of gravity and D is the channel diameter. (The Froude number is dimensionless parameter requiring the use of consistent units in this equation.) Certainly the f ';,<havior observed by the experiment is consistent with this refill characterization since no significant waterhammer events were recorded. In fact the pressure measurement in the loop seal region sees no such events.- Moreover, calculating the column rejoining pressures assuming water properties gives a pressure of 86 psi and 172 psi for the column stagnation. These bound the measured pressures. As the water column arrives at the gate valve upstream of the evacuated vessel, it is traveling at a velocity greater than the normal flow velocity since it is representative of a - pressure difference between the supply vessel and the test apparatus of about 70 psig at this point - in time. Consequently, the pressurization experienced by the coolant when it encounters the I:\\FAl\\96-89.5

5 10 large restriction ivisted by the valve is sufficient to slow the water column to the normal flow velocity. Herm the pressurization history indicates that the system pressurizes approximately 10 psig greater than that which is representative of the pressure in the apparatus under the normal flow condition. As this pressure increase propagates back through the water coolant, the coolant is slowed to that velocity representative of the normal flow. 'Ihis experiment shows that those conditions in which the supply riser remains full would experience waterhammer events during the steam voiding phase, but that these events would be in the range of tens of psi. Not surprisingly, the strongest event tends to be when the void ingresses to a loop seal and the steam void progresses to the vertical riser at the downstream end of that loop seal. However, even in this case the observed events are tens of psi with the largest measured event for all the tests being the 60 psig event illustrated in Figure 5-4. Moreover, in - the refill stage, those systems with refill velocities that correspond to a Froude number of approximate unity would not experience the stratified steam-water configuration. Thus strong waterhammer events would not be possible. 5.2.2 Frner}*nce With an Fre*= dad I*a-th of Diacha-me Pla8== As discussed earlier, the experimental configurations studied also included a loop seal in the discharge piping. This loop seal had a horizontal run at the bottom of the apparatus which was anywhere from 7 to 10 feet in length. During the voiding phase, this is a location where potential stratification of the steam-water phases could occur and perhaps induce " slugging" to create significant waterhammers.' Consequently, this part of the configuration was changed in one experimental setup to include substantially longer piping. An example of the resulting pressure transients for such a configuration is illustrated in Figure 5-5 (a, b, c and d). As illustrated by the pressure measurements at different locations in the apparatus, the measurement at the top of the discharge piping sees some waterhammer events during the voiding of the horizontal leg before the steam void progresses into the downcomer segment of the discharge piping. As the void progresses downward through the 20 ft length of vertical piping, the pressure measurements indicate that no waterhammer events occur. However, when the void progresses into the horizental piping segment, waterhammer events are again measured with IAFAl\\96-89.5

5-11 oc a o g m

_ii,

,iiij,,,,i,it, i,;,,,,,,,j,i,ji.,i 43 o ( .5 ~ .e. m u g a e i 2 W - o 4. o-m o g - o u o 8 1 1 - o ? a s f_ j u j u_ .g q_ a S i 1

e. _ m n=

e 4_- o .a f] g e g W a i H 3 v e = c g; - 8 4 oi o o$< :e w a r

E.

c. 1 r o _ o, j tn a 4 ~ _ m s g e .e_ m nu ee $. o Bi m m a n. _d L r_ n e 6__ m t_ eo 4 L o es - m n o.e I.

o. c a

_ o 3o 39 je g ,,,,I,,,,l.,,,l,,,,1,,,,I,,,ri,,,,!,,,,I,,,,a e= 091 Okt 02I 00I 08 09 Ob 02 .0 02-4, v, 915 d. Id eeME I:\\FAl\\96-89.5

5 13 g E . ij.ii.... j.i,i;..i i .j.,ijii., ii4._ g it o n 1 \\ o._ .1 w O l 1

o 1

L, o g _ o 7 m 0 .c o 'y w _ mn C o 8 a w L { z om E _ 5 v y t i L W - oi i e 7 eW o . R._. r n. s (n. .l o a l o E v .s. 3 w

p

.s i E 3$ ^ eE _ on .s I ap o 18 y 2,8 a 2,...l ..,1,,,,l ...I,,r i... t-,,li,,,l.... o 091 Obl 021 001 08 09 Ob 02 0 02-4 m DISd-.Ed. e:sM lE 1:\\FAl\\%89.5

5 13 8 l! o _i,,ijiiiiji m . j. i i i j i i i j i i i j i.. i j i i i i g i i i._ p Q .g .g o m u P y ~ ~ b ~ O d _ = wo n 1 o E r o 8 r _ g CD ~ U .g 4 g .g I ~~~ CD k .] G U s u a ~ 1 O tA e ) 5 y so r i e: w t g ~ ~ % d ~ - o e w a ~ ~ E 'E- .f .a H _~ 2 0tn, j N u - o 5 -) - e .c ,L e.g g d, a - o .a 8 5 Tt-m 5a a n. m 3 k 8 ~ r

O n.E n

.3.e n N ~ a _ O 8 7 ~ c E,,i,,,,i,,,,i,,,,i,.,.....,1,,,,i.,,,o 2 =@ 091 Obi 021 001 08 09 Ok 02 0 02-

1. A DISd..kd ea be lI' 1:WAl\\96-89.5

l 5 14 l S 1 O .,i i .. ; i..,,,,, i. (, j,.. i,,,,,,.._ m y .a .g _ o . m-O 5 ~~ w o 6 T

O O

6 y, _ O r d F U g ~_- _- O 'I i = _ Q m g U ~ .a g g ~. 2 O (n i'a b v c t C 8 = - oa u g y ,a g .E ( g g Q 8 \\n . 5, N g ~ O 1-v .. = g E M j, $ 'c g o .a x 7 ~ g m n g$ ~ l o @",s gF 3$ ~ - o =v .c y "6 ,,,,l,,,,I,,,,I,,,,I,,,ai,,,,i,,,,I,,,,!,,,,g 091 Okt 021 001 08 09 Ob 02 0 02- ? m .. D I S d -.Sd HJM i E I 1:\\FAI\\96-89.5 \\

Sal 5 magnitudes as large as 60 psi At the steam void continues to ingress through the horizontal piping and into the vertical riser to the downstream gate valve, additional waterhammer events l are recorded in the venical segment but are not transmitted upstream due to the tr. c compliance of the steam voided region. Hence, the inclusion of a longer horizontallength at the bottom of the loop seal altered the timing of events, but did not change the magnitude of the measured waterhammer pressures. 5.3 Ca-Marattaan of the Two Phaea Mew Pattern Durins Voldins l Given the particular loads observed for waterhammer events during the voiding phase, it is instructive to assess the detailed information from the thermocouples. Of particular interest, j are the thermocouples in the long horizontal segment of the extended piping configuration depicted in Figure 5 3. Specifically, the thermocouples in the flow stream give significant insights to the resulting transient two phase flow behavior during the voiding transient.,As l illustrated in Figure 5 3, thermocouples TC6, 7 and 8 are in the downcomer segment with TCs 9,10,11,12 and 13 in the horizonal segment. It la instructive to evaluate the voiding rate of the vertical dischatge leg and the horizontal configuration as well as the subtleties of the system response once the void ingresses into the horizontal portion. Figure 5-6 illustrates the measured voiding rate in the vertical downcomer for Test #6 of this series. Thermocouples 6 and 8 are separatni by approximately 20 ft., hence the voiding rate in this segment is about I ft/sec. For the 2 inch pipe used in this apparatus, the reference velocity in the Froude number is 2.3 ft/sec, I i.e. a Froude number of 0.44. Since thl is a vertical segment, the flow pattern is, of course, stable. Figure 5-7 shows the thermocouple measurements (voiding rate) in the horizontal segment for the same run. The ulal length between TCs 9 and 13 is about 20 ft. As illustrated, the void progression appears to be at essentially a constant rate through the horizontal segment with a velocity of about 0.9 ft/sec, i.e. nearly the same as that observed in the vertical downcomer Consequently, the Froude number is essentially the same and is less than that value which would tend to have the pipe run full based on those conditions developed for water entering a partially filled or voided channel (Wallis et al.,1977). Table 5-1 summarizes the measured behavior, during the voiding phase, for all the tests with steam addition in the IAFAI\\96 s9.5

i' l l j llI\\ I \\ l \\ m* 4 0 ~~ 9 ~ 6 6 ~ P~ ) e 6 6 sa h 9 p i d a 4 o v 23 0 e 1 9l g I 8 h l t CCC 1 g 1 Il1 n a ( . - ~ i ru d W E 6 l e ae 1 s 1 ~ p ) a o C o T l S la a S ~ tn E C o T 0 ~ z T n g I 7 o ',s h ~ ass 8 R I e E ~ 1, o E h s M t i M /, n I i l T e 6 M s n A o p 6 H s e r e l R p 8 u E oc 0 o T g I 6 m A re W h a I a 7 -5 e s rug iF a ~ 0 5 oO Op-g- on O n aom gWJ_$OOOrtw3e-r e i D f *.* 4 ll\\

t \\ 'l i I t si 2; i F WATER HAMMER TEST 116 (11-14 -9 6) 3 o b O l g e a i i i i e a i i e i f ~TC6-TC7~ ---- l ~ I C8~ - -- i o O c, m m l LaJ j _J go [ [ l 0 t O <.n O ,u f ( s a ra: i w l c i \\s, l y a ,/ i ,/ -._... __ j } t / i _~ _ ~un l l o 30 40 50 60 ~T'ME~ t SEC)~ Figure 5-6 Thermocouple response in the vertical downcomer during the voiding phase. t l l

jl !lll i mb re e gs mna mih rP au h D g) r ni e eis 0 0 5 0 0 0 0 0 0 t r d (p 3 2 2 6 3 1 1 1 1 a ui Ws o sV e ka r e eP h P t gn i ' e rid e o d b omV 4 5 4 3 3 3 9 5 4 u r ug 0 C 0 0 0 0 0 0 0 FNi n r e u sa D h P y g t i l l n c aa i d o d e l i e mS 0 1 9 8 8 7 8 1 0 o V up V =o ~ 1 1 0 0 0 0 1 1 1 g io e n HL h id )c t i g oe s n V/ t i f 1 r e( r u t e 5 a D m m e i e 2 5 3 0 0 0 7 6 5 l s x c b n o n 1 1 1 1 1 1 1 1 1 a o r w Ti p o ta p v A D re s b t ) O aS eP) f ( a ef l s 8 2 2 5 7 7 3 2 3 o i l 1 1 1 1 1 1 1 1 i p y e( r r a PR m mu l e S a r n u )a i si 3 7 3 7 5 3 5 5 5 ms s 5 4 5 5 6 6 3 3 3 r (p e oNP y t ic lao) nl c e e mVs 8 1 8 5 9 0 0 9 2 i / 2 3 2 2 1 2 4 3 4 orft Nte( aW ts 3 4 5 e I 2 3 4 5 6 T 1 1 1 l

5 19 extended length configuration. As listed, all of these have Froude numbers much less than unity. It also appears that the void is progressing through the horizontal segment in virtually a one-dimensional manner, which on the surface would tend to disagree with the analyses performed by Wallis et al. However, there are subtleties to the dynamics of this process that are important in evaluating the interface behavior. Dese processes are responsible for the apparent one dimensional response and are typical only of the condensation process during l voiding. Furthermore, they are also important in evaluating the response of the service water system to the imposed transient. - Figure 5-8 illustrates the measured pressures in the test apparatus along with the measured temperatures during the time that the steam void is leaving through the long horizontal run. Dere is a periodic behavior associated with the waterhammer events during the voiding of the horizontal line, i.e. the events occur approximately every one to two seconds. Furthermore, it is also observed that immediately before the waterhammer event, there is a decrease in the pressure measured in the test apparatus, both as observed on the pressure transducers and the thermocouples which are in the steam space. Figure 5-9 provides an expanded plot of the P4 measurement for the event occurring at 61.74 secs. Of interest is the pressure decrease of about 12 psi that occurs during the 100 mnec interval before the waterhammer event. Typically the measured pressure decreases appear to be in the range of about 1/2 atm and the waterhammer event is observed approximately 0.1 seconds later. These observations are consistent with those discussed by Block (1980) when summarizing the observations reported by Block et al. (1977). In Block's discussion, these were related to a " water cannon" experiment where steam was exposed to a large body of cold water, observed a pressure decrease due to enhanced condenaction that was sufficient to induce water motion back into the tube and thereby causing a strong waterhammer event when the water contacted the top of the tube. Of particular note here is the onset of a pressure decrease as the trigger for a waterhammer event as illustrated in Figure 5-10 which is taken from Block (1980). I:\\FAl\\96-s9.5

5 20 WMER 4MMER TEST #6 Cll-14-ES) O l i i 4 i R ~ [

2 $ -

~ i 2 e l g ~ G. o b k f OMMN p@ $p $ p 6 p i ^ l o ~ ~ 0 ~ I i, ,,!,,,,l I i,,, l o 750 55 60 65 70 75 80 85 90 o "T!ME "CSECP o N 'j ii j ,a igii4 g iiiji 4 o : TC9-

• 7 "TCl2-E "TCl3*

~ o e ~. ( G. e, : O w ~o a-5 $ 8 i a_ La i O o l c o x LJ I ~

8

,/ i / a r ......-...-..n.*,.......................**.......'.... o"

,,,,1

,,i,,,,I,,, ,i, ,,i,,,,i....: a 50 55 60 65 70 75 80 85 90

• TIME
• CSECP Figure 5-8 Comparison of waterhammer incidents and the thermocouple response during voiding of the horizontal loop seal.

1:\\FAI\\96-89.5

i;l\\!!\\1jll 1\\IlIl \\^ fB .esae 2 r c i _ 1__ ~ _ 6 n i e r e 9 p g I 1 d , 6 i p i e h , 8 t g I g 1 n , 6 id e c r e r 7 p g I t 1 ne 6 v e i n o i i 6 Y ta z I t C 1 r , 6 e E S + C r ~ o 5 d g I 1 e 6 ~ h t E g ~ 1 n t i w I 4 T o g h I 1 ~ s , 6 t n e v e r 3 e g I m 1 6 m a h i re s t a 2 w g I 1 i i 6 a f o t o l p 1 g I d 1 e , 6 d . b:b~______h~_~_2 n a p x E 1 6 oo O on O om om 2 o o' 9 n -5 2$mee v n_ e rug iF e s # F.*. 1<

5 32 M E T AL, PIPC ntnoise ovtmentS3Ont SPixt ISPIK 5, 0.05 - tecan orr senti g [ t y 15 7 Ek 3 E O E {m 10 - - - - = ~

yP*8M%

I4 h 5 5 0.05 OCPAt35URt!Afloff g li i i i i i i. l o i i O ioo 200 m 400 m o e a a g ovtarnt une seixt x7 (PtAn 0FF scam W p 0.05 _g3 f P ' I M Ps h O v - - - g 10 5 0.05 (f. PRES 80RIZAfleet i i i e i t i i i 0 i il i 1 i 0 soo too 300 400 soo 0 a 1 WEAME33unt SPw NE' ptanorrscush,[1/ 15 ge8 % 0.05 3 w E O E 10 4 { ~ 0CPet$$URftATest i e i t i e a 0 i ii i e-t o eo zoo aos 400 M 0 e a 3 flut (mal Tlut (ms) Figure 510 Simultaneous pressure traces from three slug impact experiments. 1:\\FAI\\96-89.5

5 23 Another important aspect to be observed in the experiments during the voiding phase is the isolation of the upstream regions due to the compliance of the void space. Once the void propagates to the horizontal loop seal, there is a large voided region sepanting the location where waterhammer events are occurring and the fan cooler coils. The experimental results show that the upstream regions do not experience the sharp pressure spikes created by the waterhammer. Therefore, during the voiding pinse, the loads on these regions are essentially the loads due to thermal expansion. 1:\\FAl\\%89.5

6-1 l 6.4-ANALYTICAL CONSIDERATIONS In the FAI experiments, the voiding phase of the various transients experience waterhammers when the horizontal runs are being voided. *Ihese pressure spikes are in the range of tens of psi with the maximum measured being approximately 60 psi. The loads appear I to occur when the voiding rate causes a velocity in the horizontal lines to have a Froude number (based on the water velocity) of about 0.5. As discussed earlier in this report, the measumd water velocities are determined by the thermocouple measurements in the center of the pipe, and for a measured velocity of I ft sec / (0.3 m/sec) the Froude is approximately 0.5. Each waterhammer transient is preceded by a small pressure decrease in the test apparatus (~ 5-10 psi), which is likely due to water attempting to drain and create a separated configuration, thereby exposing the steam environment to cold water and the cold piping wall as illustrated in Figure 61. Specifically, there is a temperature profile along the pipe wall as shown in Figure Gla. Should the water attempt to initiate a separated configuration, those regions along the upper part of the pipe wall and the water interfax would experience cold temperatures and accelerated condensation. In ccmparing the steam flows that would induce the water velocities used in the Froude number, one can obtain perspective on the conditions that influence this behavior. Let us initially assume that there is no significant condensation energy in the water phase and that all of the energy needed to establish the voiding rate is involved in increasing the pipe wall temperature. Again, for this simplified analysis, it is sufficient to assume that the pipe wall has a single temperature and is in equilibrium with the steam. While this is not exactly the case, the time required for-voiding is much longer than the pipe wall thermal time constant in the experiments. Hence, as long as the temperature profile in the pipe wall (as a function of length) achieves a similar behavior, it is equivalent to analyzing the frontal response as if there is no temperature gradient through the pipe wall. Considering that the voiding rate (the water displacement velocity U ) is determined by the rate at which the pipe wall receives energy, the I:\\FAl\\96-89.6

6-2 / // // // // / a y-Steam I Water 1 1 4 f/ // // // // // // // Qi j I i l T i I l Z 1(a) i / ondensation C // // // // /// // // // // ~ 9 Steam a Water: </ // // // // // // // // RMe71M1. coa 1 1544 1(b) Figure 6-1 Influence of water attempting to form a separated configuration in a horizontal piping segment during the voiding phase. I:\\FAl\\96 89.6

63 energy balance can be equated between that provided by the steam condensation and the wall heat capacity, 2 p, # D U, h,, = p, c, w D t (T. - T,) U, (6-1) 4 In this expression, p, and p, are the respective densities of the steam and the steel, U, is the steam velocity supplied to the test apparatus, hr, is the latent heat of vaporizathn, c, is the pipe thermai capacity, t is the pipe wall thickness, with T and T, being the respective values of the steam saturation temperature and the steel initial temperature. Solving for the water velocity l (voiding rate) results in E' E 8 U (6-2) U* = 8 (4 P,, \\ t,,c, (T - T,), l For many of the experiments performed in the FAI facility, the pressure in the apparatus during the voiding phase is approximately 1 atm. The experiments of interest were performed in a 2 inct Sch. 40 pipe such that the test apparatus diameter is 2 in, and the wall thickness is approximately 0.154 in. Considering the initial steel temperature to be 144*F (80*C) subcooled, the resulting voiding velocity is related to the steam velocity by U, = 0.017 U, ; U, = (6-3) Therefore, for a voiding rate of 1 ft/sec (0.3 m/sec), the resulting steaming velocity is 58 ft/sec (17.7 m/sec). 14 us compare this steam velocity with the corresponding flooding velocity as expressed by (Wallis,1969) b 8" P[s Uw = 3 (6-4) h where a is the steam water surface tension (0.069 lb/ft) and pris the den; ty of saturated water. For 1 atm, this results in a velocity of 59 ft/sec (18.1 m/sec). This is virtually identical to the 1:\\FAl\\96 s9.6

6-4 steaming velocity calculated from the measured voiding rate. Estimating the steam supply orifice velocity delivered by the steam assuming no heat losses and choked flow through an opening of 0.5 in, in diameter results in a velocity of approalmately 21J ft/sec (64 m/sec). nus, actual steam velocities of 59 ft/sec are clearly possible. If the steaming rate is sufficient to produce a flooding condition on the water film left behind by the condensation process, that condensate is condnually removed from the pipe wall by the steam velocity and pushed into the wtter surface. Furthermore, the resulting localized pressure decrease that precortes the waterhammer event increases the local velocity and can eventually entrain thl higher temperature water into that region which is attempting to be a significant steam cavity to promote water slugging events. Consequently, there is an inherent process that continually transports hot water from the steam water interface to that region attempting to form a stratified configuration (Figure 6-2). Thus, while small waterhammer events occur, the thermal boundary layer is continually " rebuilt" by the movement of this water into that region atempdng to separate and stratify. De net result is that the thermal boundary layer is continually replenished, with the pipe upper regions supplied by the flooding /entrainment process and the lower regions receiving energy from the hot water which attempts to drain (separate). The not result is that this system could move through the pipe in a one-dimensional manner and appear as "plus flow", even though the velocity of the water voiding rate does not satisfy the Froude number condition developed by Wallis et al. (1977). 1:\\FAl\\96-89.6 )

65 0 I .tll l \\ \\ \\' \\ \\ l t d' Q o 1 a t k 1 o o o y o y \\ m- /0 I h'$ o' ~ &f' 4 o ,2 y o-g t i o D k ! l o a i t s e N I 4 (si !i li' I:WAh96-39.6

7o1 l l 7.0 CONDITIONS WIDCH AFFECT ' IRE LOADINGS 7.1 De Infhtence of Dianched Air la the Water The Prairie Island open service water system draws from the river, passes through various components including the containment fan coolers and is returned to the river. As a result, the incoming water is saturated with alt at one atmosphere with the concentration at saturated conditions being given by Henry's Law, typically stated as l P, = K X, (71) In this equation, P, is the partial pressure of the noncondensable gas over the water, K is the Henry's Law constant and X, is the dissolved gas mole fraction. His is revealing from one particular aspect: dissolved gases always exist in cooling systems wher; the water is exposed to air and no specific actions or processes have been taken to remove the air. For cooling water systems in which the water is taken from a lake, a river, a cooling pond, etc., the incoming water would be saturated with air at atmospheric pressure. Consequently, any process which causes not steam generation, such as heat addition or flashing, would also cause air to exit from solution. Next we will evaluate the extent of dissolved gases in solution. For open systems, we can assess the influence of gas coming out of solution by determining how much gas would have to exit from solution to have a substantial influence on the loads being evaluated. Table 71 shows the values listed in the Handbook of Chemistry and Physics (Hodgman, Weast and Selby,1958) for nitrogen. To simplify the following example calculations, we assume that the dissolved gas is 100% nitrogen instead of 80% nitrogen and 20% oxygen as is the case for air. Note that the partial pressure of gas in this table is given in terms of millimeters of mercury (Hg). For a gas which is saturated at one atmosphere, this partial pressure would be 760 mm Hg. We also note that there is little difference between O'C and 20*C for the Henry's Law constant, nus, for this sample calculation we will use the value - for 20'C (68'F). 1:\\FAl\\96 89.7

72 Table 71 Solubilities of Nitrogen in Water K x 10*7 K = P/X P = partial pressure mm. of Hg X = mole fraction Temp. ('C) t=0* 10' 20' 30' 38' 40' 50' 60' 70' Nitrogen 4.09 4.87 5.75 6.68 7.51 7.60 8.20 8.70 9.20 Table 7 2 is an example of the amount of nitrogen that would be dissolved in water at a tempe ature of 68'F (20'C). Assuming that all the gas were to exit from solution, and that the local pressure was 2 psi (13,793 Pa) the volume that the gas would occupy represents a void fraction of 0.13 (13%). 7.2 Evidence of Noncondensible Ganas Exiting Frana Solution Before evaluating the influence of small gas voids on the mirure compressibility, let us consider experimental evidence of noncondensible gases leaving solution. Consider small wall cavities as illustrated in Figure 71 (taken from Rohsenow,1973), which arise due to sesatches, pits and surface imperfections at grain boundaries. For this discussion it is sufficient to consider the re-entrant cavity shape shown in Figure 7-le. This configuration simplifies the discussion since it removes the wetting angle as a parameter. With an internal pressure P in the flowing i system, water is forced into the wall cavities until the cavity radius is sufficient to support the pressure difference with the water surface tension (a). Inside the cavity, the pressure is the sum of the water vapor pressure P. (T,) and the partial pressure of noncondensable gases, if the temperature of the flowing system is approximately that of the source, the latter value would be eswtially one atmosphere (Pg). '!herefore, the largest cavity diameter (d,) that could support the imposed pressure difference is given by d, = 4a/(P - (P (T,) + P )) (72) IAFAl\\96-s9.7

73 Table 7 2 Example Calculation of Noncondensible Gas (Assumed to be N ) Dissolved in Water 2 X = mole fraction of dissolved gas P = partial pressure of the dissolved gas (in mm Hg) K = Henry's Law constant l X = P/K 7 For a temperature of 68'F (20'C) K = 5.75 x 10 with prolonged exposure to air, P = 760 mm Hg, i.e. I atm. Therefore 1 760 mob d N 1.32 x 10-s X= = 5.75 x 107 mole of water Mass fraction = X Y = 2 x 10 5 kg N (18 kg H 0 3 s 3 1 m of water = 1000 kg 3 Dissolved N mass /m = 2 x 10 2 kg N2 2 For a pressure of 2 psia (13,793 Pa), what would be the gas volume if all the N were to 2 exit from solution? 2 (8314) (293) V: = = P M, N 13,793 (28) 8 V, = 0.13 m Therefore, the void fraction would be approximately 0.13 if all the gas would exit from solution. 1:\\FAI\\96-89.7

74 i [' A x-x i_ N 1 W N i .5 1 ,o t \\ M4 %q j N b .j.x e 1:\\FAl\\96 89.7 ___j

7-5 1his interface (and all similar to it) provide the site (s) for not steam formation if the imposed pressure (Pg) decreases dramadcally (flashing), the water temperature (T,) increases due to heat addidon (boiling) or both. The activation of such cavities la given by the limit of stability condition, (.e. P. (T,) + P - P = (73) i which shows that the noncondensable gases also have an important role in the inception of net steam generation. (Equation 7-2 is the limit of stability since the cavity is stable at this condition but a further increase in the water temperature (T ) or a decrease in the imposed pressure would cause the steam air bubble to grow without bound since the surface tension forces would only decrease as the bubble grows.) Rohnenow summarizes evidence that prolonged boiling (hours) tends to reduce the noncondensible gas innuence (degas the cavities) and increase the effective wall superheat. For the open system configuradon of interest in this evaluation, there would not be such prolonged boiling condidons. As a result, the available nucleation sites would be in equilibrium with the partial pressure of dissolved gases. The above discussion focused on wall cavities but similar considerations hold for entrained particles (or motes) in the cooling water. These also have surface irregularities that can serve as nucleadon sites given the condidons of flashing or boiling in the coolant flow. Therefore, the sites for steam formation can be both along the channel walls and in the water as well. For those situadons with heat addidon in the fan cooler, the fan cooler tube walls are the preferred sites as a result of the temperature gradient from the tube wall to the water coolant. Given that the incoming water is saturated with dissolved air, what other experimental evidence da we have that noncondensible gases leaving solution would influence the net steam generation? Many two-phase critical flow studies have been performed in systems similar to Figure 7-2. Here high temperature water flows through a constant cross sectional area duct with 1:\\FAl\\96-s9.7

Ei D LOW y E BLEED X U = j DEMINERALIZER p DELIVERY LINE HIGH x DELIVERY 6A A X FILL HEATER X BLOW-LINE DOWN NITROGEN g VESSEL BLEED LINE b I ~ TEST M! TRANSDUCER SECTION I = Figure 7-2 Experimental apparatus for two-phase critical fle d= to flashing (taken from Henry, Fauske and McComas, 1970a). l

77 the coolant pressure decreasing due to wall friction. Considering only steam and water, one would expect the pressure to decrease linearly until it is equal to, or less than, the saturation pressure corresponding to the water temperature. At this condition, flashing would begin and the pressure would decrease at an increasing rate as the steam water mixture was formed. However, as shown in Figure 7 3a, the pressure profile deviates from the linear pressure decrease at a pressure signincantly greater than the saturation pressure corresponding to the water temperature. Moreover, gamma ray densitometer measurements of the two phase density (Figure 7 3b) show that the void fraction has increased to 0.4 at the location where the pressure reaches the saturation value (P (T,)). His measured behavior is due to dissolved gas in the water exiting from solution as the pressure approaches the saturation value, in this experimental system, the water was demineralized and saturated with air at one atmosphere. Hence, the air still exhibited a significant partial pressure after the water was heated to temperature of 300'F (149'C) to 400.'F (204'C). Derefore, the dissolved air influenced the incipient flashing condition and the magnitude of the steam generation. It is also interesting to investigate other suwh experiments that investigated low quantity, high velocity flashing flows. Pressure profiles from two other studies are shown (Figure 7 4) and both of these clearly indicate a departure from the linear pressure profile, typical of all water flow, before the pressure reaches P (T.). Consequently, these investigations also experienced the influence of noncondensable gases exidng from soludon. Thus, there are clear examples of the importance of noncondensable gas in the inception and magnitude of the steam formed in the processes of flashing and boiling. For those postulated accident conditions with a LOCA into the containment, the high steam partial pressure causes energy transfer to the fan cooler coils and possibly boiling in the coils. Here again the dissolved air would be important in determining the inception of steam generation and would also be liberated into the steam volume. If subsequent conditions cause the steam to be condensed, the noncondensable gases would again influence the mixture compressibility by decreasing the sonic velocity. I:\\FAl\\96-s9.7 l

78 1 74 g g g. g g g 70 6W o1 62 a' 58 pg 54 so l I I I I I I 32 28 24 20 18 12 8 4 0 LENGTH, In, I I I '58 1,0 pg 58 o P 0.8 -{ 54 0.6 52 a4 c g 50 0.2 0 0 4.0 3.0 10 1.0 0 LENGTH, In. Figure 7-3 Pressure and void fraction profiles for Run 3 of TS R7, taken from Henry, Fauske and McComas (1970m). I:\\FAI\\96-89.7

79 i I I I I i l l i 44 43 l 40 Poet RUN 24 3e (KLINGEBIEL,1964) 3e 34 38 b 1-te se ~ ~ RUN 11 Pw (UN I I ta - to 18 - 16 - I I I I I' 14 It 14. M la 10 4 4 4' 8 0 LENsTH,In. Figure 7-4 Experimental data of Klingeblel (1964) and Cruver (1963) indicating dissolved gases are coming out of solution. 1:\\FAl\\96-89.7

7 10 7.3 Impodance of Small Steam or Gas Vala on the Mhence Sonic Velocits Numerous experiments (Henry, Grolmes and Fauske,1971, Karplus,1958, Semenov and Kosterin,1964 and DeJong and Firey,1968) have demonstrated the influence of small quantities of steam in the mixture compressibility. In particular, these are illustrated in Figure 7 5 for the experimental data of Karplus (1961) for a pressure of 10 psia, the data of Henry, Grolmes and Fauske (1971) for a pressure of approximately 38 psia (Figure 7-6) and the data of DeJong and Firey (1968) illustated in Figure 7-7. In these figures, there are different models represented for expressing the influence of the void fraction (a) on the mixture compressibility and these are Homoneneous Isothermal Sonic Velocity (ay ad= I "' P (7.q (1 - a): 4, (g.,) 3 n, P p, 2 + a (1 - a) b a + Ps, Pt, p, a,2 Homoneneous Adinhatic Sonic Velocity (my ad= I (75) t YP , P: a , (g. a) h (g.,)2 ,(g.,) P t Ps, Pt, p, a,2 where a is the void fraction, y is the isentropic coefficient, P is the system pressure, n is the i isothermal coefficient (generally close to unity) with pr and p, being the respective densities of saturated water and steam. At moderate and low pressures, the liquid compressibility may be neglected for all but the very small void fractions (a = 0.01). This simplifies Eqs (7 4) and (7 5) to b (76) 2 + a (1 - a) b 'Y a, P, 1:\\FAl\\96-s9.7

7 11 1 350 i 3 P = 10 pslo g U. E 250 0 b y200 i o l w \\ o a \\ o 0 o 150 P 8 8 Eq. (3-8) o too '~---~~~s E o 50 0% I i 0~ 0.10 0.30 0.50 VOID FRACTION.G w. Figure 7-5 Comparison of the homogeneous adiabatic n odel and Eq. 7 5 with the steam-Water data of Karplus. 1:\\FAhS89.7 .)

7-12 R 4 o c 0 d 0 000 j e E .c oo So O ~ ~ o O o 4 e O O O C s3 o o m. a O ? o, 3 6o .o p q O ) W = Q t a oc 0 l3 ~ 0 .g 'O o E g a m (b p. <4 .h 6 e 00 A O O O a" E o 3 O o' w N Q 00 N a a 4 Q d N0/ 40 S31110073A NOl1V0VdOWd 30 Ol1VW i .q y 1 1:\\FAl\\%49.7

l 7-13 i i i soo l I I I I j C: TEST SECTION PRES $URE, pgie l 75o 0IQ) O 59 j A 41 l 700 O 99 f d I 6So 4 jsoo 3l5 i E C N q' O i g550 i o q\\ y' o iL O j gsoo C OOO O O O 4 C l l I a\\ i 454 to o O 6\\'8o 9tm R0 t oo \\s O O O a aa a \\k O o s M l' ~ ~ oo o d 0 o 00 o N '% v ~ 3co - og 000 i N-l ) gho i 2So. A. O I I I I I I O S 60 IS 20 2S 30 vo$ MMCT10N(@ X 40I Figure 7-7 Comparison of Eq. 82 and the steam-water data of DeJong and Firey. l 1:\\FAl\\96-89.7 -\\

7-14 and N.' I p7) cr +, (g _ ) h

• s s

Ps, 'Ihese are the homogeneous models for the substantial reduction in the sonic velocity with the oxurrence of small voids. I 'the comparison of these models with the experiments show that these represent a strong decrease in the sonic velocities (the mixture compressibility) with the presence of very small void fractions. In particular, whea the propagation velocity is compared with typical values for the propagation of sound in water (approximately 4500 ft/sec) it is seen that smril void fractions that are much less then 10%, and even for those less then 1%, can reduce the mixture sonic l' velc:ity by an order of magnitude. 'Ihere is the additional expression included in some of the l figures that considers the deviation bheen the measurements and the homogeneous isothermal expression. - Considering the virtual mass of discrete vapor bubbles (Henry, Groln u and Fauske,1971) in a water matrix, one can arrive at an expression that the'two-phase sonic velocity is represented by the function 't = 1.032 + 1.676a (7-8) a. From this further consideration, the results for larger void fractions are better represented than those expressions typically given by the adiabatic and isothermal representations. However, for those void fractions less than 10%, the representation of the homogeneous models is sufficient to indicate the large influence of small steam-gas vo;ds on tie mixture compressibility. 'Ihis substantial decrease in the two-phase mixture compressibility is important when assessing the waterhammer loads throughout a coolant systent like the fan cooler. In particular, the exiting of noncondensable gases fr9m solution creates a situation in which the condensation of steam as the system pressure increases in not sufficient to completely eliminate the existence of a gaseous phase. Of particular note is that the air cannot be driven back into solution as h W AT # 89.7 \\

7 15 efficiently as it exits solution during the imtial vaporization process. Therefore, steam condensation still results in a significant second phase, which is the composite of the remaining noncondensable gases and the steam existing at a partial pressure equal to the water temperature at the gas bubble-water interface. As a result, the " cushioning" of this far more compressible region lengthens the time over which the momentum is transferred from the water flow rate resulting from the restart of the service water pumps and its collision with the existing water slug (s). With this " soft landing", the water slug (s) is (are) accelerated over a much longer time thar. that which would be considered by a water slug impinging on a second water slug with no compressibility associated with the two slugs other than that of the water alone. Such a " soft landing" enables the acceleration to be slower and the loads to be distributed over a longer time such that the influences on turns in the piping configuration are minimized. Thus, the respective loads on piping supports and hangers are also minimited. 7.4 Nancondensible Gaans Exiting Sohntion l For the Prairie Island open service water system, the coolant is at I atm and would be I exposed to subatmospheric pressures with the interruption of service water pumping power. Given the specific pressures, this could be sufficient to force noncondensible gases to exit from solution. To examine the extent of the local void fractions that could be generated as a function of the system pressure, FAI performed simple experiments to assess the possible influences that would be created given column separation conditions. While these experiments demonstrate the behavior at mduced pressures, qualitatively the same prarmee occur with sufficient heat addition to approach local saturation. Figure 7-8 illustrates tne apparatus used to investigate the void fractions created by suhema=heric pressures imposed on room temperature water. (Note that the approximate saturation pressure for room temperature water is 0.33 psia.. In these experiments, the pressures were substantially greater than this value such that the observed void fractions are almost entirely made up of air exiting from solution.) The results from these experiments are illustrated in Figure 7-9. As illustrated, depressurizing room temperature water hegins to exhibit a void fraction when the system pressure is decreased to approximately 6 psia with the void fraction reaching about 0.01 (1%) 1:\\FAl\\96-s9.7 y

7-16 C 3 Scale ~ ~ Manometer = l = 1 F = Level To Vacuum Increase Pump \\ ~ v a _ : _._ _ --~~- -- n .0-o j

0' O

. Water 0 0 0 O RH971025.COR 1 10-97 Figure 7-8 Test apparatus for observing noncondensible gases exiting solution at subatmospheric pressures. I:WAL96-89.7 I e

w Gi Df 0.06 i i i i i i ea D 0.05 l O o Test 1 A Test 2 ts 0.04 O Test 3 d .9 h 0.03 A W f V G A O >o 0.02 O A o g d D A I 0.01 g D A o f e O A 0.00 O 1 2 3 4 5 6 7 Pressure, psia Figure 7-9 Measured void fractions for room temperature saturated 'witn air at one atmosphere, at subatmospheric pressures. , ~,.

7o18 with a pressure of 3 psia. As discussed earlier in this section, such void fractions have a - substantial influence on the mixture pressure wave propagation velocity (sonic velocity), thus this has a substantial bearing on the evaluations for systems that would undergo column separation as well as those systems where noncondensible gases would be forced from solution as a result of heat addition. Obviously, fw the design basis conditions considered in the Prairie Island analyses, both depressurization and heat addition are important mechanisms and should be part of the evaluations. I I:WAM6-s9.7 b .a

8-1 8,tL.WATERHAMMER IDADS FOR COLUMN RFJOINING DURING REFILL WITH RESIDUAL GAS 8.1 Description of the Approach As with any voided system, the return to service of the components must be able to accommodate the refill velocity. The increase in local pressure due to column rejoining can be

===a=*ad with the calculation APw = p c,, U/2 (8-1) In this expression p is the local coolant density, c, is the velocity of sound in the water-gas mixture residing in the tube after column separation, and U is the refill velocity. The factor.of 2 represents one column rejoining with another water column which is static. If the water-gas mixture encounters a solid boundary which brings the entire water column to rest, the above expression remains the same except the factor of 2 is removed. The discussion of two-phase sonic velocity experiments shows that the two-phase behavior is important in considering the pressure and the rise time of the pressure increase. The principle questions are a) how do small voids influence waterhammer pressures, b) what is the value that should be used, as well as c) what is the experimental basis. An important element in this assessment is the influence of noncondensible gases that have exited from solution. Consequently, the relevant experimental database must come from experiments in which the water is essentially saturated with air at 1 atm. Considering the available database, there are only a limited numbee of experiments in which pressure increases have been monitored for different twophase flow conditions. One such set of experiments are those reported by Sweeney and Griffith (1992) in an apparatus that is illustrated in Figure 8-1. In these experiments, flow was initiated with a pressure difference of 10 psi between the upstream tank and the atmospheric discharge with the principle variable being the temperature of the water as it left the upstream tank. As shown in this figure, water was added to the tank and heated by I:\\FAl\\96-49.8

82 g - Steam 'or aosring 314 #efer 9 A d 4ge m,,, ......,,,,n, 3 ~.,,,,u,,2e,. L. "?s fem ge r e f ur e wdh v.i.e q Void Wo.Sutin, $1stien Pressur. Trans duce r 0 1 f.or.in-aj 3 l s I a.nv.i... @ @. @ ir moei a.no I M' A Figure 8-1 Schematic of the test apparatus showing the void measuring section in place. For ( the pressure-time readings that section was not there. Different lengths were accommodated by changing the lengths of the straight sections of pipe. I: TAB %89.8

8-3 steam addition, ne waterhammer transient was initiated by closing a downstream valve in approximately 40 msec with the peak measured waterhammer pressure reported as a function of the upstream temperature. This information, illustrated in Figure 8-2 as well, shows that the waterhammer pressures can be several hundred psi and are not substantially different for the two different lengths of pipe considered, i.e. 24 feet and 40 feet. As seen in Figure 8-1, two quick closing valves were used in the experiment for some of the experiment's void fraction measurements, which are shown in Figure 8-3. It is observed that the general trend is that the void fractions increase as the temperature increase, but the variations in the data are substantial. The role of these variations is discussed later in this section. Given this information, what kind of interpretation can be developed in terms of the two-phase behavior and the propagation of pressure waves through the two-phase mixture (sometimes called the sonic velocity)? For small void fractions, the two-phase sonic velocuy reduces to C,= _T (82) n l a Pt, where y is the isentropic coefficient for the gaseous mixture. To assess the developed waterhammer pressures, for this mixture encountering a rigid boundary, which is the type of experiment performed by Sweeney arki Griffith, this expression can be substituted into the Joukowski equation as the mixture sonic velocity such that APw = p, C, U, (8-3) n When addressing the pressurization of a one-component (steam-water) mixture, condensation is an important element of the mixture response. For adiabatic (no heat addition) flashing flows which begin as saturated or subcooled liquid, the pressure in the pipe after the valve is closed only needs to increase to the stagnation pressure to result in condensation of the steam. Hose situations in which the flow is completely stopped by a single pressure pulse have substantially greater pressures than that necessary for complete seam condensation. Thus, we I:\\FAl\\96-69.8 x.

8-4 Mr !.51

• 40 f t. save

@ 74 ft.sete

1. #4 if. $lte set

. I4 M.3:00tel 6 b Calculated Upper Bound y (Lowest Void Fraction) a = F 1 '= 4 ~ oo pgd

• I.79 "e

v E g 5 l s p l I 38 i 200 = E I l e Calculated Lower Bound. i (Highest Void Fraction) m a e-0 t t-L 4 m 30 9 tie 220,,330 240 ASO t t-t- 100 14 0 130 .g fempetetste Figure 8-2 Comparison of the peak pressure reported by Sweeney and Griffith for the 24 and 40 ft (7.3 and 12 m) sections of pipe and calculated values for the residual void model. The bounding lines result from the measured void fraction boundvies. 1:\\FAl\\96-89.8 -- J

8-5 30s 00 = a ) e e en i e 40 4 e e l' e i e l 3 h e e, e 4c. e ,e e ao = ee e eEsserimem l t; ~ 4 Y i t ios no iso 'C fossereture Figure 8-3 Measured void fractions reported by Sweeney and Grifnth and the boundary values used in these calculations. I:\\FAl\\96-89.4

8-6 can consider two components of the pressure pulse, the first which results in condensation (and is quite small), and the second which stagnates the resulting flow (and is much larger). At the end of first pressure pulse, the void in the mixture is virtually zero. But as will be discussed, a small residual void is left due to noncondensible gases and the small steam mass necessary to be in equilibrium with the noncondendble gases. At this juncture, the density of the two-phase mixture is given by P. " (1 - a,) p, = pc (8-4) where the term a, is the residual void after steam condensation has occurred. The two-phase mixture velocity prior to the condensation condition for adiabatic flashing flows can be approximated using the homogeneous two-phase frictional pressure drop model such that U, = < 2 AP D,'

• 2 AP D" (8.5)

=< , fL Pu ,fL (1 - a,) pg, In this expression, f is the flow frictional coefficient (taken to be 0.02 in these analyse:;), D is 3 the hydraulic diameter of the test apparatus, L is the length of the pipe investigated and op is - the void fraction under the steady-state condition. Furtharmore, the pressure differential operating on a given flow segment in the apparatus used by Sweeney and Griffith is represented as AP. The velocity of the two-phase mixture after the condensation component of the pressure increase is given by U, = (1 - a,) Up (8-6) This in essence is the velocity of the fluid that must be stagnated as a result of the remaining pressure increase. This is similar to the modeling' performed by Karplus (1961) when investigating the pressure amplification, considering compression waves sufficient to condense steam voids. In practical terms, this is the " refill velocity" for the water filling the voided region of the pipe under steady-state conditions. I:7Ah96-49.8 l

8-7 As the pressure wave propagates through the mixture, we can represent the pressure in the sonic velocity expression as the average of the initial and final pressures. Since the waterhammer pressure is much greater than the initial pressure, this essentially reduces to P ~ 1 APw (8-7) 2 Substituting these values into the Joukowski equation results in a calculation of the waterhammer pressure that would result from the stoppage of a flowing steam-water mixture. b U,2 (88) APw=2a i Of particular note is that the waterhammer pressure varies as the square of the refill velocity which arises due to including the compressibility of a two phase mixture. In the above expression, the one remaining term to be 9ecified is the appropriate s that represents the system compressibility once condensatior. has occurred and a residual void exists. This particular valu must come from experimental data since it involves dissolved gas coming out of solution as well as some small steam mass remaining in the gas bubbles. To address this, the data reported by Sweeney and Griffith were used to determine the value for A. Specifically, the highest peak pressure measured was used to determine a value used for all of the analyses discussed below. This value for i was determined to be 0.00$ assuming y = 1.4, i.e. a very small void fraction. However, this small void fraction has a significant influence on reducing the impact pressures. When companng the results to the data developed by Sweeney and Griffith, the two-phase void fraction in the test channel under steady-state conditions determines the appropriate refill velocity. As shown in Figure 8-3, there is considerable scatter in the void fraction measurements. To address this, the bounds of the void measurements were used as shown in this figure. Specifically, at a given stagnation temperature, both the minimum and maximum L void fractions were analyzed to assess the respective waterhammer pressures. Through this, the calculated lines shown in Figure 8 2 were developed; as illustrated, these provide a good 1:\\FAl\\96-89.8

8-8 characterization of the observed pressurc increases for the experiments. It is to be noted that the value of i = 0.005 was used for all of the calculations shown in Figure 8 2. l The approach discussed above represents the complete stoppage of a flashing two-phase mixture For those site.ations in which column rejoining occurs, one water column encounters another and the average velocity of the two columns is one-half the initial water velocity. As a result, the waterhammer pressure is one-half of that represented by the complete stoppage of a water column. Hence, the above expression for column rejoining is given by 1 AP,= $' U,2 zg,9) c 4a We can compare this expression to the plant experience at HB Robinson (CPL,1996). In these experiments, column separation and rejoining tests were performed with a refill velocity of approximately 23 ft/sec and a peak pressure of about 550 psi was observed. Substituting the refill velocity into the above equations results in a pressure of 500 psi, which agrees well with the plant experience, nerefore, the plant experience is consistent with the observations reported by Sweeney and Griffith, as well as the representation of the residual void fraction of 0.004 to characterize the resulting mixture compressibility. It should also be noted that at these higher velocities, the waterhammer pressures calculated assuming a residual gas component and those calculated using all water properties are approximately the same. However, they differ substantially at lower velocities. There are several other experiments that h;ve measured the pressure increases resulting from waterhammer events. Typically, these have been performed to investigate the response of steam generator feedwater lines, and consequently have differences with the experiments carried out for typical fan cooler discharge piping. However, these can be compared to the general approach for considenng the influence of small residual voids. Appendices A, B, C and D show the comparisons with the maximum pressures measured in the different experiments. As illustrated, this characterization is consistent with the peak pressure measured in the various geometries. Note that all evaluations were performed with i = 0.005. 1:\\FAl\\96-49.s

91 9.0 CONSIDERATIONS OF DISSOLVED GAS IN THE VOIDED REGION When considering the refilling of a steam voided region once the service water pumps had restarted, an important consideration is the noncondensible gas that exists in the steam void and the influence that it can have on cushioning the impact as two water columns approach one and another. This has generally been termed the " soft landing". In foriaulating this assessment, it is important to evaluate the total amount of steam has been generated since this represents the amount of nonconcensible gas that could reside in the steam bubble and which must be compressed as the two water columns approach each other during the water acceleration transient. A model has been developed to addras this hydrodynamic transient and is documented in Fauske & Associates report FAI/96-116 (FAI,1996) (The fundamental equations which are solved in this model are described in Appendix E). This model represents the inertia of the two water columns, the possible heat transfer from the steam bubble region as it.is compressed and the diffusion of steam through the noncondensible gas layer as the partial pressure of the noncondensible gas becomes significant (later stages of the collapse). Considerations of the plant specific features related to acceleration of the inlet water column by the pumps, the movement of the downstream water column away from the impact region as the two columns approach, the influence of heat transfer out of the steam bubble region, etc. can be represented on a plant specific basis. In contrast to the residual void calculation discussed in the previous section, this model does not consider the compressible characteristics of the liqrid column that is causing the collapse. Thus, the actual plant behavior is likely the combination of the two mechanisms. Therefore, the considerations provided in this section are not to be interpreted as an either/or, but rather as a description of two different mechamsms with each applying in the plant system. However, with the complexities associated with such a detailed compressible flow model, it is easier to evaluate the effects in isolation and then make recommendations associated with the combination of the two effects, f I:\\FAl\\96-89.9

9-2 Of particular note in these assessments is the evaluation of the noncondensible gas mass fraction in the steam bubble at the time that the collapse process is initiated. As discussed previously, the noncondensible gas mass fraction in solution where a system saturated with air at 1 atm is approximately 2 x 10-5 For those parts of the coolant that would be completely vaporized, all of this gas would be volatilized and released to the steam bubble. However, the progression of the steam bubble is greatly influenced by the condensation on the cold pipe wall that is in the discharge piping. Certainly the pipe must be heated to the bubble saturation temperature for the steam bubble to continue its ingression into the discharge piping. While there is a temperature profile through the piping, the thermal time response of the pipe wall is substantially less than the overall bubble progression transient. Therefore, as long as a similar temperature profile is developed in the pipe wall, this can be interpreted as essentially a one dimensional increase in the pipe wall temperature. Evaluations of typical pipe wall thicknesses show that the extent of energy required to promote heating of the pipe wall is an order of magnitude greater than that which is represented by the steam bubble itself. Therefore, the total l mass of water that is vaporized to create a steam bubble ingression is a least an order of magnitude greater than the mass fraction would have been created based upon the steam bubble l volume itself. Hence, in the plant evaluations, this increased mass fraction must be represented. Another imnortaat element in the system response is the characterization of the temperatures surrounding the steam bubble that would cause the condensation process to occur. Here again, the pipe wall is of major importance since this represents a boundary for the condensation process and one which has been heated to essentially the saturation temperature of the steam bubble at the time that the refill process has been initiated. Furthermore, the. downstream water interface has been heated by the condensation process and also is " thickened" by the condensate which has been formed due to the heating of the pipe wall. Moreover, the refilling water which is moving through the fan cooler unit would receive additional energy transfer as it flows through the cooler and would also be substantially greater in temperature than the cold water typical of the service water inlet temperature to the fan cooler region. Lastly, as the water column moves through the fan cooler and the voided discharge piping, the front of the water would receive energy from the voided piping and this would tend to " thicken" the boundary layer in this water column. Hence, the assessments for the plant system should I:\\FAl\\96-89.9

9-3 consider these combined processes to assess a meaningful temperature to represent the condensation process in the collapsing bubble. In this regard, if the downstream water column does not move substantially before the two columns are accelerated to the same velocity, then the temperature for this region should be very close to the initial saturation temperature in the steam bubble when the refill process was initiated. Figure 91 illustrates the calculated pressure increar in the gas volume (in Pa) as two water columns approach and Figure 9-2 illustrates the calcula'ad velocities of the column moving through the fan cooler (Ul) (in m/sec) and that which is initially stagnant in the downstream piping (U2). Figure 9-3 illustrates the relative movement of the two water columns (in meters) as a result of the dynamic pressure calculated due to condensation and compression of the noncondensible gas and steam. (In Figures 9-1, 9-2 and 9 3, the abscissa has the units of j seconds.) - As illustrated, the peak pressure increase calculated is approximately 0.58 MPa (84 psi), but the major result derived from this assessment la that the rate of rise is over about 1 sec. Hence, the dynamic component associa*ed with substantial load distribution throughout the structures, as would be deduced from an acoustic transmission of such a pressure pulse through a water filled system, is virtually eliminated. Derefore, the net loads on various components of the system are hardly different from the steady-state loads that would be deduced. - As a result, the net consequence of the pressorization shown in Figure 9-1 is that the pipe is stressed to this level, which is well within the design basis considerations for the pipe. Those other dynamic processes related to acoustic transmission of pressure pulses through piping arrangements in essence should take approximately 105 of this pressure increase over 100 msee as the pressure / force imbalances that would occur as a result of the dynamic response. With the analysis of the void compression and collapse in the 8-inch pipe, the additional calculations, associated with the compression of the gas bubble and the inclusion of the noncondensible gases that would be resident in this bubble, result in a relatively small pressure increase and a long rise time. His again is typical of what would be expected when the bubble collapse is due to movement of me steam void into the colder piping regions. I:\\FAl\\96-s9.9

9-4 Q g iiiiiiii g iiiii iii j i. iiiiiii j iiiiiiii. j iiiiiiiii j iiiiiii g 4 E 1 S 3i s s E 3 8 =. 8 u e a g b O Q e 9 F r z E-2 "p s ii..

- al.i_ggg!{n{

t-ir!g e iv - 3

b g

.E .s ill ay- =_ m ggg$ga a Bbbl g

• 5-z Er

~ ~ w

1. 9sdl3~

gg!!.gsV{,ggjj" die! e ~ .E ~5-33 vij]d3 3 e i Bi*fe n.up* a 5.a N iiig a!=.!,1biagg E i i. s i se n,99. +v 'l,ii!!!!_ili.li!;!!!',i l -i i 5 ! i i H i t i c d iI i i hii:s ! 5 g ~,, i, i i i i l i i ii i i i i i l i i i i i, i, i l i i i i i i i i, l i i, i e :.. l i i i i i.., P e O 9 5 b E C I O f pl* od Od-d 1:\\FAl\\96-89.9 l

\\ ,1ii!!1 ,1j \\)Ili'l \\l ljll v* )2U ( n mu l g_ 8 o .== E =7 :E E_ E E Z g E :2:E = E : E~r E:E.: -5 I f c I r i te I a i i w I i l tn i l a v 1 i i 7 n ga I s I t '/, 1 e / 1 h v I f l t N 1 o 1 t i s i 6 a \\ ht \\ f f dn i a I l I I i ) I, i g i I I l I i U ,,,I i ( I i I I 5 r 1 I / i e l / I loo I c I M" c. n i n .A o" E .s a i u E .e0 i f n = t W ea e* W te I o F" t.e f.g l M e cln.o c n n h i ets* 4 I t fcu u sulecuFso* t i T m .SsMs F o:r s 't, rctc ' i o t ( g t5c saa,' r s.e i n r os e " etr i f o se

s. tac'2'",: " 5sn l

m st i ts5 m E a n vMom.Ou.i.i="oG(et s s Lo< a Ic 1 i i e l s Ue. 4, s lll R I L t u i Pt IOs u i sat ir,=s:d s l P' e 1 l 3 o u a sr. 5dte I c mal a eet W"s.e.ie.",!'. sa s nr OA 1 r L t ntf. 39m e r. e a.ll R

n ' i0c 2-vt 1

s51 J, l t t s #"Prlue a I i o e:u Osu s 4o.E a v'4".*' h WD U-w e I m e s 1 e e. 9 ,2 ..U=:f 0e l y I o0 0 = - 0e l h t a.E a<0 I c t 0'o i .ll le0 2 f 0o. ll: lo0s 1 o a 0 ll o.l ? cts 6 :s 0.c :'."t 3#s o. s'.' :aess 2 o l ae 2.. l t i r A l c l o e2:,r ).,. n :a :: r.*I. If. e s l l s" a Ae f yOe l e v l r i i l e 1 t l a I w l d I e 1 t l l 1 a u .I 5 E:_:E EiLgE:~:E=1E2E 1 c g==

~ -

l 0 aC nj n.m m g N D.~ ~ D-o o. i i i ND. a 2-9 e rug iF eiBy,.. s o f1

9-6 C gainio in o niijin u o u j uin inigiinisiu jiuinin j uining sc b-E o s 8 s t.0 s 5 's b P s s e \\ 2 \\, '+ 7 T D E i 2 I g i 2 I A S I [ E fr t 7 o 11 I a1 i 2 asd 1 Li, I i r in.qa!!!!3 !1 l, .g ill. e m

r. a

manli, ggg 3

i g i idi!!!!!i!!" !! l l u laln,-!!.i.t an..sa. r at j =.& en m isil__1! d s$ 1 a i !!!ilill be n,9 vi E = sii!!!!il@9 l -i =g aa-oa d I g b $$$iiilladiIiMiN l g ,h 1 1 4 IfIttttttltttttttttlttttttttt t t t tii tit lt tIIt t t t tlt t t t tii t tlt t t t t !It h g SE OE SE 02 St 01 5 0 EX'IX IMAl\\96-89.9

9-7 Table 9-1 Assigned Parameter Values for a Generic Evaluation of Column Rejoining in an 8" (0.2 m) Pipe Parameter Symbol Value Units Pipe Flow Area A 3,14 x 10 2/0.34 2 m /ft2 Pipe Circumference C 0.63/2.1 m/ft Mass of Liquid Column 1. m 3200/7040 kg/Lbm Mass of Liquid Column 2 m2 1600/3520 kg/Lbm Pressure less Coefficient for K 71 i Liquid Column 1 Pressure Loss Coefficient for K 10.5 2 Liquid Column 2 - 8 Supply Pressure P, 8.4 x 10 /122 Pa/ psia Initial Axial Length of Bubble X.(, 20/66 m/ft 2 Liquid Temperature Tr 393/248 K/*F Inert Gas Mass Fraction Y 2 x 104 i i In summary, the peak pressure that is observed in this calculation is much less than that which would be calculated from the Joukowski equation assuming column rejoining at this pressure in an all water system. Moreover, the Joukowski equation implies a rise time less than a maec. 'Ihis analysis shows that these loads are far too conservative and should be viewed in terms of pressure increasing approximately 1 sec. This is a large reduction in the loads on. ii P P ng supports, hangers, etc. It is always helpful to benchmark model calculations to available experiments. One such set of experiments are the MIT bubble collapse experiments. These are benchmarked to the modeling approach used for plant analyses in Appendix E. As illustrated in the appendix, this approach bounds the pressure increases-observed from the experiments. Hence, those assessments used for the plant analyses, if anything are a conservative representation of the dynamic response that would be observed far the imposed conditions. 1:\\FAI\\96-s9.9 9

10-1 l 10.0 _ CONCLUSIONS-To determine the influence of void formation in the Prairie Island service water fan cooler supply and discharge piping under DBA conditions, waterhammer experiments were performed for a variety of possible conditions. These experiments were scaled using a Froude number for refill and incorporated both 1-inch and 2-inch diameter piping for the particular configuration of interest in the Prairie Island design. Specifically this incorporated a loop seal in the section representing the fan cooler. The variables for this test program were: refill rate, rate of steam voiding, duration of steam voiding, drain-down of the supply side piping, and size of the piping. While waterhammer activity was clearly seen in these experiments, the magnitudes of such pulses were typically in the range of tens of psi during both the voiding and refill portions of the coolant transient. Furthermore, the most dynamic of these waterhammer processes occurred with no drain-down in the supply side piping, i.e. when the supply piping was allowed to drain, the dynamics of voiding and refill processes were reduced. In all cases, the waterhammer pulses were less than those that would be associated with stoppage of the refill velocity, and in most cases, were a small fraction of this value. The inherent capabilities of the test apnaratus to measure the pressurization associated with flow stoppage verified that'should strong transients be typical'of the experimental apparatus, they . would indeed be measured. From these experimental results it could be concluded that: 1. The transients associated with voiding and refill for the Prairie Island fan cooler servin water system are bounded by those assessments based upon the refill velocity into voided piping. L\\PAE9449.10 i

10-2 2. The result of drain-down in the supply side piping for the fan cooler is a - reduction of the dynamics associated with waterhammer in the coolant piping, in particular, when comparing experiments with supply side drain down to those without, the test with drain-down experienced less dynamic transients during the refill stage. Furthermore, with the analyses presented in this report, it is clear that noncondensible gases exiting from solution have a sigt.Jicant influence on the dynamic response that would be observed in fan cooler circuits as a result of the imposed conditions. In particular, the noncondensible gases can: l substantially reduce the sonic velocity for the two phase mixture, and - cushion the deceleration of the approaching water column thereby both reducing the pressure and lengthening the time for a pressure increase. Both of these are major contributors to a load reduction in the service water systems. In particular, the assessments for residual void show that for refill velocities less than about 15 ft/sec (4.6 ndsec) the pressure increases for column rejoining will be less than one half of the pressures that would be calculated - for column rejoining assuming all water properties. - Moreover, comparisons of the residual void model with available experimental results'shows that the explanation is consistent with the observations reported for the different waterhammer experiments. Comparisons of the bubble collapse model with experimental results for bubble collapse show that the model. bounds the pre.ssure increases that were observed in the experimental system. Furthermore, these analyses illustrate that the presence of noncondensible gases twt;.w the two water columns will substantially cushion the dynamic response and in particular lengthen the time that the pressure increases imposed on the piping system. Specifically, the pressure increase is calculated to occur over about I second. Hence, the dynamic assessment 1\\FAM649.10 _.__a

10-3 for such evaluations should increase the rise time for the pressure transient to at least several hundred milliseconds, De two mechanisms considered in this report should not be considered as either/or evaluations In fact, both of these occur in all systems in which noncondensible, gases will exit from solution, nus, evaluations should consider both of these as being appropr: ate in a plant configuration, Since the net result of the residual gas model is that the pressure increases are less than a factor of two that would be obtained from assessing the dynamics from the Joukowski equation assuming all water properties, and since the extensive release of noncondensible gas cushions the dynamics by more than an order of magnitude, our recommendation is that the loads on the piping system are at least a factor of two less than those that would be calculated by refill velocities assuming the standard Joukowski equation using all water properties. Furthermore, it is our assessment that his factor of two in itself is a conservative representation (overestimate) of the actual loads that would be imposed on the piping network. 1 r t\\PAMH9.10

11-1 l

11.0 REFERENCES

Bjorge, R. W., end Griffith, P.,1984, " Initiation of Water Hammer in Horizontal and Nearly Horizontal Pipes Containing Steam and Subcooled Water", Transactions of the ASME, Journal of Heat Transfer, Volume 106, pp. 835 840. Block, J. A.,1980, " Condensation-Driven Fluid Motions," International Journal of Multiphase l Flow, Volume 6, pp,113-129. l Block, J. A., Rothe, P. H., Crowley, C. J., Wallis, G. B. and Young, L. R.,1977, "An l Evaluation of PWR Steam Generator Waterhammer," NUREG-0291. Carolina Power & Light (CPL),1996, Presentation at the NEI Industry Meeting, Dallas, October 29,1996. Cruver, J. E.,1%3, "Metastable Critical Flow of Steam-Water Mixtures," Ph.D. Thesis, Department of Chemical Engineering, University of Washington. Davies, R. M. and Taylor, G. I.,1950, " Proceedings of the Royal society", Volume 200, pp. 375-390. DeJong, V. J. and Firey, J. C.,1968, "Effect of Slip and Phase Change on Sound Velocity and Steam Water Mixtures and the Relation to Critical Flow," I&EC Process Design and Development, Volume 7, (July 1968). FAI,1996, " Experimental Data to Simulate Possible Waterhammer Loads in the Prairie Island Service Water System for DBA Conditions", Fauske & Associates, Inc. Report FAI/96-107. Fauske & Associates (FAI),1996, " Bubble Collapse During Pipe Refill," Fauske & Associates Report FAI/96-116. Henry, R. E., Fauske, H. K. and McComas, S. T.,1970s, "Two-Phase Critical Flow at Low Qualities Part I: Experimental," Nuclear Science and Engineering, Volume 41, pp. 79-91. Henry, R. E., Fauske, H. K and McComas, S. T.,1970b, "Two-Phase Critical Flow at Low Qualities Part II: Analysis," Nuclear Science and Engineering, Volume 41, pp. 92-98. Henry, R. E., Grolmes, M. A. and Fauske, H. K.,1971, " Pressure-Pulse Propagation in Two-Phase One-and Two-Component Mixtures," Argonne National Laboratory Report, ANL-7792. i L\\ pal \\M-M.ll

i 11-2 Hodgman, C. D., Weast, R. C. and Selby, S. M.,1958, Hawthaak of Chemistry and PhysirJL A n Av.Refamnce Rank of Chemleal and Phvalcul Data,40th Edition, Chemical Rubber Publishing Company, Cleveland, Ohio. Izenson, M. G., Rothe, P. H. and Wallis, G. B.,- 1988, " Diagnosis of Condensation Induced Water Hammer", NUREG/CR-5220, Creare TM-1189, Volumes 1 and 2. Karplus, H. B.,1958, "'Ihe Velocity of Sound in a Liquid Cor.taining Gas Bubbles," Illinois Institute of Technology Report IIT Report C00-248. Karplus, H. B.,1%1, " Propagation of Pressure Waves in a Mixture of Water and Steam," Armour Research Foundation Report ARF 4132-12. Klingebiel, W. J.,1964, " Critical Flow Slip Ratios of Steam Water Mixtures," Ph.D. Thesis, Department of Chemical Engineering, University of Washington. Kreith, F.,1960, Princioles of Heat Transfer, International Textbook Company, Scranton, PA. Kutateladze, S. A.,1972, " Elements of the Hydrodynamics of Gas Liquid Systems", Fluid Mechanics / Soviet Research, No.1, Vol. 4, p. 29. Nuclear Regulatory Commission (NRC), 1991, "An Integrated Structure and Scaling Methodology for Severe Accident Technical luue Resolution", NUREG/CR-5809 EGG-l 2659 (Draft for Comment). l Nuclear Regulatory Commission (NRC),1996, " Assurance of Equipment Operability and Containment Integrity During Design Basis Accident Conditions," NRC Generic Letter 96 4. l Rohsenow, W. M.,1973, " Boiling," Section 13 of Handbook of Heat Transfer. Edited by l Rohsenow and Hartnett, McGraw-Hill, New York. Roldt, R. M., 1975, "Stenn-Water Slugging in Steam Generator Feedwater Lines," j Westinghouse Research I.aboratory Memo 74-7E9-FLINE-M1. Semenov, N. I, and Kosterin, S. I.,1964, "Results of Studying the Speed of Sound in Moving Gas-Liquid Systems," Teploenergetika, Volume 11, pp. 46-51. Sweeney, E. J. and Griffith, P.,1992, " Water Hammer Due to the Sudden Stopping of a Flashing Flow," Transactions of Om ASME, PVP - Volume 231, p.127. Wallis, G. B.,1969, One-Dimensional Two-Phase Flow. McGraw-Hill Book Company, New York. l i m m m.it

i 11-3 f Wallis, O. B., Crowley, C. J. and Hagi, Y.,1977, " Conditions for a Pipe to Run Full When 1 Discharging Liquid Into a Space Filled With Gas", Transactions of the ASME, Journal of Fluids Engineering, Volume 99, pp. 405-413. 4-0 4' J 4 i 4 i E 4 2 i I:WAh9689.11

A-1 I j APPENDIX A' 'Am Interpretation of the Westinghouse Steam Generator Waterhamuner Experknents j Westinghouse has perform:d a set of experiments to investigate the nature of steam water l slugging in feedwater lines (Roidt,1975). The fundamental experimental results for wave overpressures were characterized in terms of the length of a straight section that would permit slug formation, acceleration and waterhammer. Figure A-1 illustrates a feedwater line i configuration between the feedwater manifold and the steam generator. Of particular interest is that region shown as the " test section length', which runs between a downward turning elbow and the Tee where the feedwater piping connects witn the feedring. If we consider the waterhammer pressures that could be developed for systems with small gas bubbles, the waterhammer pressure is given by U* (A-1) APw= L 2a, where y is the isentropical efficient, p, is the water density, a is the effective void fraction in the water and U, is the water velocity. The effective void fraction must come from experimental information, but other experiments on the stagnation-of a flashing mixture (Sweeney and Griffith,1992) indicate that this would have a value of about 0.004. Of particular interest here is the functional Wies that would be expected if this were applied to the Westinghouse experiments. For this experi, mental system, a water slug is formed in the test apparatus and accelerated from the Tee towards the downward turning elbow as a result of steam condensation in the liquid slug. For the sake of this assessment, it is sufficient to assume that the water slug is of a constant length (L) and this slug must be accelerated through a distance equal to the assumed bubble length (x3). As will be discussed, this initial bubble length is considered to vary as the length of the test section is varied. For this analysis, it is sufficient to assume that the pressure IAFAl\\96-89.A

a-4aA.Asim. Ah4#_ md..edae-Ep-as e 2M eda A.AEA_un.mLQnw4aast_w-,4_A=.e.A .es A'h+.' ..mK.42m6 g4 - em aw a dr me-h mg.h.p3 s m m. pp m .m, aa a man,ma L.s wea..mae a434.w,,,cw.w.c.- A-2 E i .e. r Ji 'C t3 /1 l / \\ /wx g e s i 5s N l m s 3 i = l [ _4 i I x 1 53 .2 b g E = a w f T ) ",w c f 6 4 AM \\f N. r v 4 J3 1 4

A3 l difference accelerating the water slug also remains constant during the acceleration period. l Thus, the velocity developed in the slug is given by E i 2 x, APo j U,, = (A2) Pr Ls where APo is the driving pressure difference and I i 4 s the slug length, dubstituting into the above equation, the waterhammer pressure can be expressed i APw = _ M AP (A3) Y o 2a L [ For this experiment, the driving pressure difference is determined by the steam generator [ pressure and the effective saturation temperature of the incoming water, which is essentially the same for all experiments. Furthermore, the slug length is determined by the role up of the water mass into a slug geometry and should also be virtually the same for all the experiments. Therefore, this equation for waterhammer can be characterized as i l APw=Cx, (A-4) i-where C is treated as a constant for the given experiments and involves all the parameters in Equation (A-3) except for the initial bubble length. 'Ihis result suggests that the waterhammer l pressure varies linearly with the initial bubble length, which as mentioned above, is assumed to vary directly with the test section length. The consequence of this assumption is that the calculated waterhammer pressures would be linearly dependent upon the test section length. Within the experimental scatter, measured waver overpressures reported by Roldt have a linear variation with test section length. In summary, this assessment results in a test of the two-phase representation for f-explaining the waterhammer pressures. The experimental observations are consistent with the f' two-phase representation of waterhammer loads, i.e. the waterhammer pressure varies as the square of the impacting velocity. i 4 - IAFAhE89.A i ~

A-4 References Roldt, R. M., 1975, " Steam-Water Slugging in Steam Generator Feedwater Lines," Westinghouse Research Memo 74-7E9-FLINE-M1, Westinghouse Proprietary Class 2. Sweeney, E. J. and Griffith, P.,1992, " Water Hammer Due to the Sudden Stopping of a Flashing Flow," Transactions of the ASME, PVP - Volume 231, p.127. i i i r 4 i J l I:\\FAh489.A l l

B-1 APPENDIX B An Interpretation of the MIT Steam Colman CoEmpse Experknents A set of experiments illustrated in Figure B-1 have been reported as part of an MIT thesis (Persins,1979). In these experiments, a column of steam was exposed to a reservoir of cold water by quickly opening the v.dve separating the two. Depending upon the water temperature, the subsequent condensation event caused waterhammer pressures like those illustrated in Figure B-1. One important aspect shown in this figure is that for temperatures above 60'C, there was virtually no pressure increase observed. 'Ihe details of the condensation process are complex and the possibility of the breakup of the water surface near the interface and enhancement of interphase heat transfer is a key consideration. However, one can assess the functional dependencies observed such as the variation of the peak pressures as a function of the liquid subcooling. Let us consider the rate at which steam is being consumed as a result of the condensation process. The resulting steam i l velocity is given by D (B-1) U= 8 p, A hr, i where Q is the heat removal by condensation, p, is the steam density, A is the pipe cross-sectional area and h, is the latent heat of vaporization. The heat removal rate due to f i condensation can be expressed by I Q = h, A (T. - T,) = h, A AT. (B-2) where h, is the effective heat transfer coefficient representing the condensation process, T, is sub s the water subcooling. i j the saturation temperature, T,is the bulk water temperature and AT In this characterization, the effective heat transfer coefficient for condensation includes all of the influences of liquid breakup at the interface, entrainment, etc. Similar simplified approaches have been used to represent high velocity steam discharge into water and the effective heat 1:\\FAl\\96-89.B

B-2 2 )t 400 ~'" --==_-_ 350

• e 57.s*

300 I a. - 250 st l x -=t-@ _..._,,N. g,200 - s Calculated S ed Response for ? 'a Two Phase i So ,e e* \\ Mixture o e \\\\ e\\ ** 100 \\ e e.ge e g \\ 50 \\e s \\N-i O = a 20 ,40 60 80 100 Temperature ('C) Figure B-1 Peak steam bubble collapse induced pressure measured in the apparatus shown at the upper right. m

l B-3 removal rate (Kerney et al.,1972, Weimer et al.,1973, Westendorf,1970 and Grolmes,1968). Substituting into the expression for steam velocity we arrive at a calculation of the steam and water velocities due to the condensation process U' = k = U, (B-3) Ps h,, Considering the role of noncondensible gases in the steam, as well as those that would be near the liquid front, considerations of the twophase mixture sonic velocity that would govern the waterhammer process results in a calculation of the pressurization event that is given by IE AP,=< _f, U,* (B-4) ,2 a where y is the isentropic coefficient, p, is the water density and i is the effective void fracti,on ic the water column. Other experiments indicate that this effective void fraction could have a value of 0.004, which substantially reduces the sonic velocity to be used in the impact calculations and also lengthens the time over which the resulting load is imposed on the surface. For the system to be addressed in Figure B 1, the column is stagnated, hence, the formulation for bringing the column velocity to zero is used. Substituting the expression for the effective velocity as a result of the condensation process results in E' APw= (B-5) L 2a, P h,,, ~ t where all the details of the condensation and two-phase response are embedded in the effective heat transfer ccefficient (h ) and the effective void fraction (i). If we assume that these parameters are the same for all the experiments and that the only parameter which changes substantially is the subcooling, we can write Equation (B-5) as APw = C ATi (B-6) l 1:\\FAh96-s9.B

B-4 L i i where C is a constant involving all of the parameters in Equation (B-5) except for the water 1. subcooling, nis characterization tells us that the waterhammer pressures observed for this configuration would vary as the square of the water subcooling, nis result can be compared to the experimental data in terms of the functional variation by fixing the constant C based upon a measured pressure increase of 410 psi (1.8 MPa) at a water subcooling of 70'C. Given this, the resulting functional dependence of the impact pressure as determined by the water subcooling is illustrated in Figure B-1. As shown, this provides a reasonable characterization of the peak pressures in the experimental apparatus, nis exercise shows that the considerations of waterhammer associated with a two-phase mixture behavior is consistent with the observations from this steam bubble collapse experiment. In particular, the characterization using a two-phase sonic velocity results in a waterhammer calculation that is proportional to the square of the impinging velocity. The resulting data from this particular experiment apoear to be consistent with such a representation. References Grolmes, M. A.,1968, " Steam-Water Condensing Injector Performance Analysis With Supersonic Inlet Vapor and Convergent Condensing Section," Argonne National Laboratory Report ANL-7443. Kerney, P. J., et al.,1972, " Penetration Characteristics of a Submerged Steam Jet," AIChE Journal, Volume 18, p. 548. Perkins, G.,1979, " Peak Pressures Due to Bubble Collapse-Induced Waterhammer," SB Thesis in Mechanical Engineering, MIT. Weimer, J. C., et al.,1973, " Penetration of Vapor Jets Submerged in Subcooled Liquids," AIChE Journal, Volume 19, p. 552. Westendorf, W. H.,1970, "A Model for Predicting the On-set of Oscillatory Instability Occurring with the Intermixing of High-Velocity Vapor With a Subcooled Liquid in Concurrent Streams," Paper B5.4, Volume 5, Proceedings of the Fourth International Heat Transfer Conference, Paris-Versailles. I:\\FAl\\96-s9.B

C1 APPENDEX C Consparison of the Residual Vold Analysis With the Water Cannom Expeutments l The literatun related to waterhammer events includes a number of different configurations. One of those is the " water cannon" experiment (Figure C-1) in which a vertical tube is partially submerged in cold water and steam is slowly added from the top. The steam warms the water in the tube and slowly pushes the water down through the tube until it finally is exposed at the bottom of the tube to a large volume of cold water. Experimental observations show that the steam can penetrate into the water pool, cause significant disruption (fragmentation) of the interface, thereby substantial augmentation of the heat transfer area. As a result, the pressure inside the tube is sharply reduced and accelerates a water slug back into the tube where it eventually impacts on the top of the tube generating a waterhammer event. One such set of experiments were performed by Rothe et al. (1977) and are discussed by Block (1980). Figure C-1 is taken from Block and illustrates two features of the experiment that are typical to condensation-induced waterhammer events. The first is that the system initially experiences a depressurization as shown by the three figures on the left. This depressurization is approximately one half of an atmosphere and is the, driving force for the water column to re enter the steam space and be accelerated upward through the tube. As discussed by Block, this depressurization event would cause a slug velocity of 5 to 6 m/sec. This is the velocity that would then impact on the upper end of the tube. This also provides sufficient information to test the calculation for the waterhammer magnitude with residuel void in the water column. In particular, the experimental data of Sweeney and Griffith (1992) indicates that the " representative" residual void should be 1 considered as 0.004. With this value, the residual void expression for the waterhammer pressure increase (APw) caused by stagnating a low void fraction mixture is given as 5 APw = 1.75 x 10 Uw (C-1). I:\\FAl\\96-s9.C

C-2 "'"U 20 mtnoise ovenPRCssunt sPixt l5P'MC l,, 0,0 5 - terAn orr scAtti y / ] 15 w 4g N g o e_ _- _ = ,go p j h' 8MP, i4 A I k 6 5 .O.05 OtentS$URilAflott 3 g o ,ji t i I iie i i O ies 200 m 400 soo o i a a ovtAPatssuRC sping g[0 (Pt AN OFF scAus s 0.05 P*7MPc h g e N3 W h10 ag 0 % - E g4 g -0.05 E M.PAtss0RIZATlost 6 ,gh,g(, i I e i e I i i o O 80 0 too soo 4oo 300 o e a a YpE' W c'.'tAPet:ssunt semp / g 15 go 8 MP, k"g 0.05 WEA8 FF ME t .g N E3 W 3 E g '0 / i; E o w4 4 = g -o.05 8 5 ocentssumi2ATioet g I i I I i 0 i. I it I e t o e zoo 300 400 300 o e a a Tius (mst Tiut (mal Figure C-1 Simultaneous pressure traces from three slug impact experiments. 1 u

C3 where Uw is the water velocity, Given the values mentioned above, which are typical of a slug v'ith the water oensity, which would also be essentially that for a small residual void, the i calculated pressure increases would be between 4.4 MPa and 6.3 MPa. Comparing these with the measured pressure events on the right hand side of Figure C-1 shows that the peak value is somewhat higher than this, but these values.are a good representation of the event which pressurizes the tube for approximately 1 mnec as the compression wave traverses the tube to the longer end and the rarefaction wave returns. In particular, these values adequately characterize i the loading on the tube, l Consequently, this exercise indicates that the waterhammer evaluation using the residual void model is consistent with the observations from the " water cannon" experiment. This does i not provide proof that residual voids were in such experiments, but does indicate that the pressure increases observed are consistent with such an interpretation. i References i, l Block, J. A.,1980, " Condensation-Driven Fiuid Motions," International Journal of Multiphase i Flew, Volume 6, pp.113-129. Rothe, P. H., Block, J. A., Cawley, C. J., Wallis, G. B. and Young, L. R.,1977, "An i Evaluation of PWR Steam Generator Waterhammer," Creare Report TN-251, NUREG-0291. Sweeney, E. J. and Griffith, P.,1992, " Water Hammer Due to the Sudden Stopping of a Flashing Flow," Transactions of the ASME, PVP-Volume 231, p.127. 4 4 Y d I I:\\FAl\\96-9,C y-e rTv- ~

D-1 APPENDIX D Comparbon of the Residual Vold Model With the Cmare Stensa Generator Model Data To evaluate the possible waterhammer loads tnat could evolve in a steam generator feedwater line, Creare performed a set of experiments using the facility illustrated in Figure D-1. In this apparatus, feedwater we added through the feedwater inlet and began to fill the simulated feedpipe which war also filled with saturated steam from the steam generator. As the cold water began to fill the feedpipe, waterhammer events were initiated as a result of strong condensation on the water surface, roll-up of the water into slugs and impact on the end of the feedwater pipe in which the overpressure transducer is located. Prior to the onset of the waterhammer event, the depressurization transducer measured depressurizations in the local pressure of 5 to 10 psi over 100 to 200 msec as illustrated in Figure D 2, which is taken from Block et al. (1977). As shown in this figure, waterhammer events were several hundred psi with the maximum being 700 psi at a water inlet flow of I gpm. As discussed by Block (1980) such depressurizations would result in liquid velocities as in the text /(AP/p,). Consequently, this would be expected to induce velocities cf 5 to 6 m/sec. Figure D-3, which is taken from Block et al. (1977), illustrates the measurements of the waterhammer event for the various water flow rates tested. Experimental data are showr. for both metallic and ac.ylic pipes, and as expceted, the r.::tallic pipe records higher pressures. Also indicated in this figure are the waterhammer pressures that would be calculated considering residual void in the inpacted column for velocities of 5 and 6 m/sec. As illustrated, these pressures, which were based on a residual vold of 0.005, bound the highest pressures that were measured in the apparatus. Consequently, the residual void characterization is consistent with the observations from the Creare steam generator model. Refemtces Block, J. A., Rothe, P. H., Crowley, C. J., Wallis, G. B. and Young, L. R.,1977, "An Evaluation of PWR Steam Generator Waterhammer," NUREG-0291. Block, J. A.,1980, " Condensation-Driven Fluid Motions," International Joumal of Multiphase Flow, Volume 6, pp.113-129, 4 I:\\FAh96-89.D

D-3 FAI PROPRIETARY l l vt33EL PRESSURE CONTROL VALVE r//// / / / / // mxxxxxxxx WE us p / d b d ~~ ~~ ~ ~~ X % SY M ASS AIM i (XI I 3 TRAM INLET A - 12' r QEPRES3URIZATICN / / gygAg ogNgMATOR TRAN20UCER VESSEL SlWULATCM gy p / CVO PMt33URE 75,n / TRANSOUCEA g SIMULATED PttOPlPE / PRE 3SURE y_- g auet ir se l / d o II' =) THERMCCOUPLE QRAIN /

== / y g "C*-RING SEALS VitWING 0 WINDOW U3 CATION PREDWATEA 0 int.ET VESSEL WATER INLIT 4 Figure D-1 Description of the experimental apparatus as reported by Block et al. (1977).

D3 \\ 8 0. CV ERPR ENWE ' OI e 1.0 8* "3P WE(CFF d i3 9 g 1000 ". SCAL El h P

• 700ee4 400 g

/ h

p. e :

c r e00 1 $4 4400 = w 3 f

• 200 j

50tPRy3FJRt OM I I l ' a.40 " 0 0 100 200 300

• )0 0

1 2 3 4 iE(0f Of el.0 gm

• I

,, 3 s ; iC00 - l 3}. j = s00 - 0 3 s00 P * *00 set l C ~ a S$ h400 / ! "' [0EMt330RIIATO { 100 L W.10 : i\\ i ,i,,. o 0 10 0 200 300 400 0 4 1 3 4 . gr M; " Oge 1.0 gp y g 5 -.*00 $ 04CALS) ~ = 2: E s-m = h0 [ %*4309e4 ~~ "y .e " 300 =. d 4,3 200 = 5059RiSEUnl CN 5 100 b ~ i/, (, a.le i .i,,, a i 0 60 0 20 0 300 400 0 1 2 3 4 3 0 " QV ERPR E33#E @ " O g

• 4.0 gen l

"8 PINE (OPP 3 @ f'i 3 ". S00 - s

• S00 p si 2 =.

- 40 i h0

• 30 0 S

~ .S = , ), (, { 8 50tmpsuniz4, 5 IC0 ie o 0 .00 200 300 400 0 l 2 3 4 Tlast,(m seen itM E,(m soel Figure D-2 Initiating depressunzation transients and the con. sequential waterhammer events as reported by Block et al. (1977).

D-4 l 400 i i SASEUNE CONFIGUMTION 6 m/sec 0 WETAL PtPE

• == rus roo o

Calculated Values too o o o 5 m/sec 9 0, C g, see o o Q a:8 i O g 400,- AA AO o o 8 see. a e 4 e e o e 4 A A A A A O f) 200 ~ A O A g edD 100 = 9- ' C i' A 0 _ 1.0 2.0 3.0 4.0 S,0 4.0 WATER FLQW MTE, Q,(tan) OittR9pftstJtt aca vastate WsTER FLON RATES Figure D-3 Waterhammer pressures reported by Block et al. (1977).

E1 APPENDIX E Comiparison of Bubble Collapse Model with Available Test Data A number of cylindrical (slug) bubble collapse experiments were carried out at MIT with an appr.ratus similar to the one sketched in Fig. E-1. The reservoir (or tank) is open to the atmosphere at pressure P. and contains water at temperature To. At time t = 0, the liquid is allcwed to flow out at the bottom ar.d enter a vertical tube containing saturated steam at one atmosphere pressure. In keeping with our objective of testing the bubble collapse model, we make use of the following assurnptions. (1) The liquid advances downward into the vertical tube in the form of an intact liquid column. (2) The initial pressure in the steam column is equal to the equilibrium pressure P,q(To) at the water temperature. (3) The peak pressure measured in the steam column is due solely to die compression of the steam by the advancing liquid column and is achieved at the instant the steam column is compressed to a minimum size and the liquid column begins to rebound. (4) The effect of gravity and wall friction on the rate at which the liquid flows into the tube is negligible compared with the force imposed on the liquid column by the diiTerence between the atmosphere and bubble pressures. It should be noted that Assumption 2 is not consistent with the MIT experiment which was initiated only after the vertical tube was filled with steam at atmospheric pressure. One may defend the validity of Assumption 2 by presupposing that during the first bubble collapse /expnsion cycle the rebounding liquid column leaves behind enough cold liquid in the form ofliquid drops suspended in the steam and/or a liquid film on the inner tube wall to reduce the steam pressure via condensation from atmospheric to its equilibrium value at To before the I:\\FAI\\%s9.E

Eo2 Poo 4 P-e u ? ,~N )( on. QY f;;h':_~ S >.9 k_hg,; L

1. W's.

o Steam P o ugyF1044.COM 1 1748 muum Figure E-1 Schematic of steam bubble collapse appantus, Wicahg gmebg mum colum compressing steam bubble. ...m

E-3 cycle is over, During the next cycle the liquid is driven back into the steam-filled tube by the atmospheric minus equilibrium pressure difference, P. - P (To). Regardless of whether the y above rationalization of Assumption (2) is corrtet or not, this is the assumption employed in the model used to assess the pressures induced by bubble collapse in the cooling water piping network, and it is retained here. Subject to Assumption (4), the momentum equation for the penetrating liquid column is p, X d:X = P, - P (E-1) dt where X is the instantaneous length of the liquid column, #r is the liquid density, P, is the pressure at the tube inlet, P is the instantaneous bubble pressure, and t is time. At the tube entrance we have the pressure loss associated with converting the pressure " energy" P. within the tank into velocity " energy", or in mr.thematical terms g ys P,, = P, + 7 pr (E-2) Now normally we would introduce another differential equation that accounts for bubble-side mass-transfer controlled steam condensation on the tube wall and or the penetrating liquid front. It is assumed that the tube wall in the MIT experiments was insulated and/or heated to prevent steam condensation on its inside surface, During the first half of a steam bubble collapse / expansion cycle, when the liquid column is penetrating into the tube, the liquid front is stable and the surface area it presents to the steam bubble is too small to remove much steam by condensation. Thus, for present purposes we may assume that the total mass of steam in the bubble during compression is constant and the compression process is Isentropic. Thus we have the following equatiottof state for the steam 1 L P = P (T ) (E-3) y o ,L - X. where y is the ratio of specific heats for steam and the ratio 11(L - X) is the initial steam volume divided by the steam volume when the liquid column has advanced to position X (see Fig. E-1). I:WAh96-89.B

E-4 Note from Eq. (E 3) that, as previously discussed, the initial steam pressure (at X = 0) is the equilibrium pressure at the liquid temperature To, Substituting P, from Eq. (E-2) and P from Eq. (E-3) into Eq. (E-1) gives the differential equation for ths motion of the liquid column; namely I Pt XJ = P. - p, - P,(T) (E-4) A first integral of this equation may be obtained by noticing that d:X dX d ' dX ', g d'X, I dA2 (E5) dt2 dt dX ( dt, dt 2 dt where we introduced the symbol A to represent the liquid column velocity dX/dt, Thus Eq. (E-

4) becomes

.l p, X X * = P X - b +c (E-6) 2 y-1 sL - X, where c is a constant of integration De constont c is evaluated by using the initial condition that the slug velocity A = 0 when X = 0 to obtain c= (E-7) Y-1 Den Eq. (6) becomes f p, X i' = P.X - -1 (E-8) As mentioned above the liquid column velocity is zero when X = 0. The liquid column velocity is also zero when the gas (steam) is fully compressed to its maximum pressure P,,x and the column has reached its maximum length X , 'Ihe value of X, is determined from the solution of Eq. (E 8) with X = 0: 1:\\FAl\\96-89.E

E5 p, x, _ L P,(To) r L 3T t -l =0 (E-9) Y-1 ,L-X., The maximum pressure P. is related to X. via Eq. (E-3), or b P. = P (T) (E-10) y Eliminating X between Eqs. (E-9) and (E 10) gives an implicit expression for the maximum bubble pressure upon collapse as a funcdon of the 11guld temperature: 1 1 ' P= ' El P* Y -1 (E11) = P y-1 1 P (T)' ! a y(T), P (To) y P. Equation (E-11) was solved for an isentropic exponent y = 1.33 and atmospheric pressure P. = 0.1014 MPa. 'Ihe theoretical results are compared with the measured bubble collapse pressures in Fig. E-1. The theoretical curve is seen to bound the data. IAFAl\\96-89.E

E-6 i e 400 e o 350

• e 300 Is e

- 250 - e e g E se at 200 - e e o e 1 50 = e e e e e 100 - e eno e seee eo e 50 e N 0 m 20 ,40 60 80 100 Temperature (*C) Figure E-2 Peak steam bubble collapse induced pressure measurements compared with the predicted maximum pressures (solid curve). .}}