ML20127K708
| ML20127K708 | |
| Person / Time | |
|---|---|
| Site: | River Bend  |
| Issue date: | 05/18/1985 |
| From: | FAILURE ANALYSIS ASSOCIATES, INC. |
| To: | |
| Shared Package | |
| ML20127K698 | List: |
| References | |
| FAAA-85-5-10, NUDOCS 8506270543 | |
| Download: ML20127K708 (56) | |
Text
.
FaAA-85-5-10 QRGS EVALUATION OF DSR-48 ENERGENCY DIESEL GENERATOR CRANKSHAFTS AT RIVER BEND STATION Prepared by Failure Analysis Associates Palo Alto, California l
l .
Prepared for TDI Diesel Generator Owners' Group May 18, 1985 8506270543 PDR 850516 S
ADOCK 05000458 PDR 1
f TABLE OF CONTENTS
1.0 INTRODUCTION
AND SW91ARY.......................................... 1-1 2.0 C04PLI ANCE OF CRANKSHAFT WITH DOMA RECC:4MENDATIONS. . . . . . . . . . . . . . . . 2-1 2.1 Revi ew of TDI Torsi onal C ri ti cal Speed . . . . . . . . . . . . . . . . . . . . . . . 2-1 2.2 Review of Stone & Webster Engineering rorporation Torsiograph Test............................................. 2-3 2.3 Conclusions.................................................. 2-6 3.0 FATIGUE ANALYSIS OF CRANKSHAFT.................................... 3-1 3.1 Crankshaft Dynamic Torsional Analysis........................ 3-1 3.2 Calculation of Fillet and Oil Hole Stresses.................. 3-5 3.3 Endu rance Limi t for C ra nk shaf t s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-7 3.4 Factor of Safety Against Fatigue Failure..................... 3-8 3.5 Conclusions.................................................. 3-8 4.0 CRANKSHAFT ENDURANCE TEST......................................... 4-1 4 .1 Conclusions.................................................. 4-3 5.0 CRANKSHAFT FILLET AND OIL HOLE INSPECTIONS........................ 5-1 5.1 C rankpi n Fi ll et Eddy Current Exami nati on. . . . . . . . . . . . . . . . . . . . . 5-1 5.2 Crankshaft Oil Hole Liquid Penetrant Examination............. 5-1 5.3 Crankshaft Oil Hole Eddy Current Examination................. 5-1 5.4 Conclusions.................................................. 5-2 i
r-
1.0 INTRODUCTION
AND SlD9tARY An evaluation of the 13-inch by 12-inch crankshaf ts installed in the emergency diesel generators at River Bend Station, RBS, was performed to determine the adequacy of the crankshafts for their intended service. This report summarizes the results of testing, analyses, and inspections performed on the crankshaf ts in the two Transamerica Delaval Inc., TDI, DSR-48 engines installed at RBS.
The TDI part number for the crankshafts at RBS is 03-310-05-AC. The plant ID for engine serial no. 74039 is 1EGS*EG1A (EGIA) and for engine serial no. 74040 is 1EGS*EGlB (EGlB). The forging and machining of the crankshafts was performed by Ellwood City Forge Corporation.
Significant testing and evaluation of 13-inch by 12-inch crankshafts installed in the TDI DSR-48 engines at Shoreham Nuclear Power Station, SNPS, has been performed. The part number for the crankshafts at SNPS is 03-310-05-AC. Fatigue damage was discovered in the original 13-inch by 11-inch crankshafts at SNPS and the replacement crankshafts installed are of the same design as those in use at RBS. The two crankshafts have identical crank-shaf t stiffness, inertia and fillet geometry. The main differences in the two torsional systems are that the engines have different flywheel and generator inertias and a different shaf t stiffness between the flywheel and generator.
Due to the similarities of the two torsional systems the response of the crankshafts are similar, and this similarity has been demonstrated by analysis and testing.
The testing on the crankshaft at SNPS consisted of dynamic strain gage measurements to determine maximum stresses, torsiograph testing and an endur-ance test run for 107 cycles at loads equal to or greater than 3300 kW.
Dynamic torsional analysis and stress analysis have been performed on.the SNPS crankshaft. In addition, the detailed investigation of the failure of the ,
original 13-inch by ll-inch crankshaft at SNPS provides data for evaluation of tie fatigue life of the crankshafts. To utilize the data available from the evaluation of the SNPS crankshaf ts, similarities and comparisons to the RBS crankshafts will be made throughout the report where appropriate.
1-1
r The crankshaft is required to meet the recommendations of the Diesel Engine Manuf acturers Association, DEMA. In Section 2.0, the RBS torsional design calculations prepared by TDI and the results of the torsiograph tests performed at RBS by Stone & Webster Engineering Corporation, SWEC, are reviewed for compliance with DEMA stress allcmables. The relationship between torsional response and engine load is determined based on the results of the torsiograph test. The crankshafts at RBS meet the DEMA allowables for single order and combined order stresses at 450 rpm and 3500 kW.
The fatigue analysis to determine the safety margin of the crankshaf ts at RBS at 3130 kW is discussed in Section 3.0. Amendment 16 to the FSAR states 3130 kW as bounding the maximum emergency service loads. A dynamic torsional analysis of the crankshaft is performed to determine the stress levels in each cylinder of the crankshaft. This model is compared with the SWEC test data for the amplitudes of free-end vibration. Stress values at full load are determined over a range of 5% of rated speed.
The f atigue endurance limit is established for the crankshafts at both RBS and SNPS by first obtaining the endurance limit for the failed crankshaft at SNPS from strain gage test data and a review of literature. The dif-ferences between the failed crankshaft and those in use at RBS and SNPS are then assessed and the endurance limit is then scaled to account for these differences.
Utilizing the SNPS dynamic strain gage test data, the factor of safety against f atigue f ailure is calculated to be 1.39. This value for the factor of safety is established with more confidence than usual on account of the dynamic strain gage testing of a similar crankshaft at SNPS, and on account of the endurance limit established from the failed crankshafts at SNPS. The variation of the factor of safety over the 5% rated speed range ,is then reviewed. The reduction in the factor of safety at overspeed c:mpared with rated speed is found to be small in comparison to the safety margin.
In Section 4.0, the RBS load that produces stress levels equal to those in the SNPS crankshaft during the 10' cycle endurance test is determined. The applicability of the SNPS endurance test results to the crankshafts at RBS is discussed.
1-2
r- T The safety margin of 1.39 for the RBS crankshafts at 3130 kW demon-strates that the crankshafts are adequate for their intended service at loads
-not exceeding 3130 kW.-
4 i
4 i~
1-3
r-2.0 CmPLIANCE OF CRANKSHAFT WITH DB A RECOiMENDATIONS The purchase specifications for the diesel generator sets required that the recommendations of the Diesel Engine Manufacturers Association, DEMA
[2-1], be followed. These recommendations state:
In. the case of constant speed units, such as generator sets, the objective is to insure that no harmful torsional vibratory stresses occur within five percent above and below rated speed.
For crankshafts, connecting shaf ts, flange or coupling components, etc., made of conventional materials, torsional vibratory conditions shall gener-ally be considered safe when they induce a superim-posed stress of less than 5000 psi, created by a single order of vibration, or a superimposed stress of less than 7000 psi, created by the summation of the major orders of vibration which might come into phase periodically.
In March, 1975, Transamerica Delaval Inc. (TDI) performed a torsional critical speed analysis of the crankshaf ts at RBS [2-2]. References to TDI analysis in the body of this report all reference this effort. In Section 2.1, this analysis will be reviewed for compliance with the above allowable stresses. Where appropriate, data for SNPS will be included in the tables for comparison [2-3]. In September 1984 and again in May 1985, Stone & Webster Engineering Corperation, SWEC, conducted torsiograph tests on the crankshaft in EG1A at RBS [2-4, 2-5]. In Section 2.2, the test results will be compared with the above allowable stresses.
l l
2.1 Review of TDI Torsional Critical Speed Analysis l Diesel generator torques due to dynamic response are usually calculated in two steps. First, the torsional mode shapes and natural freque,ncies of vibration are calculated. Second, the dynamic forced vibration response due j to gas pressure and reciprocating inertia loading is calculated. TDI calcu-lated the response at 100% of rated load of 3500 kW.
i i
j 2-1 y .
2.1.1 Natural Frequencies The first step in a torsional critical speed analysis is to determine the natural frequencies of the crankshaft. The engine speed at which a given order resonates may then be calculated. The diesel generator is modeled as a system of lumped mass moments of inertia interconnected by torsional springs, as shown in Figure 2-1. The inertia and stiffness values are shown in Table 2.1.
It has long been standard practice in the diesel engine industry to solve this eigenvalue problem by the Holzer method [2-6]. This method has been used for at least 40 years [2-7], and thus is well established.
TDI used the Holzer method to calculate the system's first three natu-ral frequencies, which are shown in Table 2.2. The first natural frequency was found to be 38.0 Hz, which produces 4th order resonance at 570 rpm.
2.1.2 Nominal Stresses The second step in a torsional critical speed analysis is to determine the dynamic torsional response of the crankshaft due to gas pressure and re-ciprocating inertia loading. The 1st order is a harmonic which repeats once per revolution of the crankshaft. For a four-stroke engine, harmonics of or-de r 0.5, 1.0, 1.5, 2.0, 2.5. . . exist. TDI performs this calculation for each order of vibration up to 12.0 separately. For each order, the applied torque at a cylinder due to gas pressure and reciprocating inertia is calculated.
The values of this torque for each order are usually normalized by dividing by the piston area and throw radius. The normalized value for the nth order is referred to as T n. The values of T n for significant orders used by TDI are shown in Table 2.3. These values may be compared to those recommended by l Lloyd's Register of Shipping, LRS [2-8]. It is found that TDI's values are higher than LRS's values for low orders and lower for high orders. However, the stress from the measured largest single order was found to be within 5% of that computed by TDI. The response is then calculated by one procedure if the harmonic is at resonance and by another if the harmonic is away from reso-nance.
2-2 O
L
r At resonance, the torsional vibration amplitudes would increase indefi-nitely . in the absence of damping. The solution is obtained by balancing the energy input with the energy loss due to damping. TDI used an empirical form of hysteresis damping due to friction. The purpose of this calculation is to ensure that the diesel generator could be brought up to operating speed with-out undergoing excessive stresses as critical speeds are passed. Observations have shown that excessive vibration during startup does not occur [2-4, 2-5].
Away from resonance, the torsional vibrations reach a steady-state level even without the aid of any damping. The magnitude of this respor.se for each structural mode and loading order is calculated as the product of a dynamic amplification factor and an equivalent static equilibrium amplitude.
The equivalent static equilibrium amplitude is computed using a modal load and modal stif fness [2-9] for the nth order harmonic and given mode shape. The nominal shear stress, t, in the 12-inch pin of Crankpin No. 8 for each order is then calculated from the dynamic torque, T, using i = Tr/J, where r is the pin radius and J is the polar moment of inertia.
TDI calculated the response for the first three modes and plotted the results for only the first mode since higher modes produce much smaller stresses. The nominal shear stresses for the significant orders are shown in Table 2.4. It is seen that the largest single order stress of 3321 psi at rated load and speed for the 4th order is well below the 5000 psi DEMA allow-able.
TDI does not calculate the associated phase angle with the response of each order, so that it is not possible to calculate the combined response.
The measured combined response will be compared with the allowable in the next section.
2.2 Review of Stone & Webster Engineering Corporation Torsiograph Te'st The purpose of the torsiograph test of the emergency diesel generator is to measure the angular displacements of the forward end of the crank-shaft. These displacements are then used in conjunction with a dynamic tor-sional analysis of the crankshaft to confirm the results of the vibrational 2-3
analysis and to assess the maximum stresses in the crankshaf t. The torsio-graph response may be decomposed into components corresponding to each order and the peak-to-peak response may also be obtained. The torsiograph test is usually performed in two stages. The first stage is performed without load at variable speed and is used to determine the location of critical speeds.
Critical speeds may also be determined while operating at a fixed speed and observing the frequency content of the response. The second stage is per-formed at rated speed of 450 rpm with variable load, and is used to confirm the forced vibration calculations. These tests collect data during steady state operation. Data may also te obtained during transient (start-up, coast-down, misfiring) operations to assess the response of the crankshaft during these events.
Two torsiograph tests were performed on EG1A at RBS by SWEC. The first test, which was performed in September 1984, concentrated on collecting data during steady-state operation [2-4]. The second test, performed in May 1985, collected data during transient operation and repeated the variable load portion of the prior test to determine the response after the engine had been run in [2-5]. The operating history estimated from test engineers logs for EG1A and EG1B at the time of these tests are summarized in Tables 2.5 and 2.6. The results of these tests are discussed and compared with the results of the torsiograph test performed at SNPS [2-10] in the following sections.
The torsiograph response of the crankshaft at full load for each test is shown in Table 2.7. This comparison indicates that EG1A is better balanced now than at the time of the first torsiograph test. This is indicated by the lower half order response recorded during the second torsiograph test. The total response of the engine also shows a decrease. This decrease in response is typical of an engine af ter run-in. At the time of the second test EG1A had 120 hours0.00139 days <br />0.0333 hours <br />1.984127e-4 weeks <br />4.566e-5 months <br /> of operation at 3500 kW or greater, 87 more than at the time of the first torsiograph test. The response of the engine during the secon'd test is indicative of the future response of the engine. Torsiograph data presented in this section will be taken from the second torsiograph test.
2-4
2.2.1 Natural Freauencies The frequency content of the torsional vibration signal indicated the torsional resonant f requency of the system was 38.1 Hz [2-4]. This value is in excellent agreement with TDI's computed value of 38,0 Hz.
2.2.2 Nominal Stresses at Full load The torsiograph response of the cranksnaft at full load for significant orders along with the peak to peak amplitude is shown in Table 2.8. The amplitude of nominal shear stress may be estimated from the free-end vibration by assuming that the shaft is vibrating in its first mode. This assumption is typically used to estimate maximum stresses from torsiograph vibration data.
However, a more accurate calculation is made in Section 3 in which the crank-shaft stresses between each cylinder are determined and used in calculating the fatigue margin of the crankshaft. Assuming first mode response, the nominal shear stress in Crankpin No. 8 is 9351 psi per degree of free-ena rotation. The amplitudes of nominal shear stress are presented in Table 2.8. The results indicate that the largest single order has a stress of 3348 psi which is well below the DEMA allowable of 5000 psi. The total stress of 6546 psi is also shown to be below the DEMA allowable of 7000 psi.
The measured response at 3500 kW is in general agreement with that calculated by TDI and shown in Table 2.4, although the measured values are somewhat higher than the calculated values.
A comparison of RBS and SNPS results indicates that the response is similar for the crankshaf ts as expected. The fourth order is the largest order for both crankshafts, and is slightly higher for RBS because its fourth order resonant speed is closer to the operating speed. The five and a half order is smaller at RBS because its resonant speed is further away from the operating speed than it is at SNPS. Even though the fif th order at RBS has a resonant speed of 455 rpm, its magnitude is still quite small and there are six other orders with greater magnitude. Thus, the fif th order response is not large enough to be of concern.
2-5
2.2.3 Nominal Stress at Reduced loads Torsiograph test data is typically obtained at about four or five oifferent load levels. The relationship between torsional response and load may be obtained by plotting the torsiograph test data versus engine load.
The torsiograph test data for the variable load tests at RBS and SNPS are plotted in Figures 2-2 and 2-3. From these plots it was determined that a
, linear relationship between torsional response and load existed. This allows the determination of stresses at reduced loads based upon stresses at a 3500 kW load and the slope of the torsional response versus load curve. From the RBS data it was determined that there is a 7% change in torsional response for a 10% change in load. This relationship also holds for the SNPS torsiograph test data. To determine the torsional stress at a load L (kW), it, from the torsional stress at 3500 kW, t,,99, the following equation can be used L
=x )
1L 3500 (0.3 + 0.7 3500 s
Using this approach the stress levels at the 3130 kW load tnat produces stresses similar to those in the SNPS crankshaft during its endurance test were calculated. The values are tabulated in Table 2.9.
2.3 Conclusions The following conclusions can be stated:
e The TDI Holzer calculations of the natural frequency of the crankshaft are in good agreement with test data.
e The TDI forced vibration calculations at 3500 kW are reason-ably accurate and show that the single order stresses comply with DEMA.
e The torsiograph test data at 3500 kW shows that the combined order stresses comply with DEMA at the rated speed of 450 rpm.
i e The torsiograph test data shows that there is a linear varia-tion of crankshaft response with load.
u
! 2-6
Section 2 References 2-1 Standard Practices for Low and Medium Speed Stationary Diesel and Gas Engines. Diesel Engine Manufacturers Association, 6th ed., 1972.
2-2 Yang, Roland, " Torsional and Lateral Critical Speed Analysis: Engine Numbers 74039/42 Delaval-Enterprise Engine Model DSR-46 3500 KW/4889 BHP at 450 RPM for Stone & Webster Engineering Corp., Gulf States Utilities." Transamerica Delaval Inc., Engine and Compressor Division, Oakland, California, March 5,1975, Revision - May 4,1977 2-3 Yang, Roland, " Proposed Torsional and lateral Critical Speed Analysis:
Engine Numbers 74010/12 Delaval-Enterprise Engine Model DSR-48 3500 kW/4889 BHP at 450 RPM." Transamerica Delaval Inc., Engine and Com-pressor Division, Oakland, California, August 22, 1983.
2-4 Bercel, E., "Crankshaf t Torsional Vibration Measurements DG1A River Bend
! Nuclear Power Station," Stone & Webster Engineering Corporation, September 1984.
2-5 May, 1985, SWEC Torsingraph Test Report for River Bend Station.
2-6 Thomson, William T., Theory of Vibration with Applications. Second edi-tion, Prentice-Hall, 1981.
2-7 Hartog, Den, Mechanical Vibrations. Third edition, McGraw-Hill, 1947, 2-8 Lloyd's Register of Shipping, Guidance Notes on Torsional Vibration Characte ristics of Main and Auxiliary 011 Engines.
2-9 Craig, Roy R. , Jr. , Structura; Dynamics: An Introduction to Computer Methods. Wi l ey , 1981.
2-10 Bercel , E. , and Hall, J.R. , " Field Test of Emergency Diesel Generator 103," Stone & Webster Engineering Corporation, April 1984.
l
[
r 2-7 l L
TABLE 2.1 STIFFNESS AND INERTIAS FOR TOI HOLZER ANALYSIS Inertia Inertia Stiffness location (lb. ft. sec2) (ft. Ib./ rad)
Front Gear 6.8 58.1 x 10 6 Cylinder No. 1 49.2 84.7 x 10 6 Cylinder No. 2 47.9 84.7 x 10 6 Cylinder No. 3 47.9 84.7 x 10 6 Cylinder No. 4 47.9 84.7 x 10 6
, Cylinder No. 5 47.9 84.7 x 10 6 Cylinder No. 6 47.9 84.7 x 10 6 Cylinder No. 7 47.9 84.7 x 106 Cylinder No. 8 50.1 76.9 x 10 6 Flywheel 426.5 (1100.1)* 309.7 x 106 (276.8 x 106)
Generator 4976.1 (2650.4)
Where RBS and SNPS structural parameters differ, SNPS values are indicated in parentheses [2-3].
e 2-8
+ a- _-,,.--,v.,
- , ,m..- -
w.,-,e-_, , . - = - . ,,-,..,v.- , . --....,n- _ _ , - - -r_,--.y-,, -, , , , n .-y-,-,,----
(-w TABLE 2.2
- TORSIONAL NATURAL FREQUENCIES FROM TDI ANALYSIS AT.3500 kW-Mode Natural Frequency (Hz)
RBS SNPS
[2-2] [2-3]
1 38.0 38.7 2 -107.0 92.9 3 146.5 116.7 2-9
. f. ,
- r., - ;
z
, .,/ .
r( .
,, TABLE 2.3 ;
/ ' TORSIONAL LOADINGS FOR TDI ANALYSIS AT 3500 kW
~
, ,, . Order TorsionalLoading,T,(psi) 3
[2-2]
1.5 129.5 2.5 71.7
't t
3.5 . it "
42.8
, 4.0 27.7 4
4.5 23.8
~
j 5.0 17.4
, 5.5 12.8 r
i i
s f
1 I
y l
l f
I
]
i 2-10
c-TABLE 2.4 SINGLE-ORDER NOMINAL SHEAR STRESSES FROM TDI ANALYSIS AT 3500 kW Order Amplitude of Nominal Shear Stress (psi)
RBS SNPS 1.5 1599 1606 2.5 1069 1064 3.5 466 452 4.0 3321 2980 4.5 644 565 4
5.0 947 348 5.5 797 1080 DEMA Allowable for Single Order 5000 5000 i
2-11
t i
t TABLE 2.5 OPERATING HISTORY FOR EG1A
//15/84 8/29/64 b/4/S5 Preservice First Second Load Factory Crankshaft Torsiograph Torsiograph Inspection Test Test
>3500 4 Same as 4 4 3500 26 Factory 27.4 116 3130 < load < 3500 -- --
67 2625 < load c 3130 8 15.6 91 1750 < load < 2625 8 12.1 21
<1750 4 13.8 28 TOTAL 50 50 72.9 327 2-12
.m.-
TABLE 2.6 CPERATING HISTORY FOR EG1B ,
//15/84 8/29/84 5/4/85 Preservice First Second Load Factory Crankshaft Torsiograph Torsiograph Inspection Test Test
>3500 3 Same as N/A 3 3500 32 Factory 93 3130 < load < 3500 1 64 2625 < load < 3130 --
61 1750 < load < 2625 4 24 <
<1750 1 30 TOTAL 41 41 275
.~
2-13~
TABLE 2.7 COMPARIS0N OF SWEC TORS 10 GRAPH TEST RESULTS ON EG1A AT 3500 kW LOA 0 Amplitude of Free End Rotation (degrees) from SWEC Torstograph Data Order 1984 1985
[2-4] [2-5) 0.5 0.128 0.093 1.0 0.002 0.004 1.5 0.174 0.177 2.0 0.001 0.001 2.5 0.132 0.132 3.0 0.001 0.001 3.5 0.055 0.056 4.0 0.365 0.358 4.5 0.071 0.069 5.0 0.080 0.068 5.5 0.086 0.086 6.0 0.007 0.007 Combined response 1/2 peak to peak 0.757 0.700 2-14
r l TABLE 2.8 NOMINAL SHEAR STRESSES CALCULATED FROM SWEC TORSIOGRAPH TEST AT 3500 kW LOAD Amplitude of Free-end Amplitude of Nominal Rotation (degrees) Shear Stress (psi)*
[2-5] [2-10]
0.5 0.093 0.056 870 535 1.0 0.004 0.005 37 48 1.5 0.177 0.171 1655 1635 2.0 0.001 0.001 9 10 2.5 0.132 0.130 1234 1243 3.0 0.001 0.001 9 10 3.5 0.056 0.058 524 555 4.0 0.358 0.325 3348 3108 4.5 0.069 0.064 645 612 5.0 0.068 0.034 636 325 l 5.5 0.086 0.127 804 1214
! 6.0 0.007 0.008 65 76 l
l DEMA Allowabler for a Single Order 5000 5000 l
Combined Response j 1/2 peak to peak 0.700 0.693 6546 6626 l
DEMA Allowable for combined response 7000 7000 1/2 Peak to peak Amplitude of nominal shear stress is calculated to be 9351 psi per degree of free-end rotational amplitude for RBS [2-2] and 9562 psi per degree of free-end rotational amplitude for SNPS [2-3].
2-15
TABLE 2.9 RIVER BEND CRANKSHAFT STRESSES AT 3130 kW Amplitude or tiominal Order Shear Stress (psi) 0.5 806 1.0 34 1.5 1533 2.0 8 2.5 1143 3.0 8 3.5 485 4.0 3100 4.5 597 5.0 589 5.5 745 6.0 60 Combined response 1/2 Peak to Peak 6062 2-16
b o
~
w m n y e o k m E
N d d d d d d d 6
- o c c c c c c c c 3 c o
. I I I I I I I I e E E E E 0 S E E E 0 0 0 0 0 0 0 0 -e
, ['c f [ [ [' [' [ [ U i;
D ,
r ,
L L Ji i d, c
U L L h[j L-d 1 Rotational inertia .
Torsional spring I l.
Figure 2-1. TDI dynamic model of crankshaft.
1 FaAA-85-5-10
0.9 i i i i i i i
,- 0.8 -
O Total response -
en 4 4th order response O-
- u. o 0.6 -
,: [ -
$h 0.5 -
fQ 3e 0.4 -
_ 7% change in amplitude f or a 10% change in load
<g 0.3 - u
^'
0.2 -
c -
m w
c 0.1 - -
u.
0 '
O 500 1000 1500 2000 2500 3000 3500 4000 LOAD (kw)
Figure 2-3. Shoreham torsiograph test data [2-10].
FaAA-85-5-10
3.0 FATIGUE ANALYSIS OF CRANKSHAFT in Section 2.0 it was found that the RBS crankshafts satisfy the DEMA nominal stress recommendations at rated speed for 3500 kW. The stresses for a single order were considerably below the 5000 psi that is recommended as an allowable. However, the stresses for combined orders were close to the 7000 psi that is recommended as an allowable. While the DEMA limits are believed to contain an intrinsic (though unspecified) safety margin, a fatigue analysis of the crankshaft was undertaken to determine the true margin.
First, a dynamic torsional analysis of the crankshaft is performed to determine the true range of torque at each crank throw. This model is com-pared with SWEC torsiograph test data for the amplitudes of free end vibra-tion. The model may then be used to predict the nominal torsional stress over the speed range of 450 rpm 5%.
Second, the maximum actual stresses in the crankpin fillet are deter-mined from the dynamic strain gage test data from SNPS and the differences between the stresses at SNPS and RBS. The actual stresses in the oil holes are computed from the nominal torsional stress and the stress concentration factor.
Third, the fatigue endurance limit is established for the RBS 13-inch by 12-inch crankshafts by first obtaining the endurance limit for the f ailed 13-inch by 11-inch crankshafts at SNPS, and then assessing the differences between the two crankshafts. The endurance limit is compared with values provided in the literature.
i l Finally, a factor of safety against fatigue failure is computed.
3.1 Crankshaft Dynamic Torsional Analysis -
l 3.1.1 Torsional Model l
FaAA developed a dynamic torsional model of the crankshaft to overcome
! limitations in TDI's conventional forced vibration calculations. For I
instance, the TDI method does not compute the phase relationship between the 3-1
various orders or modes, so it is not possible to compute the true summa-tion. The actual maximum stress is a direct result of this summation. Fur-thermore, the TDI method always predicts maximum stress in Crankpin No. 8, which is generally true for a single order in the first mode but not true for the combined response of all orders and modes.
The dynanic model developed used the same idealized lumped inertia and torsional spring model as the TDI analysis (Figure 2-1 and Table 2.1) with one additional spring placed between the generator and ground to represent the effect of the grid on dynamic response during synchronous operation. This spring constant was found to be 2.053 x 106 f t.-lb./ radian based on generator specifications. This constant is set close to zero to represent RBS emergency bus operation.
The first three torsional natural frequencies for the RBS crankshaf t are shown in Table 3.1. The natural frequencies are in agreement with those computed by TDI and that measured by SWEC. There is also a rigid body mode which has a natural frequency of 3.0 Hz when the engine is connected to the grid, and is 0 Hz for RBS emergency bus operation.
When the diesel generator is running at a given speed and power level, the forced vibration problem is steady-state where both load and response repeat themselves every two revolutions of the crankshaft. To model the dynamic response, a modal superposition analysis [3-1] was used with harmonic load input. The calculation of the harmonic loads will be discussed in the next section.
3.1.2 Harmonic Loading To calculate the harmonic loading on a crankshaf t it is necessary to consider gas pressure, reciprocating inertia, and frictional loads.. The gas pressure loading was obtained from pressure versus crank angle data. This pressure was measured at SNPS by SWEC [3-2]. The pressure was measured in Cylinder No. 7 by inserting a probe through the air start valve. A top dead center, TDC, mark for Cylinder No. 7 was simultaneously recorded by a probe on the flywheel. The pressure data at 3500 kW was reduced by FaAA to obtain the pressure curve snown in Figure 3-1.
3-2 e
The torque produced by this pressure may then be calculated as a func-tion of crank angle. The mean value of this torque should be the torque required to produce 3500 kW divided by the mechanical efficiency. A mechani-cal efficiency of 1.0 was obtained, rather than the expected 0.88 to 0.90.
The dif ference is probably explained by either the pressure measurements being too low or by the TDC being shif ted. Peak pressures were measured in all the cylinders to ensure that all cylinders were balanced. Normally, the excess torque above that required to run the engine at 3500 kW is dissipated by friction.
The reciprocating mass of the connecting rod and piston was found to be approximately 820 lbs. This mass causes reciprocating inertia torque on the crankshaft. The effect of this torque was combined with the gas pressure torque.
The total torque was then decomposed into its sine and cosine harmonics corresponding to each order. These torque harmonics were used in the steady-state analysis. The magnitude of the torque harmonics are normalized by dividing by the piston area and throw radius. The resulting normalized torques for the most significant orders are shown in Table 3.2.
3.1.3 Comparison of Calculated Response With Test Data The response aue to the first 24 orders and all 11 modes is calculated using modal superposition with 2.5% of critical damping for each mode. The actual value of damping used has little effect on the response since the orders are not at resonance at 450 rpm.
The calculated amplitude of free-end displacement is compared to the SWEC test measurements at RBS [3-3] in Table 3.3. It is seen that the agree-ment is close for all significant orders. The vector summation listed repre-sents half the maximum -peak-to-peak displacement range. This confirms that the pressure versus crank angle data used provides a good representation of the pressure loading at RBS.
3-3
The model also calculates the range of torque at each crank throw, from which nominal shear stresses (t = Tr/J) are calculated and shown in Table 3.4 The maximum stress occurs between cylinder number 5 and cylinder number 6.
3.1.4 Calculation of Nominal Stresses at Offsoeed Conditions The dynamic torsional model of the crankshaf t may be used to predict the stresses over the speed range 450 rpm 15%. Over this speed range the maximum stress always occurs between cylinder number 5 and cylinder number 6. The maximum amplitudes of torsional stress are shown in Figure 3-2 and Table 3.5 for both RBS and SNPS at 3500 kW. It is seen that there is good agreement between the values computed by FaAA and those computed using Dr.
Sarsten's computer program COMHOL2 [3-4]. At 450 rpm the stress at RBS is somewhat higher, 7357 psi, than that at SNPS, 7006 psi. At overspeed, 472.5 rpm, the stress at RBS is higher than SNPS since the fourth order critical speed is closer to the operating speed at RBS than it is at SNPS. On the other hand, at underspeed, 427.5 rpm, the stress at RBS is lower than SNPS because the five and a half order critical speed is further away at RBS than it is at SNPS.
3.1.5 Calculation of Nominal Stresses at Reduced Loads In Section 2.0 it was shown that the torsional vibration of the crank-shaft is linearly related to the engine load. Torsiograph data for both RBS and SNPS snowed that this relationship leads to a 7% change in vibration amplitude for a 10% change in engine load. In addition, the dynamic strain gage test at SNPS [3-2] showed that the relationship between crankpin fillet stresses and load is also linear. Figure 3-3 indicates that the crankpin measured stress data shows an 81/2% change in stress for a 10% change in engine load.
This linear relationship may now be used to compute stresses at loads below 3500 kW. The stresses in Figure 3-4 and Table 3.6 are computed for a load of 3130 kW at RBS and 3300 kW at SNPS. The load of 3300 kW at SNPS corresponds to the load used for the 107 cycle entrance test on DG103 at 3-4
I SNPS.
The load of 3130 kW produces stresses at RBS that are comparable to stresses at 3300 kW at SNPS. These reduced stresses are calculated assuming a 7% stress reduction for a 10% load reduction. Using 7% leads to a comparison which gives relatively higher stresses for RBS than if 81/2% were used.
The stresses at 3130 kW computed in the above manner will be used to compute the fatigue margin of the crankshaft.
3.2 Calculation of Fillet and Oil Hole Stresses The fillet of crankpin number 5 of a 13-inch by 12-inch crankshaft at SNPS was instrumented with strain gages and tested under operational condi-tions at 450 rpm and 3500 kW [3-2]. The fillet geometry of the SNPS crank-shafts is the same as that of the RBS crankshafts as specified in TDI drawing 03-310-05-AC. Figure 3-5 shows typical structural dimensions for the crank-shafts.
A dynamic model of the crankshaft confirms that crankpin number 5 undergoes the greatest stress. Three-dimensional finite element models of a quarter crank throw were used to determine the location of highest stress, both within the fillet and around the crankpin [3-5]. The strain gage rosette was placed in this highest stressed location.
The strain gage results from the SNPS test at 3500 kW are shown in Table 3.7. The calculation of principal stresses from the strain gage results are also shown.
To account for the simultaneous effects of shear and bending, the stress _ state is represented by an equivalent stress using Sine's method
[3-6]. For a biaxial stress state, the equivalent alternating stress, S ga' and equivalent mean stress, Sqm, are given by: ,
S *
(ba -b b +
qa 1
a ia 2 a2 N2 and S *S +S qm mg m2 3-5
l l
where S and S are the alternating components of principal stress, and a3 a2 S
m3 and S m are the mean components of principal stress. From the SNPS test 3
at 3500 kW [3-2], the equivalent alternating stress, S a,q and equivalent mean stress, Sqm, on crankpin number 5 were calculated to be:
S qa
= 24.5 ksi S
qm = 4.8 ksi Equivalent stresses, S qa and Sq ,, are those alternating and mean uniaxial stresses that can be expected to give the same life as the given multiaxial stresses.
To determine the actual fillet stresses at 3130 kW in the RBS crank-shafts the value of S qa needs to be factored by the differences in nominal torsional stress. Thus, the equivalent alternating stress at RBS at 3130 kW is given by S
qa
= 24.6 = 23.9 ksi (7006) since the nominal stress at this load is 6813 psi and the nominal torsional stress at SNPS at 3500 kW is 7006 psi.
Tfie stress in the oil holes may be calculated from the nominal tor-sional stress using a stress concentration factor of 4.16. This factor is based on theoretical solutions for a plate with a hole in shear [3-7], and a
- finite element model to account for the blend radius.
l The nominal torsional stress in the main journals is computed from its adjacent crankpin by accounting for the difference in the radius of the jour-l nal. The maximum oil hole stress is determined to be in main journal number 6 and is 22.3 ksi.
j 3.3 Endurance Limit for Crankshafts j The endurance limit for the RBS crankshafts was based on an analysis of the failed 13-inch by 11-inch crankshafts at SNPS and confirmed by a review of l the literature.
l 3-6
The number 5 crankpin fillet of a 13-inch by 11-inch crankshaft at SNPS, which had a fatigue crack, was instrumented [3-8]. The strain gage data and resulting principal stresses for this crankshaft at 3500 kW are shown in Table 3.8.
The equivalent alternating stress, Sga, and equivalent mean stress, S
qn, were calculated to be:
S qa = 33.7 ksi S
qm = 10.9 ksi At the time of the fatigue failure in DG102 the crankshaft had experienced at least 273 hours0.00316 days <br />0.0758 hours <br />4.513889e-4 weeks <br />1.038765e-4 months <br /> at equal to or greater than 3500 kW. Counting one stress cycle for every two revolutions of the crankshaft, this leads to about 4x los cycles. The operating history for the three SNPS crankshaf ts are shown in Table 3.9.
A review of S-N data for material similar to the crankshaf t indicates that a typical S-N curve is represented by the notched wrought 1040 data shown in Figure 3-6 [3-103. This data was 'used to determine the shape of the S-N curve. The curve was then shifted on the vertical scale to match the data from the full scale crankshaft failures at SNPS. Using the shape of the curve in Figure 3-6, the 107 cycle endurance limit for this mean stress was deter-mined to be 32.4 ksi. The ultimate tensile strength for these crankshafts averaged 96 ksi.
The crankshafts at RBS have a minimum ultimate tensile strength of 94 ksi. The endurance limit scales approximately linearly with ultimate tensile strength. On this basis, the endurance limit for the RBS 13-inch by 12-inch crankshaf ts is calculated from the SNPS 13-inch by 11-inch crankshaf ts and is shown on the Goodman diagram [3-11] in Figure 3-7.
Additionally, a review of the literature was conducted to fina other fatigue data for large diameter crankshafts. Nishihara and Fukui [3-12]
reported full scale fatigue test results for a large diameter crankshaft of a similar material, with an ultimate tensile strength of 94 ksi and found an 3-7
endurance limit of 35 ksi . This is in excellent agreement with the value of 35.7 ksi at zero mean stress calculated by FaAA.
3.4 Factor of Safety Against Fatigue Failure The factor of safety against fatigue failure for the RBS crankshafts at 3130 kW is computed to be 1.39 from the Goodman diagram in Figure 3-7. The stress point corresponding to crankshafts at SNPS at 3300 kW is also shown on the Figure.
The f actor of safety computed above is for operation at 450 rpm. At of fspeed conditions the ractor of safety may be adjusted to account for the difference in nominal torsional stresses. Thus, at 5% overspeed, the factor of safety would be approximately 9% less since nominal torsional stress is 9%
greater. However, continuous operation at offspeed conditions is prevented by the control of the governor.
3.5 Conclusions The following conclusions may be stated:
e The nominal torsional stresses in the RBS crankshaf ts at 3130 kW are approximately equal to those at SNPS at 3300 kW over the speed range. The stresses in the crankshaf ts at RBS are lower at underspeed and higher at overspeed than those at .
SNPS.
. The actual maximum crankpin fillet stresses at RBS at 3130 kW may be determined by scaling the strain gage test data from SNPS at 3500 kW.
. The endurance limit for the RBS crankshaft may be determined from the fatigue data from the 13-inch by 11-inch crankshafts at SNPS. This data is consistent with other fatigue data for large diameter crankshafts.
e The factor of safety against fatigue cracking of the RBS crankshafts at 3130 kW is 1.39. This factor of safety pro-vides sufficient margin for safe operation at loads up to 3130 kW.
3-8
Section 3 References 3-1 Timoshenko, S. , D.H. Young, and W. Weaver, Jr., Vibration Problems in Engineering. Fourth edition, Wiley, 1974.
3-2 Bercel , E. , and Hall , J.R. , " Field Test of Emergency Diesel Generator 103," Stone & Webster Engineering Corporation, April 1984 3-3 May,1985, SWEC Torsiograph Test Report for River Bend Station.
3-4 Nervik, N.R., and A. Sarsten, "COMHOL2 User's Manual," Division of Internal Combustion Engines and Marine Engineering, The Norwegi an Institute of Technology.
3-5 " Evaluation of Emergency Diesel Generator Crankshafts at Shoreham and Grand Gulf Nuclear Power Stations," Failure Analysis Associates Report No. FaAA-84-3-16, May 22, 1984 3-6 Fuchs, H.O. , and Stephens, R.I . , Metal Fatigue in Enaineering. Wiley, 1980.
3-7 Peterson, R.E., Stress Concentration Factor. Wiley & Sons, New York, 1974.
3-8 Bercel, E., and Hall , J.R. , " Field Test of Emergency Diesel Generator 101," Stone & Webster Engineering Corporation, October 1983.
3-9 " Emergency Diesel Generator Crankshaft Failure Investigation Shoreham Nuclear Poser Station," Failure Analysis Associates Report No. FaAA 10-2.1, October 31, 1983.
3-10 Structural Alloys Handbook, Volume 1, Battelle's Columbus Laboratories, page 45, 1981 Edition.
3-11 Collins, J.A., Failure of Materials in_ Mechanical Design. Wiley, 1981.
3-12 Nishihara, M., and Fukui, Y., " Fatigue Properties of Full Scale Forged and Cast Steel Crankshafts," Transactions of the Institute of Marine Engineering. Series B on Component Design for Highly Pressure-Chargea Diesel Engines, London, January 1976.
3-9
TABLE 3.1 NATURAL FREQUENCIES FOR DSR-48 13-INCH BY 12-INCH CRANKSHAFTS AT RBS Mode. Natural Frequency (Hz) 1 38.0 2 107.0 3 146.5 TABLE 3.2 TORSIONAL LOADING FOR FaAA ANALYSIS AT 3500 kW Order Torsional Loading, Tn (Psi) 1.5 112.0 2.5 77.0 3.5 48.0 4.0 33.0 4.5 26.2 5.0 21.2 5.5 15.5 3-10
+p--,,---w-.. .y_, e
__f., , y9 m. -y-- --yw i-.er-._7
I i
TABLE 3.3 t.
FREE-END VIBRATION LOAD FOR DSR-48 13-INCH BY 12-INCH CRANKSHAFT AT RBS AT 3500 kW Order Am litude of Vibration (degrees)
FaAA Analysis SWEC Test [3-3]
- 0.5 0.070 0.093 1.0 0.001 0.004 1.5 0.180 0.177 2.0 0.001 0.001 2.5 0.145 0.132 3.0 0.001 0.001 3.5 0.064 0.056
, 4.0 0.397 0.358 4.5 0.079 0.069 5.0 0.048 0.068 5.5 0.094 0.086 6.0 0.013 0.007 6.5 0.013 --
7.0 0.002 --
7.5 0.001 --
8.0 0.014 --
Vector Sumation 0.733 0.700 l
i 3-11
TABLE 3.4 TORSIONAL STRESSES FOR DSR-48 13-INCH BY 12-INCH CRANKSHAFT AT RBS AT 3500 kW Amplitude of Nominal
' Location _ Shear Stress (psi) 4th Order Total Between Cylinder No.1 and 716 3068 Cylinaer No. 2 Ber . ween Cylinder No. 2 and 1347 3293 Cylinder No. 3 Between Cylinder No. 3 and 1951 4842 Cylinder No. 4 Between Cylinder No. 4 and 2516 5666 Cylinder No. 5 Between Cylinder No. 5 and 3030 7357 Cylinder No. 6 Between Cylinder No. 6 and 3484 6191 Cylinder No. 7 6etween Cylinder No. 7 and 3867 5977 Cylinder No. 8 Between Cylinder No. 8 and 4176 5903 Flywheel 3-12
TABLE 3.5 C0tiPARIS0N OF SHOREHAM AND RIVER BEND CRANKSHAFT STRESSES AT 3500 kW Engine Speed Max Amolitude of Nominal Torsional Stress (psi)
(rpm) SNPS RBS FaAA Sarsten FaAA Sarsten FaAA Sarsten 427.5 428.0 7203 7050 7098 7077 436.0 437.0 7072 --
6999 7074 445.0 444.0 6962 --
7169 7206 450.0 450.0 7006 7090 7357 7427 455.0 454.0 7137 --
7574 7606 464.0 464.0 7500 --
7779 7893 472.5 473.0 7694 7850 8059 8172 3-13
TABLE 3-6 COMPARISON OF SHOREHAM AND RIVER BEND CRANKSHAFTS STRESSES AT REDUCED LOADS Engine Speed Max Amplitude of Nominal Torsional Stress (psi)
(rpm) SNPS at 3300 kW RBS at 3130 kW 427.5 6915 6573 436.0 6789 6481 ,
445.0 6684 6638 450.0 6726 6813 455.0 6852 7014 464.0 7200 7203 ,-
472.5 7386 7463 ,
3-14
TABLE 3.7 STRAIN GAGE RESULTS FROM SfiPS TEST AT 3500 kW WITH 13-INCH BY 12-INCH CRAtiKSHAFTS [3-2]
Strain Gage Maximum Minimum 5-1 (Compression) -195 pc 288 uc 5-2 (Bending) 695 pc -410 pc 5-3 (Tension) 737 pc -610 uc Major Principal Stress 26.2 ksi 4.9 ksi Minor Principal Stress -2.9 ksi -18.7 ksi l
l J
3-15
TABLE 3.8 STRAIN GAGE RESULTS FROM SNPS TEST AT 3500 kW WITH 13-INCH BY ll-INCH CRANKSHAFTS [3-8]
Strain Gage Maximum Minimum 5-1 (Compression) 1118 uc -707 uc 5-2 (Bending) 773 uc -459 uc 5-3 (Tension) -389 uc 266 uc Major Principal Stress 35.4 ksi 12.5 ksi Minor Principal Stress -4.2 ksi -22.0 ksi
. JL 3-16
e TABLE 3.9 OPERATING
SUMMARY
FOR SNPS EMERGENCY DIESEL GENERATORS WITH 13-INCH BY 11-INCH CRANKSHAFTS [3-9]
DG101 DG102 DG103 S/N 74010 S/N 74011 S/H 74012 TDI Factory Test Hours 128
- 30 40 110% 16 19 20 100% 180 254 249
> 75% 260 323 427
> 50% 51 75 77 Un'known (< 100%) 11 17 5 Total hours 646 718 818 Number of Starts
- 536 266 236 Hot Restarts 5 5 2 Engine S/N 74010 was used to qualify the R-48 product line for nuclear standby service and, therefore, experienced more factory test hours, including 300 starts.
3-17
PRESSURE (P S I G)
. , so .
t 300 -
inseI e.ij i
inst ';
\
l I "j ~ \
sd
,,c .
l _
-36C -180 (f00) 180. 36C.
CRANK ANGLE (DEGREES) l Figure 3-1. Measured pressure versus crank angle for SNPS at 3500 kw load.
FaAA-85-5-10
. i . i i . . . . i i i 8.0 -
m#~y{ ,
fo
$Di j_ 0
$ 6.0 5
/_< o Fa A A SNPS - 3500 kw
{hui 3.0
- Sarsten's program, SNPS - 3500 kw a = Sarsten's program, 2@F RBS - 3500 kw x _: 2.0 4<
lE z Note: Maximum stress 2 occurs between z 1.0 cylinders 5 & 6 I f f f f i 1 i i t t t j 430 440 450 460 470 ENGINE SPEED (rpm)
Figure 3-2. Comparison of RBS and SNPS nominal stress over the speed range of 450 rpm 5%.
t 1
FaAA-85-5-Io I
t 8.2 %
LESEPC 30000 -
o was0s es%:iet-,
a vauwsw sata =
c st%0:NO s?otss O To*out s-taa 8.5%
0 m
=
1 20.0v> -
w
.a y . 8.6%
o a w
C e
e-E 8.5 %
E d
e-S Change in C 1***'**
~
c stress for a 10%
change in load i.
e ee . .
C 1750 2700 2000 3*00 3600 2000 POWER KW Figure 3-3. Measured stresses as a function of load for SNPS.
FaAA-85-5-10
i i i i i i e i i i i i 8.0
+
w 7.0 'g,[ M 6.0
=
m 6
mm s.O O$
uJ a:
O $ 4.0
$ .a a RBS - 3130 kw g$
- SNPS - 3300 kw
< O 3.0 m
EF 2 .0
~
Note: Maximum stress occurs between yy '
cylinders S & 6 2_ z j 1.0 2 .
I I f I f f f f f f f f 430 440 450 460 470 ENGINE SPEED (rpm)
Figure 3-4. Comparison of RBS and SNPS nominal torsional stresses at reduced loads over the speed range of 450 rpm 5%.
FaAA-85-6-Io
~
.oos
, .cos I 5/16 i 3/4 R 1.1875 Detall A g 1/4 R
, ! to' , i
& 3 l 1/16 - - +
s-c
-3 8 e.
u:
EE EE o 4
O
?I hI O T C tD n
l 17.sn g l l l i
is sn e.o n g ,
'N \
l 1 \
l 1 t
i ;
- /- 1 1. 5 R i j , /
I t #
1 I /
i /
g g
,/ l .-
/ I I I
/ I
/ I l '
_ _ t _ __ _ _ _ I__
\ Detail A i
\ , l
\ i N I
\
l \ 6 .5 R e.sn I '-- - 1 I u
l is.2 s n ~ 4.s -
10.5 :
Figure 3-5. Typical structural dimensions of the RBS crankshafts.
{
l Fa A A-85-5-10 ta .
_ -x.----. - - - - - - -
WROUGHT 1040 C-0. 39, Mn-0. 8 2, St-0. 25, P-0.020,S-0,037 NORM 1650F, T 1200F Fw = 90 KSI, e = 26. 5 %
CAST 1040 C-0. 39.Mn-0. 78, Si-0. 4 7 P-0.023,S-0.039 NORM 1650F T 1200F Fm = 94. 2 KSI, e = 24. 51 UNNOTCHED NOTCHED WROUGHT O O CAST D D 55 NORht4UZED AND TEMPERED 50 ! '
L%
45
- v. 4 0 "
35 k 30 6
25 ' ' ' ' ' ' '
10 4 5 6 10 10 10 7 CYCLES TO FAILURE Figure 3-6. Endurance limit for 1040 steel.
l l
1 FaAA-85-5-10
40 . . . . . , . . .
De = 35.7 k si- f-S N P S 13"x 11", 3 5 00 kw -
e oSNPS - 3300 kw
= oRiver Bend - 3130 kw e 30 -
v>
> @ 2 3.9 ;a ~
2 c: 23.6 "l $
4 20 -
Factor of 3g safety for -
-z RBS Endurance limit for oP w< -
crankshaf t = 1.39 [RBS cranksht f t -
2 m
10 -
O ' ' ' ' ' ' '
O 20 40 60 80 l 100 EQUIVALENT MEAN STRESS (ksi) UTS - 94 ksi Figure 3-7. Goodman diagram for River Bend crankshafts.
4 A
, FaAA-85-5-10
4.0 CRANKSHAFT ENDURANCE TESTING The NRC Safety Evaluation Report on the TDI Diesel Generator Owners Group Program Plan, [4-1] provides that where TDI diesels are to be operated at brake mean effective pressures exceeding 185 psig, the applicant should submit information demonstrating that components, which are of the same design, have operated successfully for at least ten million load cycles which meet or exceed the maximum emergency service load requirements for the subject engines. The following comparison can be nade with an endurance test per-formed on one of the SNPS crankshafts. Alternatively, a design may be quali-fied by testing a prototype to 107 cycles at the rated load where additional data is available to demonstrate an adequate fatigue margin. In this case the endurance test is beir,g used to provide additional assurance of the adequacy of the design even though the test by itself would not represent suf ficient assurance.
A 107 cycle endurance test was run on a DSR-48 engine crankshaft with 13-inch main journals and 12-inch crankpin journals at a load of 3300 kW* at SNPS. The crankpin fillet and oil hole geometry is identical for the crank-shafts at RBS and those at SNPS. The crankshafts were made according to TDI drawing number 03-310-05-AC.
While the crankshafts are of the same design for the two stations, the flywheel and generator are somewhat different. This difference leads to a small change in the torsional stresses for the crankshafts as discussed in previous sections of this report. To account for these design differences the load at RBS that produces the same torsional stress as a load of 3300 kW at SNPS may be calculated. This calculation is based on the linearity of stress with load which has been established in Sections 2.0 and 3.0. Based on a 7%
change in stress for a 10% change in load, a load of 3071 kW at RBS produces the same stress as a load of 3300 kW at SNPS, as shown in Figure 4-1.
The encurance test has been stated to be nominally at 3300 kW. However, Dr. Pischinger (FEV) performed a cumulative damage analysis of the crankshaf t based on the actual loads experienced during the 745 hour0.00862 days <br />0.207 hours <br />0.00123 weeks <br />2.834725e-4 months <br /> endurance test and found the load to be equivalent to 3505.kW [4-2].
4-1
The material specifications for the crankshaf ts at RBS and SNPS were the same. Crankshaf ts at both plants were required to meet ABS Grade 4 which has a minimum ultimate tensile strength, UTS, of 83000 psi. The UTS measure-ments for RBS [4-3, 4-4] and SNPS [4-5, 4-6, 4-7] are shown in Table 4.1. The lowest measurement at RBS is 94000 psi and at SNPS is 100800 psi. The crank-shaf ts at both stations meet the minimum requirements of ABS Grade 4 and have UTS values within the expected range for this material . The lowest measure-ment at RBS is about 7% lower than the lowest measurement at SNPS.
The fatigue limit for these crankshaf ts is approximately proportional to the UTS of the material. Thus, the f actor of safety in the SNPS crank-shafts at 3300 kW is approximately 7% higher than in the RBS crankshafts at 3071 kW. However, this effect is small in comparison to the factor of safety in the RBS crankshafts.
The crankpin fillets at SNPS are shot peened whereas those at RBS are not. While no credit was taken for the effect of shot peening in computing fatigue margins, it is possible to estimate the maximum effect that it could have had on the crankshsft endurance test. The stresses in the most highly stressed main journal oil hole is computed to be about 7% less than that of the most highly stressed crankpin fillet. Thus, if shot peening has more than a 7% effect on the crankpin fillets, the oil holes will govern crankshaf t life. This 7% effect is quite small when compared to the safety margin.
The endurance testing at SNPS along with RBS actual experience are plotted against the crankshaft S-N curve in Figure 4-2. It is seen that for the RBS crankshafts at 3071 kW, the stresses are significantly below the l actual SNPS endurance test of 3500 kW. The margin between the RBS stresses at j 3071 kW and the S-N curve for the RBS crankshaft corresponding to the lowest l UTS measurement is adequate to ensure reliable operation of the RBS crank-shafts for their intended service.
4-2
t v
4.1 Conclusions The following conclusions may be stated:
- The crankshafts at RBS are geometrically identical to those at SNPS e At 450 rpm the torsional stresses at RBS at 3071' kW are equal to those at SNPS at 3300 kW e Although the lowest UTS at RBS is lower than that at SNPS, the material specifications are the same and the material meets the requirements of ABS Grade 4.
e The advantage of shot peening is limited to about 7% since the oil holes are not shot peened.
Section 4 References 4-1 U.S. NRC Safety Evaluation Report, Transamerica Delaval, Inc. Diesel Generator Owners Group Program Plan, August 13, 1984 4-2 U.S. NRC Atomic Safety and Licensing Board hearing on Shoreham Nuclear Power Station, Docket No. 50-322-OL, Tr. 28419-20.
4-3 American Bureau of Shipping, Report on Castings and Forgings for the crankshaft in engine 74039, 4-6-80.
4-4 American Bureau of Shipping, Report on Castings and Forgings for the crankshaft in engine 74040, 31-7-80.
4-5 Krupp Stahl, Schmiede Und Bearbeitung for the crankshaft in engine 74010, 17-10-83.
4-6 American Bureau of Shipping, Report on Castiags or Forgings for the crankshaft in engine 74011, February 4,1983.
4-7 American Bureau of Shipping, Report on Castings or Forgings for tFe crankshaft in engine 74012, February 4,1983.
I 4-3
TABLE 4.1 COMPARISON OF RBS AND SNPS ULTIMATE TENSILE STRENGTHS Engine . Ultimate Tensile Station Designation Serial No. Strength (psi) i RBS EG1A 74039 94000 97000 RBS- EGIB 74040 99500 101000 SNPS DG101 74010 100800 100800 SNPS DG102 74011 101800 106500 SNPS DG103* 74012 100800 106500 ABS Grade 4 Requirement 83000 SNPS 107 cycle endurance test at loads equal to or greater than 3300 kW was performed on this engine.
i a
T ,
4 4
m
- 4-4
(
1 l
J 8000 , , , ,
e m RBS w
c SNPS F 6726 - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
m '
i j 6000 -
5 l l 2
O ll i ,
t a ! l O* I i w E 4000 l
o o 8 D h" t- !li i
- ' i Q. I I 2 '
< l 2000 -
2 l
I l -
i o i e 2 '
l x i !
< 3071 kw i i e 2 '
0 ' ' '[! 3300' kw O 1000 2000 3000 4000 LOAD (kw)
Figure 4-1. Loads at RBS and SNPS that result in equal
. crankshaft stresses.
e-
- Fa A A-85-5-10
c 50 , , ,
as i m M .*..'-.. S-N curve for 13"x 12" crankshaf t at E S-N curve for f ailed y 40 -
13*x 11* crankshaf t at ' " - SNPS (UTS = 101 ksi)
' * . . ...' ~
g SNPS (UTS : 96 ksi) SNPS 13* x 11' z . .. . " at 3500 kw us . .
< S-N curve for RBS crankshaf t ' ' *
(UTS = 94 kai) ,,"""*"--"--
4 2 -
a 30 -
O SNPS 13"x 12" actual onduranco test (3500 kw) ~
us RBS actual experience at 3500 kw o a O UA z RBS
, p o i < 3071 kw z 20 -
m SNPS 13'x12" nominal endurance test (3300 kw) uJ F-
_J u.
O 10 -
m -
- O a
.H
_J n.
2 m < o I i 1 10 4 10 5 6 10 7 I 10 10 8 g STRESS CYCLES (13,500/hr at 450 rpm) u'.
Figure 4-2. Comparison of RBS and SNPS crankshaft stresses and experience with S-N curve.
(
5.0 CRANKSHAFT FILLET AND OIL H0LE INSPECTIONS Crankpin journal and main journal oil holes and crankpin fillets on diesel engines EG1A and EG1B at RBS were examined to identify any crack indi-cations. The inspections were performed by FaAA personnel in July 1984. The operating history of engines EG1A and EG1B prior to these inspections were summarized in Tables 2.5 and 2.6. The results of the inspections are dis-
~
cussed below. The recommended acceptance criteria for eddy current inspec-tions are given in References 5-1 and 5-2.
5.1 Crankpin Fillet Eddy Current Examination Eddy current examinations were performed on the crankpin fillets on cylinders No. 5 governor end and cylinders no. 6 through 8 governor and gener-ator ends on both engines [5-3]. FaAA NDE 11.6 Revision 0 eddy current test procedure was used. The eddy current recording criterion was any crack-like indication exceeding 25 percent of the signal obtained from the 1/16 inch long by 1/ 32 inch deep simulated crack in FaAA crack standard PA07395-831228. No recordable indications were observed.
5.2 Crankshaft Oil Hole Liquid Penetrant Examination Liquid penetrant examinati ans were performed on the main journal oil hole radius of Journals No. 7 through No. 9 and on the crankpin oil hole radius of crankpins No. S through No. 8 in both diesel engines [5-4]. FaAA NDE 6.4 Revision 0 liquid penetrant examination was used. No linear (crack-like) indications were observed.
5.3 Crankshaft Oil Hole Eddy Ctr rent Examination Eddy-current examinations were performed on main journal oil? holes in journals No. 7 through No. 9 and crankpin journal oil holes in crankpins No. 5 through No. 9 in both engines [5-5]. The crankshaft oil holes were inspected to a depth of 3.0 inches. The acceptance criteria for the individual oil holes is specified in Reference 5-2. Main Journal No.10 was not inspected since it is not a through-depth oil hole.
5-1
_,, ._ iy m, . - - - - --r
FaAA NDE 11.3 Revision 0 eddy current procedure was used. The crack reference (FaAA PA07395-74073) is a simulated oil hole machined in a segment of crankshaft and containing a simulated crack of 0.040 inch long by 0.020 inch deep.
Eddy current indications were observed exceeding the recording threshold at various depths in all of the oil holes inspected. An interpreta-tion of the results of this test was made utilizing the most recent FaAA eddy current recording criteria for oil holes (NDE 11.3, Revision 2). Using this procedure only 5 of the 28 oil holes inspected had recordable indication
[5-2]. The recommended oil hole acceptance criteria is detailed by cylinder location and broad ranges of depth from the journal surface (0" to 1",1" to 2", and 2" to 3").
Further evaluation of the allowable notch sizes for the five oil holes at the depth at which the threshold readings were exceeded was made [5-5]. The results indicated that for these oil holes the readings are within the allowable criteria.
5.4 Conclusions The following conclusions can be stated:
. The eddy current inspection of the fillets on EG1A and EG1B showed no recordable indications were present.
e The results of the liquid penetrant inspections of the main journal and crankpin journal oil hole radii indicated that no crack-like indications were observed.
e A review of the results of the eddy current inspection of the main journal oil holes on EG1A and EG1B indicates that all readings are within the acceptable criteria.
References '
5-1 Memorandum to John C. Xammeyer, DR/QR Program Manager from Paul R.
Johnston, Crankshaft Design Review Task Leader, dated 7/26/84, re: Oil Hole Inspections for DSR-48 Crankshafts at River Bend.
5-2 Memorandum to Duane Johnson, FaAA NDE Manager, from Paul R. Johnston, Crankshaf t Design Review Task Leader, dated 7/9/84, re: Inspection of DSR-48 Crankshafts of River Bend.
i 5-2
_ , . , ._., m - * r "" ' ~
'N '"-#
~ _
.i 5-3 FaAA Eddy Current Examination 1 Report Nos. 840710-101, 840710-201,.
840711-202, 840711-203, dated July, 1984.
5 FaAA. Liquid Penetrant Examination Report Nos. 840715-603, 840715-604 dated July, 1984.
5-5 Hand- Calculation " Review of NDE Crankshaft Inspection Reports,"
FaAA-SP-85-5-10.
s 5-3
.-