ML19351F872

Vol 4 of Analysis of Hydrogen Control Measures at McGuire Nuclear Station.

ML19351F872
Person / Time
Site:	Mcguire, McGuire
Issue date:	02/17/1981
From:	DUKE POWER CO.
To:
Shared Package
ML19351F865	List: ML19351F864 ML19351F869 ML19351F872
References
NUDOCS 8102200410
Download: ML19351F872 (57)

Text

.

O DUKE POWER COMPANY AN ANALYSIS OF HYDROGEN CONTROL MEASURES l

AT MCGUIRE NUCLEAR STATION VOLUME 4 l

l l

l i

l I FEBRUARY 17, 1981 l

8/02.7_oo W D

~

6. EQUIPMENT SURVIVABILITY 6.1 Introduction 6.2 Necessary Equipment 6.3 Fenwal Tests 6.4 Equipment Survivability 5.5 Conclusions l

l i

I I

(

6-1 l

i

I 6.1 Introduction The accident at TMI-2 demonstrated that hydrogen could be generated in amounts greater than required to be considered for design purposes by 10CFR 550.44 and that hydrogen combustion could occur inside the containment of a pressurized I water reactor (PWR). The major concern over hydrogen combustion inside a PWR I

containment is that the resultant pressurn may cause a breach of containment with subsequent release of radioactivity to the environment. The structural analysis presented in Chapter 4 demonstrates that the McGuire containm nt can withstand the pressures generated by a TMI-type accident. Chapter 2 presents an analysis of temperatures and pressures resulting from the burning of hydro-gen generated by a TMI-type accident at McGuire.

A secondary concern over hydrogen combustion inside containment is the effect the resultant temperatures may have on equipment located inside containment.

Results from the Fenwal tests indicate that the thermal effects on equipment exposed to hydrogen burns are not overly severe. Duke Power Company has also conducted an analysis of the environmental effects of hydrogen burns on equip-ment. This equipment evaluation is provided in this chapter.

6.2 Nece:sary Equipment The following two functions must be performed during a postulated TMI-type event: 1. achieve and maintain the reactor coolant system (RCS) in a safe shutdown condition and; 2. maintain containment integrity. A review was per-f formed to determine what equipment located inside containment is needed to I assist in the performance of these two functions. Table 1 provides the results of this review. For the purposes of this review, all support equipment neces-j sary for the listed components to function (e.g. cables and junctions) are con-i sidered to be a part of the listed component. Further discussion of this review l

is provided below.

6-2

{ -- - . - _ - _ . . - - _ - . - . - - . . - . - . - - _ - - . -

6.2.1 Safe Shutdown To achieve and maintain safe shutdown following a reactor trip, four key func-tiens must be performed. These functions are: 1. circulation of reactor coolant; 2. removal of residual heat; 3. boration and makeup of reactor coolant; and 4 control of RCS pressure. This section addresses the methods by which these functions are performed and !he necessary equipment located inside containment.

Circulation of Reactor Coolant Circulation of reactor coolant is provided by natural circulation with the reactor core as the heat source and steam generators as the heat sink. Feed-water to these steam generaters is provided by the normal feedwater system or the auxiliary feedwater system. Steam release from the steam generators travels via main steam and turbine bypass piping to the main condenser where it is cooled, condensed and returned to the condensate and feedwater systems.

Alternate steam release paths are the main steam safety valves and power operated relief valves. Any two of the four reactor coolant loops and steam generators can provide sufficient natural circulation flow.

The equipment located inside containment which is necessary to assure circula-i tion of reactor coolant is listed belcw:

l

1. Steam Generator Water Level Transmitters

2. Pressurizer Water Level Transmitters .

3. RCS Loop Resistance Termperature D?tectors

4. Core Exit Thermocouples All portions of the RCS, the feedwater system, the auxiliary feedwater system, and the mais steam system which are located inside containment contain no l

l active or passive ccmponents, other than those listed above, which are required 6-3 l

(

to operate to assure circulation of reactor coolant. (See McGuire FSAR Figures 5.1-1, 10.3.2 ', 10.4.7-3, and 10.4.7-4.)

Removal of Residual Heat Residual heat is removed via the steam generators with feedwater supply and steam release accomplished as described above. Those portions of the RCS, feedwater system, and auxiliary feedwater system, and the main steam system which are located inside containment and are required to support the removal of residual heat are listed above. (See McGuire FSAR Figures 5.1-1, 10.3.2-1, 10.4.7-3 and 10.4.7 4.)

Boration and Makeup ,

Boration is accomplished using portions of the chemical and volume control system. Four weight percent boric acid from the boric acid tanks may be supplied to the suction of the reciprocating charging pump (RCP) by the boric acid transfer pumps. The RCP injects the borated water into the RCS via the normal charging path. The TCP discharge pressure is high enough to lift the pressurizer safety valves, if necessary, to assure adequate boration and makeup flow.

An alternative boration source is the 12 weight percent boric acid contained in the boron injection tank (BIT). The contents of the BIT can be delivered to the RCS by aligning the discharge of the RCP to this tank while the suction is aligned to the boric acid tanks.

A portion of the normal charging line is inside containment. This line contains two air operated isolation valves which isolate redundant charging paths to the RCS and are designed to fail open upon loss of air or power. Therefore, if either valve is or;en an adequate charging flow path is available.

An auxiliary pressurizer spray line branches off the charging line inside 6-4

containment. An air operated valve is provided to control and isolate flow through this line. This valve is designed to fail closed upon loss of air or power, thus preventing undesirable RCS depressurization while charging the system.

Those portions of the alternate boration paths which are located inside con-tainment contain no active or passive components required to support this func-tion which would be adversely affected by hydrogen combustion.

All other portions of systems required to support boration and makeup are locatec outside containment and will not be exposed to hydrogen combustion.

(See McGuire FSAR Figures 9.3.4-1 through 9.3.4-5, and 6.3.2-1.)

RCS Pressure Control RCS pressure control is achieved through the use of the pressurizer safety valves and those portions of the chemical and volume control system described above. The reciprocating charging pump discharge pressure is high enough to force RCS pressure as high as pressurizer safety valve set pressure, if neces-sary, to maintain the RCS in a subcooled condition.

A failure analysis of the charging flow paths is provided above. The pressurizer safety valves are self contained spring loaded valves and are immune to the effects of hydrogen combustion.

In order to accomplish RCS pressure control, adequate system isolation must be achieved. Potential leak paths as a result of hydrogen combustion effects are:

1. Pressurizer Power Operated Relief Valves Each power operated relief valve (PORV) is norm:11y closed and is casigned to fail closed upon loss of air or lo:s of power.

I 6-5

2. Residual Heat Removal Letdown Lines The residual heat removal letdown line is isolated by two normally closed electric motor operated valves in series. These valves are located inside conta1nment. They are interlocked to close automatically when RCS pressure rises above 530 psig and cannot be opened until RCS pressure drops below 385 psig.

3. Reactor Vessel Head Vent System This system contains redundant parallel paths, each path containing two solenoid actuated valves in series for isolation. These valves are normally closed and are designed to fail closed upon loss of power.

4. Letdown Line This line is provided with triple isolation capability consisting of two air operated valves in series, and a third set of air operated valves which isolate the three parallel letdown orifices.

All of these valves are designed to fail closed upon loss of air power.

5 Excess Letdown Line This line contains three normally closed air operated valves in series. These valves are designed to fail closed upon loss of air or power.

With the exception of those components previously mentioned, no active com-ponents located inside containment are required to maintain pressure control.

! Spurious operation or failure is discussed in Section 6.2.3. (See McGuire FSAR Figures 5.5.7-1, 9.3.4-1 through 9.3.4-5, 6.3.2-1, 5.1-1, 5.1-2, and 5.5.7-1.)

6.2.2 Containment Integrity I The McGuire containment is described in Section 4.1 of this report and discussed in great detail in Section 3.8.2 of the McGuire FSAR. Chapter 4 presents a t

I i

l 6-6 l

conservative analysis which demonstrates that the ultimate pressure capacity of the McGuire containment is 67.5 psig. To minimize the challenge to containment integrity lean hydrogen concentrations must be assured.

Maintaining lean hydrogen concentrations throughout containment requires the air return fans, hydrogen skimmer fans, and the hydrogen igniters. The air return fans assist in providing a turbulent atmosphere which assures the rapid and com-plete mixing of hydrogen with steam and air. In addition, the air return fans provide a source of oxygen to the lower compartment, thus allowing controlled hydrogen combustion to occur there. The hydrogen igniters provide a means of maintaining lean hydrogen concentrations through controlled combustion. The hydrogen skimmer fans prevent hydrogen accumulation in several dead-ended com-partments in the lower containment by drawing from these rooms and discharging to the upper compartment.

Those components which are located inside containment and are necessary to main-

- tain containment integrity are:

1. Containment Air Return Fans

2. Hydrogen Skimmer Fans

3. Hydrogen Igniters No other active components are required to function.

j 6.2.3 Consideration of Spurious Failures l A preliminary review of the effects of spurious actuation of components in con-tainment caused by hydrogen burning has been performed. For the natural circula-l l

tion of reactor coolant and removal of residual heat, there are no components for which spurious actuation would be a concern. The 11tegrity of the reactor coolant system boundaries and the integrity of containment isolation were evaluated.

6-7

}

The integrity of the reactor coolant system boundaries is important to assure that RCS pressure control can be maintained or to assure that an existing breach in the system would not be aggravated. The components of significance in this case are the residual heat removal system letdown line isolation valves, the reactor vessel head vent system isolation valves, the letdown line isolation valves, the excess letdown line isolation valves, and the power-operated relief i

valves. For all of these components except the power-operated relief valves, the i

only spurious failure mechanism identified was the result of very specific shorting of the control circuits. Since all cables to these components are enclosed in steel armor no spurious failures of these components are expected.

For the power-operated relief valves, the spurious failure modes would be a result of damage to the pressurizer pressure transmitter. The pressurizer pressure trans-

mitters are similar to the differential pressure transmitters listed in Table 1 and are expected to survive the effects of hydrogen burning. Therefore, spurious movements of the power-operated relief valves are not expected.

Spurious failures of containment isolation valves are not expected to occur since manual actuation by the control room operators are required to reposition these valves after containment isolation reset. Electrical cables to all applicable 4 components are steel armored and cable failures are not expect:o. In any event, a redundant isolation barrier exists outside containment for all penetrations.

One additional concern with spurious failure would be the ability to maintain l

l a charging path into the reactor coolant system. The only spurious failure.

i mechanisms identified were the result of very specific shorting of control ca bl es . These cables are all steel armored and are expected to survive hydrogen burning.  ;

I 6-8 i

6.2.4 Eouipment Function and Location The function and location of those compenents listed in Table 1 are briefly described below.

Containment Air Return Fans Two 30,000 cfm fans are provided to assure the rapid return of air to the lower compartment following the initial blowdown resulting from a LOCA. This function not only tends to dilute any hydrogen that may be generated, but also allows hydrogen combustion to occur in the lower compartment by supplying a source of oxygen. Burning hydrogen in the lower compartment results in reduced pressures because the ice condenser is available as a heat sink. The air return

! fans are located in the upper compartment and discharge into the lower compart-ment. A more detailed description of the air return fans is provid.'1 in Section 6.6 of the McGuire FSAR.

Hydrogen Skimmer Fans ,

Two 3,000 cfm fans are provided to prevent the accumulation of hydrogen in restricted areas of the lower compartment by cor.tinually drawing air out of these:

areas and discharging it to the upper compartment. The hydrogen skimmer fans are located in the up; .- compartment. A more detailed description is provided in Section 6.6 of the McGuire FSAR.

Hydrogen Igniters See Chapter 3 of this report for c escription and location of the hydrogen igniters.

Steam Generator Water Level Transmitters The steam generator water 'evel transmitters provide indication of the availa-bility of the primary heat sink. Level indication is provided for each steam generator. The level transmitters for each steam generator are located in tne 6-9

_ _ -~ , - - , _ _ _. _ .._- . - _. _ ._ . _ . - , _ _ _ _ -.

- . . .- - -_ = _ _ - _

i lower containment in a separate accumulator room (that is, the steam generator "A" transmitters are located in accumulator room "A"; "S" in room "B", etc.).

A more detailed discussion is provided in Chapter 7 of the McGuire FSAR.

i Pressurizer Water Level Transmitters Indication of the water inventory within the RCS is provided by the pressurizer water level transmitters in conjunction with RCS pressure. These transmitters are located in the lower containment in two separate accumulator rooms. A more i

detailed discussion is provided in Chapter 7 of the McGuire FSAR.

i R 3 Temcerature and Pressure The core exit thermocouples, the RCS resistance temperature detectors (RTD's),

and the RCS pressure transmitters provide information for determining the degree of subcooling in the RC5. Cables for the core exit thermocouples are routed through the lower containment to a junction box located in the instrumentation room and then to the electrical penetration. The RTD's are routod through the lower centainment directly from their loop location to their .lectrical penetra-tions. The RCS wide range pressure transmitters are located in the annulus and are therefore not affected by any hydrogen burns that may occur inside 4 containment.

)

6.3 Fenwal Tests The purpose of the Fenwal tests was basically to determine the ignition per-r fontance characteristics of a glow plug hydrogen igniter in a series of l

environments representative of various areas within :ontainment during a TMI-type l

i event. However, other useful data can be extracted from these tests. This t

secti;n discusses the effects of these environments on the components utilized in the Fenwal tests and how these effects relate to necessary equipment inside the McGuire containment.

l 6-10 1

i

- . . _ _ . . _ . . _ _ _ _ . . - , . _ , _ - . _ _ _ _ _ . ~ . . _ . . . _ . _ . _ . _ . . . _ . . _ . . _ _ . _ . . . . _ . _ ~ . . _ _ - .

Table 2 provides a list of compcnents that were used in the Fenwal tests to characterize igniter performance. After exposure to a wide variety of anviron-4 ments and hydrogen burn sequences, only three component failures were observed.

With exception of the components discussed below, no signs of degradation were exhibited by any of the items listed in Table 2.

TVA Igniter Assembly Nearly all internals showed some signs of surface scorching. This scorching is I attributed to a small opening in the ignitor box near the box lid. The opening was such that it was not covered by the lid gasket. It is postulated that hot combustion product gases entered through the opening and caused the scorchiag.

Credence is lent to this hypothesis by the path of scorch marks on tha 1.1ternal wall of the box eminating from the opening. Another potential path of hot gases into the box was through the box opening fo- the transformer power cable. This opening was not sealed. Other evident effects from exposure to the various tests wer e a hardening of the lid gasket and a slight corrosion of the box exterior surface. At the time of its replacement with the Duke igniter assembly, the TVA igniter still functioned well, demonstrating no adverse effects on its operating performance.

Duke Igniter Assembly The only visible effect of its exposure to hydrogen burn environments was a

" melting" of a small portion of the box lid gasket. This " melting" appears to be the result of hot gases flowing into the space between the box lid over-lap and the box exterior wall. This distance in the area of the " melt" is on the order of 3/32 inch, where the distance in the unaffected areas is approxi-mately 1/32 inch. The gasket material, Neoprene, has a listed maximum service temperature of approximately 3000F. Its reaction when exposed to the hot 6-11

- . , - - . y. - - , , . . ,,v- - - ,.- - - . , ,,y-, , . i.,ym,,%..e,,,-,.y,-..,,_g-w.,. ,,,.,.,,,,-,,,y,,,, ,,.-.wy-,.....v,.., . . -,_-r.,,,,,-w y- .---..... , - ,

combustion gases is not unexpected. However, in spite of the " melt," a seal was maintained as evidenced by no scorch marks on any of the assembly's internal surfaces. (The opening for the transformer power cable was sealed by wrapping aluminum foil aiound the conduit protrusion from the box and the cable.) For both the Duke and the TVA igniters, power was supplied by cable wrapped in a steel mesh armor wich a plastic covering. No visible effects on these cables were observed. The cable performanc > was not affected by the various environ-ments to which it was subjected.

Wood Block Used as a mounting for the glow plug fan, this block experienced a thin browning over most of the block surface. This result was expected, but has no impact on McGuire since wood is not present inside the McGuire contaimnent.

Thermoccuole Lead Wires After the tests of Phase II, Part 3, it was observed that the Teflen insulation I had " burned" off a portion of the wires. This " burning" obviously resulted in an electrical short. Whether this was the result of one burn or its exposure to 30 burns is unknown. However, with a maximum operating temperature of about 5000 F for Teflon and a melting temperature of about 6500F, the effect appears to be a result of chronic and not acute exposure. This hypothesis is based on the fact that the average air temperature was much higher in the Phase I tests than in the Phase II tests. If an acute exposure was the failure mechanism, the failure should have occurred in Phase I.

In any event, this failure l

mechanism cannot occur in McGuire since all necessary cabling is armored. -

i Fan Motors Two failures in the three fan motors used were observed throughout the Phase I and Phase II tests. In both cases the failure mechanism was detachment of a 6-12 i

L

- 2 soldered wire connection. (Test site personnel have verified that the solder used was normal electrical solder, not silver solder.) Whether the failure is a result of chronic exposure, as may be evidenced by fan motor No.1, or an acute exposure, as may be evidenced by fan motor No. 3, is unknown. Arguments could be made for both possibilities. However, the point is moot since solder is not used for electrical connections inside of the McGuire containment.

Electrical connections inside of the McGuire containment are made by wrapping the wire around a lug and tightening with a lug nut. As an added protective measure, electrical connections are enclosed in a protective housing, whether it be a junction box or an instrument casing. The only other observation regard-ing the fan motors was a slight oxidation film on fan motor No.1. This film did not affect the fan motor performance.

In Phase II, Part 4, of the Fenwal tests, several components typical of components found inside of a reactor containment where placed inside the test vessel. These components are listed in Table 3. Additionally, the transmitter casing, the limit switch, and the solenoid valve were instrumented with thermocouples to provide iruerior air and exterior surface temperatures. Also note that the TVA igniter was instrumented the same way for all the tests in which it was used. (A detailed discussion of the Phase III, Part 4 results is presented in Chapter 5.)

0 As expected, the data shows that aT varies from about 20 F at 8 percent hydrogen 0

to about 500 F at 12 percent hydrogen. A aT of approximately 50 F appears to also

' be a good average for the additional instrumented components in Part 4. Since no attempt was made to prevent hot gases from entering any of the instrumented components, it is reasonable to expect these t.T's to rise if the components were sealed to prevent hot gases from entering. All necessary equipment in the McGuire containment is sealed. Thus, these data demonstrate that temperature differentials could exist across equipment casings during short-term thermal transients. Of the other components placed in the test vessel, the only visible 6-13

effects of exposure to hydrogen burns were a very light oxidation film on the paint samples and two " scorch" marks on the unarmored cable. The oxidation film on the paint samples was expected. Since the coatings used inside of the McGuire containment are not combustible, there is no concern over a prolonged exposure to hydrogen burns. The two " scorch" marks are probably better described as scrape marks. There were no indications of insulation melting.

It is possible that the marks were unobserved on the cable prior to its place-ment in the test vessel. In any event, even if the marks resulted from exposure to the hydrogen burn, this has no impact on McGuire since all necessary caeling inside the McGuire contaimnent is armored.

6.4 Equioment Survivability Three environmental concerns must be addressed in evaluating the ability of necessary equipment to survive hydrogen burns. These concerns are: 1) the flame front itself or secondary fires initiated by the flame front, 2) pres-sure effects due to the hydrogen burn, and 3) thermal effects due to the hydrogen burns. The effects of these three environmental considerations are discussed below.

Fire The concerns regarding fire are twofold. The major concern is the actual burning of any part of a necessary component as a result of either the hydrogen flame itself or secondary fires resulting from the hydrogen flame.

The second concern is the added thermal effects on necessary equipment that may result from secondary fires. In addressing these two concerns, Duke Power Company has reviewed the combustibility of the necessary equipment listed in Table 1 and has condicted a preliminary rev :ew of the likelihood of secondary fires resulting from hydrogen burrs.

t

6-14

This review has shown that all necessary equipment is protected from fire.

The transmitters and cable junctions are protected by steci casings. All necessary cable is armored and therefore protected. All cable interfaces with transmitter casings or junction boxes are sealed, thus preventing a flame from propagating to the inside. There are a few small cables located inside containment which are either not armored or do not have armor as the exterior surface. Cable in the first category is non-safety and has either flame resistant or flame retardant surfaces. An evaluation is currently underway to determine the ability of these unarmored cables to wittstand the effects of hydrogen burning. The effects of any predicted cable to.iures will also be evaluated. An example of the second category of cables is the core exit thermocouples. The core exit thermocouple cables are NSGA radia-tion and flame resistant cables. These cables consist of a twisted pair of chromel-alumel thermocouple extension wires, each with a 6 mil nylon impregnated glass braid insulation with abestos fillers. These are then wound with a 2 mil polyester binder tape and a No. 20 7/w T. C. drain wire.

This is followed by 1) a 2 mil aluminum polyester shield tape; 2) two layers of 10 mil FR asbestos; 3) No. 34 galvanized steel armor; 4) a 2 mil polyester tape; and 5) a 45 mil asbestos braid jacket impregnated with flame and abrasion resistance compound type 472 NSGA. The nominal outer diameter is 0.402 inches. Exposure to a propagating hydrogen flame front is expected to result in only a slight scorching of the outer jacket.

This review also indicatec that secondary fires will not result from hydrogen burns. As mentioned above, the small amount of cable in containment that does not have armor as an exterior jackets is unable to sustain or propagate a flame. The foam insulation in the ice condenser is sealed and therefore not exposed to any hydrogen burns. In any event, the insulation is unable to 6-15

-. . . - . _ . . ,--.. . _ _ _ . . _ _ _ ,-_ .. __ _ _ _ . _ __ _ .~.

sustain or propagate a flame. Therefore, the only heat sources in containment would result from the RCS blowdown and any hydrogen burns which might occur.

Pressure The capability of necessary equipment located inside containment to withstanc the pressures generated by hydrogen burns has been evaluated. A comparison of the cressure profiles in Section 2.2 Figures 5-8 with the containment design basis analysis provided in Section 6.2.1.3 of the McGuire FSAR shows that the pressures are cuite comparable. The fact that the equipment is qualified to cressures much higher tnan sho n in Section 2.2, Figures 5-3 adds further assurance tnat survivability is achieved.

One additional concern that arises frcm the pressure transients is the effect of the upper-lower compartment differential pressure on tne air return fans and the ski =er fans. The specific concerns are whether the fan will trip on overload and whether the fan blades will deform and result in fan failure. A -eview of this situation by the fan manufacturer and Duke Power Company resulted in tre conclusion that the fans will survive significant differential pressure effects.

a 4

T'o cases were considered in this evaluation. A case where all hydrogen burns take place in the lower compartment and ice condenser and a case where several burns occur in the upper compartment as well as burns in the lower compartment and ice condenser. For the former case a maximum differential pressure of approximately 2.5 psi is realized with the higher pressure existing in the upper compartment. For the latter case a differential pressure of approximately 4.1 pst is achieved. Although significant hydrogen burning in the large volume of the uprer compartment is not expected, this case provides a more limiting situation and is considered for information purposes only. -The higher pressure j in the latter case also exists in the upper compartment. In both cases the differential pressure exists for a very short period of time.

5-16

, .- - - --. -- .. . . .. - = _ .

The first concern is that a differential pressure would cause overload trips of the air return fans and skimmer fans. In both cases the higher pressure is in the upper compartment. This results in a tendency for the air return fan to be unloaded or driven as a generator. Either situation could result in an overload trip of the circuit breakers for these fans. Since the ski-:mer fans circulate air frCm the lower to the upper Compartments, the opposite effect will be realized on the skimmer fans, that is, the fans will tend to slow down or stop. This situation could also result in an overload trip of the circuit breaker.

A review by the fan manufacturer of the 2.5 psi differential pressure case resulted in their engineering judgment that the fans would continue to operate through the transient.

A review of the fan control circuits shows that these fans start automatically during an accident situation causing significant containment pressure buildup and that for TMI-type accidents they will te running well before the enset of hydrogen generation. While overload trips are included in the circuit breakers for these fans, the breakers have an autcmatic reset feature that closes the circuit breakers within a few minutes in the event they have tripped. There-fore, even if these fans were to trip on overload, they would automatically restart within a few minutes. Additionally, the overload conditions would alarm on the plant computer system providing indication to the operator of the situation.

Since the important function of the air return fans is to provide circulation and mixing of the gases and to provide oxygen to the lower compartment to allow burning in the lower compartment and ice condenser, a momentary trip of these fans is acceptable. The skimmer fans ventilate the restricted areas of the lower compartment. A momentary trip of these fans is also acceptable.

6-17

l The second concern involves the physical deformation of the fan blades which would cause danage to the fans. Analyses by the manufacturer concluded that the air return fan would survive differential pressures well in excess of 10 psi and that the skimmer fans could survive differential pressures in excess of 25 psi.

For the air return fans the failure mechanism would be a failure of the bolt holding the fan blade into the rotor hub. This would cause the blade to con-tact the housing. The failure mechanism for the skimmer fans would be the deformation of the fan housing bringing it into contact with the mosing rotor.

Temperature The computer code CLASIX was used to investigate the response of the McGuire containment to hydrogen burn transients. (See Section 2.3 for a discussion of the CLASIX code.) Input parameters include mass and energy releases from the postulated break, containmat conditiors prior to hydrogen generation, containment geometry, and various ccntainment related system parameters. (A detailed discussion of the McGuire CLASI A inalysis is presented in Section 2.2.)

I The results of the base case analysis are shown in Section 2.2, Figures 1-8.

These figures show a series of ten hydrcgen burr.s occurring in the lower com-partment with one burn occurring in the ice condenser. The analytical models presently in CLASIX are very conse.rvative in that all heat sinks with the exceptions of containment spray and the ice condenser are neglected. Preliminary CLASIX results indicate that in the event hydrogen burns do occur, air tempera-tures inside containment cou.d exceed levels presently considered in the design basis accident (OBA) analysis. In view of this, Duke Power Company has performed an analysis on the thermal response of the components discussed in the previu section when exposed to the thermal conditions resulting from hydrogen burns.

6-18

_- _ _ _ _ _ _ _ - _ _ _ _ . - _ _ _ _ _ _ . _ _ _ ~ . _ _ . . . - _ _ _ , _ _ - _ _ _ . _ . _ . _ _ _ . . _ _ _ - . _ ,

Due to the conservative nature of analytical models presently in CLASIX, any equipment temperature calculations based on CLASIX would not be realistic.

Equipment response to thermal transients has been enveloped by calculating:

1. the short-term response from radiative heat transfer from a propagating hydrogen flame, and;

2. the long-term response by use of an energy balance.

The effect of radiative heat transfer from a propagating hydrogen flame was investigated by selecting a Barton transmitter as typical of temperature sensitive equipment in the McGuire; containment. Radiative heat transfer from hot combustion products behind the flarac front was also considered. The flame and transmitter were modeled as parallel discs with the same center line. These discs were assumed to initially be 25 feet apart. Emissivities of 0.1 for the flame and approximately 0.6 for the transmitter were assumed. The gases follow-ing the flame were assumed to be 30 percent steam and 70 percent nitrogen. Since nitrogen is essentially transparent to radiation, only steam was considered in determining the emissivity of the hot gases. An emissivity of 0.3 was assumed for the gases. Due to the geometry associated with the location of the trans-mitters, a flame diameter of one-half the room width was assumed. Assuming an initial ambient temperature of 160 F0 and combustion at 10 percent hydrogen, a flame temperature of 2053 0 F is obtained from Figure 1. The temperature distribu-tion of the combustion products behind the flame front was determined by calculating the heat flux at a given temperature. This distribution is shown in Figure 2. The temperature-heat flux relationship of the combustion products is shown in Figure 3.

As these figures demonstrate, the temperature decreases very rapidly behind the flame front with the heat flux decreasing at an even greater rate. A con-servatively slow flame speed of 2 feet per second is assumed.

6-19

Conduction is expected to rapidly districute the incident energy throughout the entire casing. When the incident energy is only distributed throughout ,

the whole thickness of the front plate, a front plate temp 9rature of approxi-mately 210cF is obtained. The realistic assumption that all the incident energy is distributed throughout the entire casing results in a calculated casing temperature of approximately 1780F. (In this case, the interior sur-face temperature would be approximately 178cF, and the heat trarsfer to the internals would be negligible. ) Figure 4 shows the time cependence of the casing temperature.

The calculated casing temperature is considered to be an upper bound due to several conservatisms in the analysis. Most significant among these conserva-tisms is the assumed geometry of the transmitter with respect to the flame. By locating the transmitter face parallel to the flame front, the radiation view factor is increased significantly. In reality, the transmitters are located on walls, making the transmitters perpendicular to and not parallel to the flame front.

If the effects of a series of 10 burns were considered to be cummulative, the total temperature reached at the end of 10 burns would be approximately 3400F.

However, when compared to the long-term energy balance discussed below, this value is found to be conservative. Therefore, this calculation is conservative and can be considered an upper bound on the temperature of the transmitter casing.

To determine the long-term thermal effects, an energy balance calculation was performed. Knowing the total mass of hydrogen consumed and the heat sinks inside containment, the temperature rise of these sinks was calculated. As an added i

6-20

. - . _ ..-._,,,.7__,- --. . , .-- -. -, _ , - . - , . - _ . - , _ _ . , . - - - _

conservatism, only structural steel was considered as a heat sink. Calculated temperatures in areas of typical transmitter locations are aporoximately 3000F.

This result verifies the conservatism in the short-term calculation and is well within the temperatures presently considered in the design basis accident / main f

steam line break (DBA/ASLB) analyses.

6.5 Conclusions As a result of the analyses and tests described in this chapter, the following i

conclusions are drawn:

1. The Fenwal tests demonstrate that various components can with-stand and operate during hydrogen burns.

2. The two functional failure mechanisms observed during the Fenwal tests are not applicable to McGuire due to the station design.

3. Calculated pressure transients resulting from hydrogen burns are ccmparable to those obtained frar DBA/MSLB analyses.

j 4. The calculated interior surface temperatures are comparable to those obtained in DBA/MSLB analyses.

5. Sufficient analytical and test data exists to conclude with reasonable assurance that all necessary equipment inside McGuire containment will function as intended during hydrogen burns.

Efforts are co.;c...ulng to refine and reevaluate the analyses described. If in the future, new information calls into question the survivability of a neces-l sary component, Duke Power Company will take appropriate reasures to ensure that compenent's ability to function as intended.

6-21 r

i L

TABLE 1 Necessary Ecuiccent Containment Air Return Fans (1)

Hydrogen Skimner Fans (1)

Hydrogen Ignitors (1)

Steam Generator Water Level Transmitters (2)

Pressurizer Water Level Transmitters (2)

RCS Loop Temperatures (2)

Core Exit Ther occuples (2)

(1) Required for containment integrity.

(2) Required for shutdown and cooldown.

TAELE 2 Components in Fenwal Vessel For Phase I and Phase II Tests Eouicment Number of Test Excosures TVA Igniter Over 30 7

Duke Igniter 20 Wood Block (4" x 4" x 5 1/2")

l Ther occupies 40 Thermocouple Lead Wires a) unprotected 30 b) wrapped in aluminum foil 6 5

Spray Noz:le Fan Motors 20 a) No. 1 20 b) No. 2 1

c) No. 3 f

I i

i l

l

l TABLE 3 Components in Fenwal Vessel For

' Phase III, Part 4 Tests Ecuipment Number of Test Exposures Paint Samples 1

a) on concrete blocks 1

b) on metal slabs 1

SX-Type Cable Unarmored, Insulated Cable 1 3

Namco Limit Switch 3

Asco Solenoid Valve Barton Transmitter Casing 5 Miscellaneous Wiring 1 Fischer Rr_gulator 1 [

i

1 1

FIGURE 1 f

PRESSURE AND TEMPERATURE AFTER

)

llYDR0 GEN-AIR COMBUSTION, CONSTANT VOLUME AND ADIABATIC i

i '

i i '

i i i '

- 8

/

/

/

I 2500 -

INITIAL CONDITIONS 7

P = 100 kPa 21 ATM '

T = 298 K = 25C i

AIR SATURATED WITH WATER VAPOR

,/

/ -

6 2000 -

x 5

j 3 1500 -

n

/

h 'o y

/

/ 4~"

2 -

w TEMPERATURE / -

l 1090 - ', .

/ m 3 ?o

_# /

PRESSURE '

500 ,/ p

- 2o

/ o a

4 I I I

• I i 1 1

, I , I _. a ,

12 16 20 24 28 0 4 8 INITIAL HYDROGEN CONCENTRATION V/O ,

. -__ - . - . - _ _ .- . , -. .~ .. -_ .. -

1 m

1

+

l i - e 9

l

- N 1

I g r

=

w a -

5 u t.-

v o r cc

< w w c.;

4" M w mo N - E C.

u c <w w m - a Cc m w N LO

= >

c c: is N - *w5d

- w c n s w ce r  % w s -  ::= w w w z w ~

C cc a w

c. <

t.

w e i

M l

C r e i

l L

. N l

1 i i i I I i I i I i e i o o e o o o e o e o e o o o

o o

a o

w 8

n o

e o

.n o

a o

m .N o o o

c.

o e

o o

o n

N - - - - - - - - - -

i J o- 3801VB3dW31 l

_ - _ _ _ . . ~ . ..__ ~ __ _ _ _ . , _ _ _ _ , . _ _ _ _ _ _ _ _ _ _ , , _ _ __

- e.

_ N.

l w

_ a u

<x

=

x- -.

cx x 6

y "z C en c= c

=w -

< h-M =c e. i wr - w w = _ - e z r=

= - .

e ru c 6 MN x y

> c.

i s.

w 25 ._

66

& Ln uH *

=c -

m

_ c.

N l

l l

l l

N l I I l I  !

• N

( c o o C. C. .

\ J 5 4 m ~ - o i

g; cas/019 - X013 1Y3H

8 1 1 i

a 6

1 s

4

_. ' 1 i

2 1

_ G N

. I i R

U D

E M I 0 I

1

. T N R S s U D v B N 4 E N i

O C

_ R E E E U G S R T O U A R -

_ G R D 8 8 I E Y E F P H M M I E E I T N O I R

E T

_ T I

M I 6

S N

A R

T I

I 4

I I 2 i

. _ _ 0 0 5 0 5 0 5 8 7 7 6 6 5 1 1 1 1 1 1

_ 'o i

R$i! "

7. CONT,..NMENT VESSEL LOCAL STRUCTURAL RESPONSE TO HYDROGEN DETONATION 7.1 Overview 7.2 Introduction to Structural Analyses

7. 3 Modeling Decisions: Loading, Boundary Conditions, Structural Geometry 7.4 Ecuation Solving and Computer Codes 7.5 Models 7.6 Results 7.7 Coments 7.8 Conclusions 7.9 Canbustion and Explosives Researen Inc. Report l

l 7-1

7.1 Overview The possibility of a hydrogen detonation inside the McGuire containment resulting from a postulated TMI-type accident at McGuire was investigated by Combustion and Explosives Research Inc. of Pittsburg, Penn. The accident flow rates of hydrogen, steam, and air within the containment were considered as a function of time. Hydrogen ar.d steam leave the reactor coolant system as a high speed jet which readily entrains the lower compartment atmosphere. This jet entrainment promotes mixing which is odequate to preclude the fortnation of i detonable mixtures in either the jet stream itself or in localized areas of the containment. The operation of the containment air return fans was con-sidered in this analysis. The results of this analysis are sumarized in Figure 7.11, and a ccpy of the report is provided in Section 7.9.

Duke Power Company has performed analyses to consider the effects of local detonations in the McGuire containment even though investigations show detonable mixtures of hydrogen will not be formed as the result of a postu-lated TMI-type accident at McGuire. These analyses provide further assurance that the McGuire containment can withstand the effects of a TMI-type accident and are discussed in the following sections.

7.2 Introduction to Structural Analyses An investigation of the McGuire Containment Vessel has been undertaken to detemine the structural response to localized hydrogen detonation. Both the upper containment (dome) and lower containment (cylinder) were analyzed for ,

A the pressure transient associated with a localized hydrogen detonation.

1 full description of the McGuire Containment Vessel is provided in Chaoter 4.

l 7.3 Modeling Decisions: Loading, Boundary Conditions, Structural Geometry l 7-2 l

ywmw .,,pewywa m - c -- *ygpm-yp. gwm+ ) gggr ye e am.34--.g., .,r.%39 p. p -,y.9-eyi'T'-M*'*97WMMM-vWWWpwWu'-+tc'7"-M*TW~V*t'z'm't 7N'ye w g9 -p e e ifsic"Ng--wgp 9*'W

E The fluid-structure interaction associated with a detonation loadina was

> idealized as a series of surface loading time histories. These forcing functions interact with the structure only in the vicinity of the detona-ti on. Factors such as angle of wave incidence and variation of pressure with distance from detonation center lessen the effect of the pressure wave on remote structural elenents. These topics are discussed on pp.

276-282 of Reference 1.

Several triangular load-tine histories were investigated. A maximum pressure of 200 psic with a time duration of 500 usec. was assumed. Each of these was J

assumed to originate from detonation of a 6 ft. diameter hydrogen sphere ad-jacent to the containment. All detonations were assumed to take olace within 3 ft. to 6 ft. of the containment shell.

Mathematical models representing portions of the dome and cylinder were used to pennit a more detailed analysis of local response. The influence of the imposed boundary conditions was verified by comparing results with those of an overall model analysis. Appropriate correlation was obtained which verified the assumed boundary conditions.

Axisynmetric geometry was used in the analyses with Fourier series expansions representing variable circumferential loads. The vertical stiffeners were i

j smeared on the cylinder model with an increase in mass density to account for the extra mass of the vertical stiffeners.

7.4 Ecuation Solvino and Comouter Codes a

The equations of motion for a dynamics problem ray be solved by direct inte-gration or modal superposition. These analyses used the direct integration method since it is best suited for short duration loads which could excite 7-3

.- ~ .. - , -.

4 I high frequencies (see p. 252 of Reference 2 and p. 46 of Reference 3). A parametric study of integration step size was conducted to avoid the problems 4

I often associated with this so'ution method, numerical instabilities for ex-ample.

f A verified in-house version of the Wilson-Ghosh computer code was used in the l

analysis. Correlation for part of the deme analysis was carried out with the i

MARC computer code.

7.5 Models Figure 7.1 shows the structural geometry, boundary conditions and material

properties for the cylinder model while Figure 7.2 illustrates the features for the dcme model. Note that symetry of structure and symetrical loading

! in the meridional direction allowed the 3 panel cylinder model to be reo-resented by 1 h panels and syrvnetric boundary conditions. The axisynynetric i dome model simulates local response to a detonation anywhere in the hemi--

I scherical dcme away frcm the springline.

Figures 7.3 and 7.4 illustrate loading time histories I, II, and III. Load-3 ings I and II vary only in the time of occurence of peak pressure. Both of i

these forcing functions were assumed to act on all affected elements (6 ft.

meridional wave front) simultaneously. On the other hand, loading III represents an advancing spherical wave front which reaches outer elements at a later time. Each affected element or node has a different load-time history l

relevant to the instant of detenation. A wave velocity of 6000 ft/sec was used to calculate the lag time for outer nodes. This loading was apph'ed to the l

dome mcdel only.

(

l9 l 7-4 i

(

Figure 7.5 shows the 50 Fourier term representation of the circumferential variation for loadings I and II (cylinder model). Note that this variation extends over approximately a 10o (s10 ft) base with an increasing intensity toward the middle. The meridional variation is restricted to the afore-i mentioned 6 f t. length.

7.6 Results two important paraneters influenced interpretation of results. The effect of integration step size can be seen in Figure 7.6. Studies of dome and cylinder results indicated an optimum time increment of 50 microseconds with a maximum of 100 microseconds.

The second major factor in reviewing numerical results involved assuring that adequate response duration was evaluated. Maximum displacements for short duration loads usually oc:ur after the removal cf the forcing function. Hence, at least one period of free vibration response must be examined to insure max-imum values have been found. Figure 7.7 shows the periodicity of the 'z' dis-placement for r. ode 1 of the dome model (load-history I).

Table 1 shows the maximum stress and displacement results for the time histories which have been investigated. These values were extracted from up to 4000 time step results per analysis and reflect an upper bound for stresses (maximum values r

i for membrane stress resultants and bending stress resultants often occured at different times). Stress values were tabulated for the outer fibers and mid-surface to emphasize the influence of bending. The maximum values for both models occurred in the element at the center of the load well after the forcing function has ended. The stress intensities calculated are those associated with i the listed fiber stresses.

l 7-5 l

i i

Figures 7.8, 7.9 and 7.10 illustrate displacement patterns for the dome and cylinder models at various time increments. These patterns indicate that several modes are participating in the response. Also, the highly localized nature of the dome response is clearly obvious.

Dynamic load factors (DLF), defined as the ratio of dynamic to static deflec-tions, were determined for the done and cylinder to draw an analogy to single degree of freedom systems. The DLF for the dome model was computed as 0.0887 and that of the cylinder section as 0.0191. These values were obtained from i loading I results and compare favorably to hand calculated values.

7.7 Comments The use of direct integration in our analyses does not exclude the use of rodal superposition. An efficient eigenvalue extraction routine would enable an

analyst to solve for all the frequencies associated with the structure and evaluate the participation of each mode.

l l

Care should be exercised when comparing McGuire results with those of other l

l containments. For example, a 1/2 inch plate thickness has an associated section I modulus of t2 /6 or 0.04167 in2 . A 3/4 inch plate has a corresponding value of I

0.09375. Thus, the 1/2 inch plate would have bending stresses which would be 2.25 times greater than the 3/4 inch plate. All analysis work neglected the effect of strain rate upon material properties. This is a conservative assurp-tion.

7. 8 Conclusions The following conclusions can be drawn from these analyses:

1. The structural responses due to hydrogen detonation are localized.

7-6

2. The calculated local stress intensities are substantially below the actual material yield strengths.

3. The resulting stress intensities are within the allowable limits (37.0 ksi < 48.0 ksi) for emergency conditions as specified by the 1968 Edition of the ASME Code including addenda thru the l

summer of 1970.

l l

l l

l l

l 7-7 l

1

)

7. 9 Combustien and Exclosives Research It c. Recort COMBUSTION AND EXPLO51VES RESEARCH. ,!NC.

. O L I \' E R D U II. u l N G .

P I T r S I) U R G li u u . l' E N N S i' L\* A N 3 ^

DTRNARD LfWi> T ie p ea. Mi3n33 DrLA RARLOVITZ Jcr.uary 26, 1% 1 c.63. A 4 4,... c0MDc:

1 Mr. David Coeser l P4B Systems Division  !

Westinchouse Flectric Con.

P. O. Dcx 355 Pittsburgh, Pa. 15230 RE: Jet M1.xing of Stoca - H ' Air 2

Dear Mr. Goeacr:

l In the following we consider the question whether detonabic mixturcs ccn ever be fomed in the McGuire containment.

1. Assuming that 1100 lbs. of 14 (MI-2-type cccident - 501, Zr.)

is evenly distributed in theMcGuire containment and all the sitecm is condenced, the concentration of hydroden in the hydrogen-nir mixture is 16.5%. T. tis is below the detonation limit of 18% hydrogen in cir.

2. With the components of the McGuire contnin:r.cnt in operation, j 1.c. ignitors, recirculation fans, ice condenser, centainment j sprays, hydrogen starts to burn shortly ofter it is relcnced

! and the nnxinws. hydrogen concentration cannot execed approximately 10%.

l 3a. Accordirc to the McGuire cecnE1o dated Ilov. 11, 1980, l the steem flow rate has a minimum of 22 lbs'./sec. nt tits.c =

l M50 sec., nt the sr=c time the H, flov is 0 75 lbs./sec.

l 2c air ficv rate from the two deck fans is 1000 CF/sec.

The ec= position of the unifem mixture is:

l H., 9 1%

i c Steam 29 6%

l Air 61 3%

Tne percentnge of stoca in the stccm-nir mixture is 33%.

We percentage of hydrogen in the I! 2-ste m jet in 23 5%.

l 7-8 1

L

• e Mr. David Coccer January 26, 1981 Pa6e 2. {

3b . The hydrogen flow rate has a re.xinu. cf 1.05 lbs./se ,

at ti=c = L750 acc. at the nn=c tir.e the stece i b re :

is le 1bs./sec. 'Ihe ccarposition of the unifor r_1.xtee is:

E g 9.?S f

Stes= 46.6% l Air 44.2[o {

Percentage of stern in the stca.> air =1.xture is 51"..

Per:cntage of % in the II2 -st as Jct is 16. p.

t

'i'he attached dincram shows the fin ==r.bility limits an:1 detenatics 11=1 s of H3 -stecm-nir r.Jxtures as a functics of the percenta6e of cr.cc= in the stcar,nr.d air =1xture.  !

Se hydregen percentacc in the Ib cterc jets (no air) is shavn en the riCht hand cide of the dincra= for the above tvo ecces. i 'Ihe L ccncentration in the final r.1xtures is shown nent the center in th'd lever part of the diacrc=. Trcnsition frc= the original I4-sten =: Jets to the final =1xtures by entrain =ent of air is linear and tilis is shavn by the straight lines connecting ther.c points in the diagrra. It is scen that while these lines cross into the f1r. nble region near the nose of the curve, they never enter the detonable range. I 1

I Very truly yours,

t COGUSTICH AND D~.?I4SIVE3 FISTARCH, II!C.

,i

.-Q .i lt i

l .. ) %. . - ,. . ' 'L . J. . . . .

Bernard Icvis

~

Bela Karlovits Encicsure I

i i

l 1

\

7-9 .

l

l

-o or a _

l .

l  :' .'

I i

I

\ i 70 '

i. .

. 93J t1 IIA w m as /c 60 L -- --

i

- t- --

%  : i a' '

l i

.: go . - - _ . . . . . .

l 5 ,ejle w % G/c Ls Nf

~

Tla.lnsua',Ic '

l

<ifn -

i s -

und Oef o uec /c - . i so _.____ _ _ _.J-... ,

-- .. . j t

u . . -

23 5 7 N' j 20- ~ yyy& \L L.@ U - l ~~ ~

/4 s sy;, -

to . - ....n._

E t

I e

~- e

/o0 #b a to 20 00 90 30 40 20 80

,CO

[/gd % 8 t) d f C & hg

• A s' y* J/! / g / so g* C, 1

e Is' . . . ~

w Fl etyn wvi et ls Ie l i.%1 ifi et H t, d.crin, w Itc11 m -4tr & W ate..

l l

7-10 l

t l [

e ..__-y v. , _ . , - -.~.-, 4.m , . - - - - -- . ~ , - - -

l Ma x 00 k/in z Max 40 k/in 2 Stre It t i kMd ibx. Inte9 MODEL LOAD-TIlli tilSTORY tbx Displ. Step Size 0 uter Middle Outer Middle NIddI" F,ibers Surface fibers Surface Outer Surface (in.) (microsec) fiber 31 .0 28.4 Z Displ.

DOME I 8.95 9.26 37.0 9.1 0.438 25 i

-13.1 -9.9 29.3 27.1 Z Displ-DOME II 8.63 8.87 35.0 8.75 0.420 25 l -12.0 -9.4 31.6 '

Z DOME III

, 8.94 9.19 37.7 9.1 0 436 10

-13.8 -10.1 9.97 10.0 R Disp 1-CYLINDER 1 4.20 1.95 16.2 3.6 0.326 100

-1.57 -6.1 1

i l TABLE 1 STRESSES AND DISPLACEMENTS i

Table 2: Refe rences

1. Introduction to Structural Dynamics, by J. M. Biggs, McGraw-Hill Book Company, 1964.

2. Concepts and Acolications of Finite Element Analysis, by Robert D. Cook, John Wiley & Sons, Inc., 1974.

3. " Dynamic Finite Element Analysis of Arbitrary Thin Shells", by E. L.

Wilson and R. W. Clough, Computers and Structures, Vol. I, pp. 33-56.

l l

l i

1 Cetenation Load i

/ Symetry Line

? - DISP. = 0.0 M"  : -

ROT. = 0.0  ;

ns,-

_w @  !

w - c, $

i ,g e  !'aterial procerties

-.i 52.* _s ,v

$ tS & *  ; g. ,b

1. = Cylincer Stifferer 164."  ;

=" $  %

ci' C+1 S: ORTH 0TROFIC ISOTMPIC l

,i i E = 1.0 psi E = 29 x 105 ps ot nl@ rr eu i C~ c8g.-

E:: = 3.25 x 10 pst v r: = 0.30 7

L. .

ci : @g. @ -

i- E . = 2.93 x 10 7psi j 7

.ca. G , = 1.12 x 10 psi i 5-

: P i v = 0.30 t u..

1. 8

@"E 1

e i

• q i$

1 i

G.

a 6

, , - e a o4 I

Axis of -

Symetry Q, I l

30. " ,3 I

r

.! e 4Z

/

/ r\ ,

/

R

c. 1 Q e /

/

u .r*

• . I

/-

1 c.

' w'l=-

1-4 d

CYLINDER PCCEL Ficare 7.1 l

-me,. . - - _%w,-., m,...- __

..%...g.,_,,mm-,..-.w-,,,.,,m.,, .,.w,.,-wr,.~.,,-,,,,_,...,,,m.,,,e.w.,,r,,.,,,,,-w._,, -~_,,,,,,--,..r~ , . . , ,

l l t

Not to Scale O' OS* n.o* t.5* g 0,2.s' ao.

lJJJ//.o-I */ ,, / .-

4 i

i. t ,  ;

D@!@!st@/@/g'/ <=...

, 1rl <

l @ @/ n O/

/ o-i I V'! lh / /g Detonation Loading b #*

' @j, \,7%

N f o.

/*kNrA-gg I

4

,, h+/

4 0

Material Properties

,h E = 29 x 10' psi

' v = 0.30 lb-sec:

0 = 7.339 x 10 u in' R

1 LOCAL DOME MODEL Figure 7.2 i

. . - . - , . . . . . - , - . . . _ . ~ , . . . , . . . . , . - _ . . _ _ _ , . , - - , . - - . , - . - - , . . . , - . - - - - - - , . . - . , , , - . . . . . , , . - . - . - - . . . , , , . , . . . . , , , . , - - , , - . . - . -

i I t i 200 -

LOAD TIME HISTORY I E

B d

E l l l l 0 100 200 300 400 500 TIME (usecs)

I l 200 C

m S LOAD TIME HISTORY II

=

d c.

3 _

~

l I I I I 0 100 200 300 400 500 TIME (usecs) i 1

LOAD TIME HISTORIES I AND II Figure 7.3 e- -y - = e v-, v-- *- w wr ye- yw--w, =,e- =r w,-*-*=,--,-w *y-,v----,,ev,wy3,---+ +--+vv ee- ,-,-y, e,%-,--=<-r, yeg-- --*-e,- ---w ---e

For Loaded Nodes Wave Velocity = 6000 ft/sec. Peak Pressure = 200 PSIG Pulse Start to Peak = 50 usec.

Node Number Pulse Duration = 500 usec.

Time (psec) 0 500 100,000

'/.  ;

_10 510

'/.  ;

g 20 520 x

A E

O 50 550 1

= .

_ 90 590

'; I 130 630 180 680 LOAD-TIME HISTORY 111 l Figure 7.4

lllflllj'1 0

gr 3

s e

l f u o P

n o

n o i e r i ts t

a al t u c,

i r nP 0.

o e sn D

t e eo ri pt

%2 -

12 ea

. 1 R n o

2'g G l

rt i

d e ee r D

m iD u r ) A 0 s u 1 s O s o1

- e L 1 A F { 0 e R 5 -

1 r g E 4 e D g

D e

I t

j

( i L

e Y t C

- a

_- n F m i 05 d

r o

i r

0 7

I l \ i l l 0-

• 6 C

o I e T r 4

+,~ l Au I n i

Ri

- - t' n

M

- e L r A e I f

m Tl l

- 0 u E R

- } 1 c r E i f t

C l- ' lu C

R I

C i 0 5 2 g-y0-3 l l! [l fl lr ljl ll

!! i

' .:i,. ' _r  !,i <l' - r ! ,

~-

- c c c c e e e e s s s s u p p u e

. 5 0 0 0 2 5 0 0 s 1 2

- p t

e A O O 0 - - _

S e

m

.i i

0 0

0 T

A C O O 3 0

0

,5 A

2

. y s p ,

t e

S 0 )

g .

0 6 e

- 0 - r6 2 0 i 1 T 7

. x f s

ore -

1. ng u 1

/ d n

o oi sF G -

y0

~

N 0 c i r

/

I e a D 5 S -

( p

_ A 1 m -

O E o L , M C

, I L

E / & T D ,

O 0 M - 0 R

E O'..- i0 1

D -

I N

L

,#O Y

C -

- 0 i

0

. 5

.g n - - - .

5 0 5 _

0 1 0 ~

2 1 u

^mtxoz~

w. nEdotdm i  % - a a< a x P4 U 5

!tii4Il! i;ii l.;:!i; i;i5ii 41 ' !1i!i , < :!,j 1 j2 '  : i)i! '11I: i, 11 :1A

zg mm. - - - -

O a

w w

w w o- --

e

-)

u-zo S$. ,.

~ 00 L

• 0 2

8 * . -

e

, = _

s

• e .

c , -

,o .

. 040'0

=

e c.

=

w .

T . w

.d . .

.=-.

=

a .' ,. ~

~

.2 . . -

gq q . - s tE 8

~

- 9=

o

• w cn

• - 0#0'0 v-

< u.

- a C.

r .

• m

.)

., - a y 2 -

% ~

- OLO*0 C ~

LOO *0 i .. 4 . i4 . 4 i4 .'4 . i b4 000'0 e .....b4

, e e e o o

w .

q M m m M R e o e o o o o o (53H]tJ1) IN3W33V7dSIG - Z

\

l l

l 1

i I

I l

/@

l i .. P@

Lead  !

t- I U 0 S 10

/

/

!  ?-

, / ,s

l I

I l^ 30 ,

ZJL 1 Displacement at 0.007 Sec, Max. Stress Time

/ P. 2 Displacement at 0.006 Sec,

?dax. Disp. Time LOCAL DOME MODEL DISPLACED SHAPES Figure 7.8

i 0.298" )

" =

Symmetry Line 7

l Axis of Symmetry I

Stiffener

<, 10 ft. between stiffeners 900 Angle for Displaced Shape i,

l 1800 -

p _. 00 1

Detonation Load 2700 l

Stiffener Line CYLINDER MODEL - DISPLACED SHAPE ANGLE = 0.00, TIME = 0.015 SEC., LOADING I i

Figure 7.9

0.017"

1. f'e #

Axis of Synnetry I

ne I 10 ft. between stiffeners l l 1

f l 90* Angle for J

Displaced Shape JS*

teo*

_T

!4 ca i

h j

Detonation

, l 23C* Load l

Stiffener Line CYLINDER MODEL - DISPLACED SHAPE ANGLE = 150, TIME = 0.016 SEC., LOADING I Figure 7.10

80 Non Flamable 70 -

w a:

b EO -

1. Flamable Limit

'O ~

f

.k

1. 5 7* 39 ,

Flamable and Detonable ci e 30 -

.5 g Detonable Limit s 23.5

--~~~ ~~ ~

5 20 ~.t. -s 16.5 8' C 5

-E' 10 - h W O

n 10 20 30 40 50 60 70 80 90 100 0

Steam in Steam and Air Mixture "

h McGuire Accident, time = 4450 sec, minimum steam flow rate, 23.5 percent H2 in H2 - steam jet.

h McGuire Accident, time = 4750 sec, maximum hydrogen flow rate, 16.5 percent H2 I" N - steam jet.

2 ,

TMI accident,1100 lbs. H ,2steam condensed.

Q)

FLAMMABLE AND DETONABLE LIMITS of

' HYDROGEN. STEAM, AND AIR MIXTURE Figure 7.11 l

l l

t t

8. EMERGENCY HYDROGEM MITIGATION SYSTEM OPERATION AND TESTING 8.1 Actuation Criteria 8.2 System Operation 8.3 Preoperational Testing 8.4 Periodic Testing 11b 8-1

8.1 Actuation Criteria Chapter 3 describes the Emergency Hydrogen Mitigation System (EHM) which has been installed in the Unit I containment of the McGuire Nuclear Station. This system will be actuated in accidenc situations which have the potential for the generation of excessive quantities of hydrogen. Excessive hydrogen generation results from uncovering the reactor core and attaining high fuel clad tempera-tures in the presence of steam. An accident which results in core uncovery necessarily involves the transfer of large quantities of mass and energy from the reactor coolant system to the containment. This transfer of mass and energy will cause an increase in containment pressure sufficient to initiate Phase B containment isolation. Therefore, an automatically initiated Phase B containment isolation provides an early indication of an accident situation which has the potential for excessive hydregen generation.

To effectively perform their intended function the EHM system igniters must be energized and at operating temperature before significant amounts of hydrogen ar e released to the containment atmosphere. Therefore, actuatiofi of the EHM system should occur once the potential for excessive hydregen generation is establishad and should not be dependent upon a measurement of tne hydrogen concentration inside containment. Automatic Phase B containment isolation establishes the potential for excessive hydrogI generation well before the release of any hydrogen to the containment atmosphere. Automat.c Phase B containment isolation also initiates the actuation of the other syst mis required to operate in conjunc-tion with the EHM system to assure containment integrity. These systems are the Containment Air Return and Hydrogen Skimmer system, and the Containment Spray system. Therefore, an autouatically initiated Phase B containment isolation will be used as the indicator for actuation of the EHM system.

8-2

l 8.2 System Operation l

The following guidelines will be used when incorporating the operation of the EHM system into the McGuire emergency procedures.

1. The occurrence of an automatically initiated Phase B contain-ment isolation signal will provide cause for actuation of the EHM system.

2. Prior to actuation of the EHM system the operator will verify that automatic initiation of the Containment Air Return and Hydrogen Skimmer system and the Containment Spray system has occurred.

3. Actuation of the EHM system will be accomplished by manually closing the circuit breakers for all of the igniter circuits at each Emergency Lighting Power system panelboard.

4 After actuation of the EHM system containment pressure and temperature should be monitored for indication of spikes representing hydrogen burning. The Containment Air Return and Hydrogen Skimmer system and the Containment Spray system should also be monitored to assure their continued operation.

5. The ERM system should remain in operation for 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> after adequate cooling of the reactor core has been well established.

The EHM system, the Containment Air Return and Hydrogen Skimmer system, and the Containment Spray system should be re-actuated upon any indication of a return to inadequate core cooling conditions.

~

8.3 Preoperational Testing j The ERM system igniters will be operated at 14 volts to assure the direct applicability of the Fenwal tests discussed in Chapter 5. Preoperational I testing will be performed to demonstrate functionability and to provide r

information for periodic testing and maintenance.

8-3 l

Preoperational testing will include a test where the igniters are energized 7

for a period of 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />. Following this 24 hour2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> period of continuous operation the voltage and temperature of each igniter will be measured and recorded. The total current on each igniter circuit will also be measured and recorded. The measured igniter voltage should be approximately 14 volts.

The allowable deviation frcm 14 volts will be specified in the preoperational i test procedure. The measured igniter temperature should be greater than

, 15C00 F. The measured total current on each igniter circuit will be used in subsequent periodic testing.

8.4 periodic Testing Periodic testing of t'he EHM system will be performed during each refueling outage. The total current on each igniter circuit will be measured and compared to the total current recorded for that circuit during the peropera-tional testing. The allowable deviation from the total current recorded for that circuit during the preoperational testing will be specified in the periodic test procedure. The temperature of each igniter will also be measured and verified to be greater than 15000F,

(

4 s

8-4

- . - . . - - . . . - . - _ . , . - _ , . . . - . - . - . . - . . . - - . . . . - . - - - - . . . . . - . - _ - . - . , , - . . . - .