ML20133A314
| ML20133A314 | |
| Person / Time | |
|---|---|
| Site: | Perry, Catawba, Harris, Grand Gulf, River Bend, Vogtle, San Onofre, Comanche Peak, Rancho Seco, Midland, Bellefonte, 05000000, Washington Public Power Supply System, Shoreham |
| Issue date: | 09/27/1985 |
| From: | Laity W Battelle Memorial Institute, PACIFIC NORTHWEST NATION |
| To: | Berlinger C Office of Nuclear Reactor Regulation |
| References | |
| NUDOCS 8510020182 | |
| Download: ML20133A314 (26) | |
Text
.
kllh September 27, 1985 Pacific Northwest Laboratories P.O. Boa 999 Richland, Washington U.S A. 99352 Telephone (509) 375-2780 Dr. Carl Berlinger Division of Licensing Tele = 15-2874 Office of Nuclear Reactor Regulation U.S. Nuclear Regulatory Commission 7920 Norfolk Avenue Bethesda, Maryland 20014
Dear Dr. Berlinger:
r
SUBJECT:
TRANSAMERICA DELAVAL, INC. DIESEL GENERATOR PROJECT - PNL'S REVIEW 0F CRANKSHAFTS FOR TDI 16-CYLINDER, V-TYPE, DSRV-4 SERIES ENGINES (SECTION 4.7 0F PNL-5600 DATED OCTOBER 1985)
In response to your request, I am submitting the subject review in advance of the technical evaluation report in which it will be incorporated.
It will be Section 4.7 of the report titled Review of Resolution of Known Problems in Engine Components for Transamerica Delaval., Inc., Emergency Diesel Generators (PNL-5600 dated October 1985). Section 4.7 has received PNL clearance for technical content.
Except for the page numbers, it is typed as it will appear in final form.
FNL's current plans call for several diesel engine consultants to review PNL-5600 before it is printed and distributed. That review has not yet occurred, because certain sections of the PNL report are still in prep-a ra tion.
However, the enclosed Section 4.7 fully reflects the views previously expressed by PNL's consultants in their reviews of the crank-shafts for the 16-cylinder TDI engines. Accordingly, I anticipate that
=
any further review of the final text will result in no substantive changes to the conclusions and recommendations documented in Section 4.7.
Sincerely, M. W W. W. Laity PNL Project Manager cc:
M. Carrington, NRC (2)
M. Plahuta, DOE (RL) gaan!RBeTgg6 Af [.e4 #'#[A S
i l
4.7 CRANKSHAFT
16-CYLINDER, DSRV-4 SERIES ENGINES PNL reviewed the action taken by the Owners' Group to evaluate the crank-
. shafts for the TDI 16-cylinder, V-type, DSRV-4 series engines at the Catawba, Comanche Peak, Grand Gulf, and Perry nuclear power stations. The conclusions drawn from PNL's review are applicable only to these installations. However, these conclusions may be extended to crankshafts of the same design in TDI 16-cy11nder engines at other nuclear installations, provided that the crank-shaft mechanical properties and the torsional crankshaft stresses are shown to be similar to those in the installations just. mentioned.
PNL's evaluation is presented following a discussion of the crankshaft reviews, analyses, and tests performed for the above-named installations by consultants to the Owners' Group and to individual licensees.
4.7.1 Component Description The crankshaft for TDI 16-cylinder, DSRV-4 series engines is a steel forging with eight crank throws driven by 16 articulated connecting rods, through which reciprocating power is transmitted from pistons arranged in two 8-cylinder banks in a V-type engine block. Each crankshaft in the above-mentioned installations is equipped with four counterweights. The crankpin journals and the main bearing journals are 13 inches in diameter, and the overall crankshaft length is approximately 20 feet 7 inches.
'The 16-cylinder, DSRV-4 diesel generators purchased for nuclear service are rated by TDI 'fo'r continuous operation at 7000 kW when loaded to an engine brake mean effective pressure (BMEP) of 225 psig at the design speed of 450 rpm. These engines are also rated by TDI to operate for 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> of every 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> at 7700 kW.
4.7.2 Design Guidelines As discussed in Section 4.6.2 of this report, the recommendations in Standard Practices (1972) of the Diesel Engine Manufacturers Association (DEMA) were used as a basis for the crankshaft evaluations performed by the TDI Diesel Generator Owners' Group and the reviews performed by PNL. The DEMA recommen-dations for torsional crankshaft stresses are summarized in that section.
1 1
.-_e
i Torsional crankshaft stresses determined through analyses and tests of 16-cylinder engines at nuclear power plants are compared in this report with the DEMA recommendations.
4.7.3 Failure History On the basis of information provided by TDI, FaAA reported (FaAA-84-3-16, 4
May 1984) that three crankshafts for 16-cylinder, DSRV-4 engines have failed since 1976. All three failures occurred in engines in non-nuclear service.
One failure occurred at Mora, Mirinesota, in February 1976, in main journal No. 8.
Two failures (one in June 1976, the other in March 1979) occurred at the Anamax mine near Tucson, Arizona. One was located in main journal No. 8; the other was in main journal No. 6.
TDI attributed all three failures to torsional fatigue cracks initiating in the oil holes of these main journals.
These engines were found to have a 4th-order critical speed at 446 rpm, which is close to the operating ~ speed of 450 rpm. Each engine was subsequently fitted with four counterweights, which lowered the 4th-order critical speed to about 430 rpm.
In addition, TDI changed the angular location of the oil holes and increased the radius where the ends of 'each hole intersect with the journal surface.
FaAA noted in the report referenced above that torsional stress is inde-pendent of angular location, and that the oil holes are areas of stress ccncen-tration. Accordingly, FaAA views the oil holes as critical sites for possible fatigue crack initiation, even with the modifications just described.
4.7.4 Owners' Group Evaluation: Grand Gulf Engines The TDI engines at the Grand Gulf Nuclear Station (GGNS) were designated by the Owners' Group as the " lead plant" 16-cylinder engines. Crankshaft stresses in these engines were evaluated by torsiograph testing and by analy-sis, and the results formed the basis for reviews conducted by the Owners' Group of crankshafts in "following plant" 16-cylinder engines. Although the 16-cylinder engines in the four nuclear plants discussed in this section are similar, differences in generators and flywheels result in differences in crankshaft torsional stresses. These differences have been addressed by the Owners' Group in the "following plant" reviews.
2
The adequacy of the crankshafts for the TDI engines at the Grand Gulf Nuclear Station was evaluated by Failure Analysis Associates for3he Owners' Group, and by Bechtel Power Corporation for Mississippi Power & Light Company, the owner of GGNS.
4.7.4.1 FaAA's Evaluation FaAA's evaluation of the crankshafts for the GGNS engines is presented in two reports: FaAA-84-3-16 (May 1984) and FaAA-84-7-10 (July 1984).
In the first report, FaAA reviews the tor'sional critical speed analysis performed by TDI in October 1975 and compares the calculated responses with the results of a torsiograph test that TDI performed on a GGNS engine before it was shipped from TDI's Oakland plant.
In addition, the results of FaAA's dynamic torsional analysis are presented. The torsiograph test that FaAA performed on an engine at GGNS is addressed in the second report.
4.7.4.1.1 FaAA's Review of TDI's Evaluation. As discussed in FaAA-84-3-16, diesel generator torques due to dynamic response are usually cal-culated in two steps. The first step is to determine the natural frequencies of vibration of the crankshaft torsional system, the corresponding engine speeds, and the orders 'of vibration that resonate at these speeds. The second step is to determine the dynamic torsional response of the crankshaft due to gas pressure and reciprocating inertia loading. TDI calculated the response at the rated engine load of 7000 kW.
By modeling the crankshaft torsional system as lumped mass moments of inertia connected by torsional springs and using the Holzer method to solve the
~
resulting eigenvalue problem, TDI determined that the first three natural frequencies of.the crankshaft are 28.8, 83.0, and 113.0 Hz. The first natural frequency produces a 4th-order resonance at 432 rpm.
FaAA comented that the response of the 4th-order resonance is important because of the proximity of this resonance to the engine operating speed of 450 rpm. As noted by FaAA, the 4th-order loading from one bank of a V-16 engine with articulated connecting rods almost cancels that from the other bank, reducing the excitation. However, FaAA also noted that this excitation is sensitive to the balance between the two banks.
3
TDI determined the dynamic torsional response of the crankshaft for whole-and half-orders of vibration from 0.5 to 12.0, each considered separately.
(The 1st order is a harmonic that repeats once per revolution of the crank-shaft. Harmonics of order 0.5, 1.0, 1.5, 2.0, 2.5... exist for a four-stroke engine.) The largest nominal shear stress amplitude computed by TDI for a single order of vibration is 1956 psi, corresponding to the 3.5 order.
This is well below the DEMA-recommended limit of 5000 psi for a single order.
TDI also performed a torsiograph test on one of the GGNS engines.
A torsiograph test measures the angular displacement of the free end of the crankshaft under various operating conditions, and is commonly used to confirm torsional vibration calculations. A typical test consists of first operating the engine without load at various speeds to locate critical speeds, and then operating the engine at its rated speed with variable load to determine the forced vibrational response.
FaAA compared the TDI-calculated responses to the results of TDI's torsio-graph test. The variable speed test showed that the first natural frequency is 28.7 Hz, which agrees well with TDI's calculated value of 28.8 Hz. The nominal shear stress amplitude for the maximum measured single-order response at the rated load of 7000 kW is 2028 psi, corresponding to the 3.5 order. This com-pares well with the stress of 1956 psi for TDI's calculated response to the 3.5 order, and is well below the DEMA-recommended limit of 5000 psi.
From torsio-o graph data taken by TDI at 7700 kW (110% rated load), the nominal shear stress amplitude for the largest single-order response (the 3.5 order) is 2366 psi.
'e FaAA also commented on crankshaft stresses for off-speed conditions. The stresses are within DEMA-recommended limits over a speed range of 440 rpm to 450 rpm +5%, according to FaAA, provided that adequate engine balance is main-tained. FaAA recommends that the engines not be allowed to operate below 440 rpm except during startup and shutdown, because of the presance of the 4th-order resonance at 432 rpm.
4.7.4.1.2 FaAA's Dynamic Torsional Analysis. FaAA noted that TDI did not calculate the phase angle associated with the response of each order, and therefore the combined response of the orders cannot be determined from TDI's results. FaAA also noted that the measurements necessary-to determine stresses 4
for the combined orders were not taken during TDI's torsiograph test. Thus, neither TDI's calculations nor TDI's measurements can be used to-compare the stresses for combined orders with the DEMA-recommended limit of 7000 psi.
To supplement TDI's calculations, FaAA analyzed the. combined response of the crankshaft using modal superposition for the first 24 orders of vibration.
This analysis was similar to FaAA's torsional analysis of the crankshafts for the TDI engines at the Shoreham Nuclear Power St'ation (SNPS).
All 11 modes of the lumped inertia and torsional ipring.model of the crankshaft were considered in the analysis.
In calculating the harmonic loading on the crankshaft, FaAA considered gas pressure, reciprocating 1nertia, and estimated frictional loads. The gas pres-sure loading was obtained from pressure versus crank angle data measured in an 8-cylinder engine at SNPS. FaAA considered these measurements to be applicable I
to the-16-cylinder engines at GGNS, because the cylinders of TDI R4-series I
engines are designed to operate under the same conditions regardless of the 1
number of cylinders.
1 To simulate the effects of cylinder bank imbalance on crankshaft response to the 4th order of vibration, FaAA's analysis assumed a 1-degree difference in timing between the two banks. Damping equal to 2.5% of critical damping was
)
. assumed for the analysis, but FaAA found that the assumed value had little effect on the calculated response because the orders of vibration are not in resonance at the engine operating speed of 450 rpm.
FaAA's calculated amplitudes 'of free-end displacement for the significant orders of vibration agree well with values measured in TDI's torsiograph test of the engine. The largest response occurs for the 3.5 order. FaAA's calcu-lated amplitude for this order at full load is 0.269 degrees. The correspond-4 ing value from TDI's torsiograph test is 0.24 degrees.
FaAA's results also include the range of torque at each crank throw, and i
the corresponding nominal shear stress amplitude. The calculated stress levels are highest between cylinders No. 3 and 4, 5 and 6, and 7 and 8.
FaAA found r
l-that the highest nominal shear stress amplitude for the vector summation of 24 i
orders at full rated engine load is 5367 psi, occurring between cylinders No. 5 l
5 w
and 6.
This value is below the DEMA-recommended limit of 7000 psi for combined orders. FaAA commented that the nominal shear stress amplitudesT or 110% load f
may be determined by extrapolation, and are also below the DEMA-recommended limit. However, FaAA did not report the values for 110% load.
4.7.4.1.3 FaAA's Torsiograph Test. As reported in FaAA-84-7-10, Fa AA performed a torsiograph test of a TDI diesel generator at GGNS at eight speeds between 410 and 470 rpm under no-load conditions, and at loads of 25%, 50%,
75%,100%, and 110% of rated load :while the engine was operated at rated speed.
The variable speed test was performed to determine the frequency of the first mode of the crankshaft torsional system. The variable load test was performed to determine the amplitude of free-end crankshaft vibration and to estimate the nominal shear stress as a function of load.
FaAA's findings and conclusions from the torsiograph test include the following:
The 4th-order critical speed is reached at about 430 rpm. Thus, the e
first natural frequency is approximately 28.7 Hz. This is in good agreement with TDI's calculated value of 28.8 Hz.
With the assumption that the shaft is vibrating in its first mode, e
the following values of the amplitude of nominal shear stress were estimated from the measured amplitudes of free-end vibration:
Full Load Overload DEMA 7000 kW 7700 kW allowable Single order (3.5) 2062 psi 2172 psi 5000 psi Combined response 4775 psi 5113 psi 7060 psi e The results of the variable load tests show that the larnest responses occur for the 1.5, 2.5, and 3.5 orders, and increase with increasing load. However, the 4th-order response increases slightly over a portion of the load range and then decreases, with the result that it is approximately the same at full load as at no load.
6
t 4.7.4.1.4 FaAA's Conclusions and Recommendations. FaAA'sjonclusionsand recommendations from the analyses and tests of the cran.kshafts for the GGNS engines include the following, as discussed in FaAA-84-3-16:
The single-order calculations performed by TDI for a GGNS engine are e
appropriate and show that the crankshaft stresses are below the DEMA-recomended limit for a single order.
The results of TDI's torsiograph test. of a GGNS engine also show that e
the crankshaft stresses are below the DEMA-recommended limit for a single order, for both 7000 kW (rated load) and 7700 kW (110% rated load).
c..
e The crankshafts are adequate for their intended service, subject to the following recomendations:
The oil holes in main journals No. 4, 6, and 8 represent a more critical stress concentration in torsion than do the crankpin fillets and should be inspected for fatigue cracks and machining
. discontinuities.
The engines should not be allowed to operate below 440 rpm except during startup and shutdown.
The adequacy of' the TDI specification for balancing of cylin-ders, in combination with the speed tolerances allowed by the governor, should be determined by torsiograph testing. These
'e tests would provide a database for the expected variations of speed and balance. With the additional data, it may be possible to eliminate inspection of oil holes.
If an engine is operated in a severely unbalanced condition, it may be necessary to reinspect the oil holes for fatigue cracks.
4.7.4.2 Bechtel's Evaluation In a report submitted to PRC (H. Denton) as an enclosure to a Mississippi Power & Light Company (J. McGaughy) letter dated February 20, 1984, MP&L-
~
described the action taken at that time on the TDI engines at GGNS.
Included in the report are the results of a crankshaft evaluation performed for MP&L by 7
the Bechtel Power Corporation.
In an independent dynamic analysis of the crankshaft, Bechtel used five sets of harmonic coefficients and'5pplied modal superposition to sum the effects of the responses. Cylinder pressures measured at Shoreham were used in determining the harmonic loading.
Bechtel's torsional analysis yielded a " single order stress" of 2389 psi and a " total average stress" of 5084 psi for the GGNS crankshafts at rated load. These values are consistent with the results obtained by TDI and by FaAA for single and combined orders, and are within the DEMA-recommended limits.
Bechtel also determined that the peak stress over a stress cycle is 6034 psi.
This, too, meets DEMA criteria.
4.7.5 Owners' Group Evaluation:
Perry Engines
. FaAA performed torsiograph tests and related analyses of the crankshafts for the Unit 1 Division 1 (U1D1) and the Unit 1 Division 2 (01D2) diesel gener-ators at the Perry Nuclear Power Plant (PNPP). Both units are TDI 16-cylinder, DSRV-4 series engines.
For each engine, FaAA obtained data at 10 speeds between 400 and 470 rpm under no-load conditions, and at leads of 25%, 50%, 75%, and 100% of rated load while the engine was operated at rated speed. FaAA also obtained transient data during fast starts of each engine from four different initial crankshaft positions, and during four coastdowns from rated speed. The initial positions r
for the four startup tests were at 180-degree intervals of crankshaft rotation to covei the full 720-degree firing cycle.
Finally, FaAA measured the crank-shaft response of each engine to operation under conditions of cylinder imbal-ance. The imbalance was. introduced by cutting off the fuel to one cylinder (cylinder No. 5, left bank), chosen on the basis of an analysis that showed it would produce the largest increase in stress.
The steady-state and transient tests are discussed in FaAA-85-4-1 (May 1985). FaAA's conclusions are as follows:
The '4th-order critical speed is reached at about 436 rpm for each e
crankshaft.
Thus, the first natural frequency of the torsional 8
4 system is approximately 29.1 Hz. This value is in good agreement with TDI's calculated value of 29.2 Hz for the PNPP engine 5~.'
With the assumption that the shaft is vibrating in its first mode, e
the following values of the amplitude of nominal shear stress were estimated from the measured amplitudes of free-end vibration:
Ampiitude of Nominal Shear Stress at Full Load (7000 kW)
~
Diesel Generator Single Order Combined Order Unit 1 - Division 1 1891 psi 4659 psi Unit 1 - Division 2 2020 psi 4642 psi i
DEMA-recommended limit 5000 psi 7000 psi The coastdown transient response is repeatable and has a maximum e
peak-to-peak amplitude of approximately 0.96 degrees, which produces a maximum amplitude of nominal stress of 3970 psi.
i A typical transient response to a 6-second fast start has a maximum peak-to-peak amplitude of 1.86 degrees. The effect of initial crankshaft position on the transient response is small. This was confirmed by a transient analysis for each of the four conditions tested. The analysis, which used ~ pressure-time data recorded during a fast sthrt on a DSRV-4 series engine at anot'her plant, indicates that a typical fast start produces a maximum amplitude of nominal stress of 7650 psi, occurring between cylinders No. 7 and 8.
This stress amplitude exists for only a few cycles on each startup.
The results of the torsiograph test indicate that the crankshafts are e
adequate for their intended service at PNPP.
The cylinder imbalance tests are discussed in FaAA-PA-R-85-06-11 (June 1985), an addendum to FaAA-85-4-1. FaAA noted that the effect of imbalance is typically greatest in the 0.5-order response. However, the response of the 4th order to cylinder imbalance was of particular interest for the PNPP engines, because the critical speed (436 rpm) for that order is close to the normal operating speed of 450 rpm.
9 5
2 ew
,g
-,.,--,..,--.m
-_e-,_
6--,y..%.,,-m, 3-c--,w
.-m--
my mn w -r e, - 7.,m m--, m m-w..._wer-,.--,,,,-,,--,-w,r-,9.-,.,-
FaAA measured the effects of imbalance at no load and at 50% of rated
~
load. Using the modal superposition model discussed in Section 4.7.4.1.2 of this report, FaAA extrapolated the effects of imbalance to full load (7000 kW). FaAA's conclus:ons include:
The U102 engine is better balanced (i.e., the pressure versus time e
curve is close to the same for each cylinder) than the U1D1 engine, as indicated by the lower 0.5-order response of U102 during both normal and imbalanced operation. However, the 0.5-order response for
'both engines was within the normal range.
The total response during the cylinder imbalance tests with no load e
is essentially the same as the response during normal balanced operation with no load.
An analytical extrapolation of the measurements at 50% of rated load e
indicates that, at rated load (7000 kW), the maximum amplitude of nominal stress would increase from 5330 psi to 6910 psi due to the imbalanced condition. This increase is not large enough to affect the adequacy of the crankshaft in eith' r engine.
e The 4th-order response showed a small increase in the U101 engine and e
a large increase in the U102 engine during the imbalance tests. The difference in the responses is due to the different balance between the left and right banks in the two engines.. However, neither response is large enough to affect crankshaft adequacy.
I 4.7.6 Owners' Group Evaluation: Catawba Engines Duke Power Company, the owner of Catawba, performed extensive tests on two 16-cylinder TDI diesel generators, designated 1A and IB, that are installed at Catawba Nuclear Station Unit -1.
In addition, FaAA performed a torsiograph test of the 1A engine. The operational tests and the torsiograph te't are discussed under the following two subheadings.
4.7.6.1 Operational Tests The engine tests performed by Duke Power Company provide substantial evidence concerning the fatigue resistance of the crankshafts under the 10
4 conditions imposed during the tests. These tests are the subject of numerous letters and reports-reviewed by PNL in the p, reparation of a technical evalua-tion report on the Catawba engines (PNL-5211, August 1984).
In summary, diesel generator 1A was operated for more than 800 hours0.00926 days <br />0.222 hours <br />0.00132 weeks <br />3.044e-4 months <br /> in a test program completed in March 1984. Over half of those hours were at loads equal or to greater than 5800 kW (83% of rated load).- Diesel generator IB was operated for more than 750 hours0.00868 days <br />0.208 hours <br />0.00124 weeks <br />2.85375e-4 months <br /> in a test program completed in July 1984 Over 600 of those hours were at loa ~ds equal to or greater than 5800 kW.
Post-test inspections of the crankshafts, including fluorescent dye penetrant tests of the lube oil holes subject to the highest torsional stresses (i.e., in main journals No. 4, 6, and 8), revealed no evidence of fatigue crack initiation.
4.7.6.2 Torsiograph Test As reported in FaAA-84-5-23 (May 1984), the torsiograph test was performed at nine speeds between 410 and 470 rpm under no-load conditions, and at loads of 50%, 75%, 100%, and 110% of rated load while the engine was operated at rated speed.
FaAA reached the following conclusions:(a)
The 4th-order critical speed is reached at about 429 rpm, indicating e
that the first natural frequency is 28.6 Hz. This agrees well with TDI's calculated value of 28.8 Hz for the Catawba engines.
With the assumption that the shaft is vibrating in its first mode, e
the following values of the amplitude of nominal shear stress were s
estimated from the measured amplitudes of free-end vibration:
Full Load Overload DEMA 7000 kW 7700 kW allowable Single order (3.5) 2079 psi 2172 psi 5000 psi Combined response 4987 psi 5071 psi 7000 psi (a) PNL notes that engine 1A was equipped with type AN piston skirts at the time FaAA conducted the torsiograph test at Catawba. The AN piston skirts were later replaced with type AE piston skirts.
In the opinion of PNL's reviewers, this change in piston skirts has no significant effect on the torsional system. Accordingly, the change does not affect the conclusions reached by FaAA.
11
4.7.7 Owners' Group Evaluation: -Comanche Peak Engines FaAA also performed a torsiograph test of the 16-cylinder TDI diesel gen-erator designated Unit 1, DG2 at the Comanche Peak Steam Electric Station (CPSES). As reported in FaAA-85-1-8 (February 1985), the test was conducted at seven speeds between 410 and 450 rpm under no-load conditions, and at loads of 25%, 50%, 75%,100%, and 110% of rated load while the engine was operated at rated speed. FaAA also obtained transient data during two fast starts and two coastdowns.
FaAA's conclusions include:
e The 4th-order critical speed is reached at about 432 rpm. Thus, the first natural frequency of the. torsional system is approximately 28.8 Hz.
This is in good agreement with TDI's calculated value of 28.9 Hz for the CPSES engines.
With the assumption that the shaft is vibrating in its first mode, e
the following values of the amplitude of nominal shear stress were estimated from the measured amplitudes of free-end vibration:
Full Load Overload DEMA 7000 kW 7700 kW allowable Single order.(3.5) 1971 psi 2064 psi 5000 psi Combined response 4544 psi 5011 psi 7000 psi The maximum peak-to-peak free-end response for the two fast starts o
e was found to be 1.64 degrees; for the two coastdowns it was 0.77 degrees. The corresponding amplitudes of nominal shear stress, calculated with the assumption that the shaft is vibrating in its first mode, are 6956 psi and 3270 psi, respectively.
12
4.7.8 PNL's Review The following PNL consultants participated in various aspects of PNL's reviews of the crankshafts for the TDI 16-cylinder, DSRV-4 series engines:
S. H. Bush, Review and Synthesis Associates e
H. Engja, Norwegian Marine Technology Research Institute o
H. M. Hardy, consulting engineer A. J. Henriksen, consulting engineer e
B. J. Kirkwood, C1venant Engineering e
o P. J. Louzecky, Engineered Applications Corporation e the late A. Sarsten, Norwegian Institute of Technology e
T. W. Spaetgens, consulting vibration engineer e
H. Valland, Norwegian Marine Technology Research Institute.
In addition, J. Spanner of PNL's Nondestructive Testing section reviewed non-destructive testing procedures used by members of the Owners' Group to examine the crankshafts.
PNL's review of the crankshafts for the T31 16-cylinder, DSRV-4 engines encompassed the following activities:
reviews of the reports referenced in Sections 4.7.4 through 4.7.7 e
regarding the analyses and torsiograph tests of the crankshafts at four nuclear power plants independent analyses performed by PNL's consultants to provide a e
basis for comparison with the torsional stresses computed for the l
Owners' Group by Failure Analysis Associates e reviews performed in the preparation of PNL technical evaluation reports for the 16-cylinder TDI diesel generators at four nuclear power plants that became candidates for operating licenses: Catawba, Comanche Peak, Grand Gulf, and Perry - The crankshaft was one of the components reviewed.
PNL representatives visited Catawba following completion of each of the two engine tests discussed in Section 4.7.6.1 of this report, and reviewed the post-test disassemblies and inspections of the 1A and IB diesel generators.
13 i
l PNL representatives also reviewed disassemblies and inspections of the unit 1. DG2 engine at Comanche Peak and the unit 1. DG1 engine at Grand Gulf.
The results of several analyses performed by PNL's consultants are summarized in Section 4.7.8.1, and compared with corresponding results of
- investigations by the Owners' Group. 'PNL's comments on the torsiograph tests performed on 16-cylinder engines at four nuclear power plants are discussed in Section 4.7.8.2.
Finally, PNL's conclusions and recommendations on the basis of all of the information reviewed are presented in Section 4.7.8.3.
4.7.8.1 Analyses Performed by PNL's Consultants 4.7.8.1.1 Torsional Stresses Below Rated Speed. Using his program COMHOL2 (Complex HOLzer version 2), Prof. Sarsten analyzed the torsional vibra-
. tion characteristics of the crankshaft for the TDI engines at Grand Gulf, over the range of 100 to 470 rpm. His results are for forced, damped vibrations during steady-state operation at certain speeds and, as such, do not represent true transient conditions during startup and coastdown. Nevertheless, his analysis predicts'the resonant stress amplitudes of the principal orders of vibration and the rpm range over which each order is significant. The harmonic coefficients used in his calculations correspond to the full-load engine brake mean effective pressure (BMEP) of 225 psig. The actual BMEP would probably be
~
less during startup, and would be considerably less during coastdown.
The results of Prof. Sarsten's analysis are summarized in Figures 4.7.1 through 4.7.3.
Figure 4.7.1 shows the amplitude of free-end vibration of the crankshaft as a function of engine rpm. Figure 4.7.2 shows nominal torsional vibratory stresses for single orders of vibration, plotted as a function of engine rpm. Figure 4.7.3 shows nominal torsional vibratory stresses for the sum of 24 orders of vibration, also as a function of engine rpm. The dynamic magnifier cited in the three graphs is a function of the assumed damping, the masses, and the associated natural frequency of the torsional system for first-mode vibration.
14
i l
l' l
l f'~
~
{
I1'I{.I- ] I ' I '-
'Ij~ '{
I-pp [ J
{ ~ I[%
9 4
'f F
}~ j
{
1 Q-d.Q {_.. _.{ p
{~P 7
f
~ _..
1--- __l-._ m.. '.y-g. b. l_
i 9
. _..L..__1._._..{.
a
{-
y
.j.._ - l_1 Q..
e L -
- Li. L.--._.
i f..p
-4 L
1.
!.p._;
. u I_' }__7 _;_.4.
'.AniUTUDE. -.. --. _.
- [ '- H. -.
. -]
.1-4.. i a [_-l-_-. -.. Ld.h d Ld ' '
i MGM13
- i. !
. 1.
.!.8
...L..
..i.
'. 4__; 4.
- ; 9.. i--_. -j -
TDI YM CRAAW3MFT.... '.-.....
- 4. 4 '., 4 4_--i-,.
- _. FRONT (ND AMMITUM_Vs innar JFEED.
_ '. _L _..,
'1..-__
fotCER DeMMD Weseelle,BMP 2HJas!,.
. _. _ SVM 0F N ORMas, DYneMtC ise0Mer/ER MW_,. _.~ '..
stesorsrarenwria.mvmuruoc,
I
l
't-
___d__
_...L.
_J l
i 3
.._. 1 - _ 7. L_.l j
l-
' _._ h ' -.._.
l _J._.d_ _._; _J4#t p//ecise assimrad Aadurand,45"see) i_.
..1 I
7
. [. -.
r*
-;... }-j.
i
.L:
4
. i p 4_;f
.. '.J.
- ; _. }.
i
' p-
- -4
,.g
- p. ;
y -- ; t-
- -+_u_r.
7.; :.i 1
..,. ;4,.~., 4.. [... g _ 'r._ ; _ t..J.
'._j -
l
_;_ _4;i.L-
!.. !. 4 l.
. 4_. '. [J
_..!__i
.._ i. ;.
m.} _ -
L i
.i y E'h+d-
-r _}y__7_ F-
- l--t-- ; -- Y - f - - l-
[ j - I-i---- l
-l-f - l 4 -
- j-- F 4
-h---
4, C-h
-fkLH--- '
..j_.j _;
- p...
_..r 2, - --
J-
._4__. p_._y_..-._ _.. p q.d m
-}-_ r-4
'l t : M- :m -+- I- - Ht---
l-
- ii
-i ; -!
- } i P-r-
-- -i-Fi-
- i. rb--
i lW
,L
?_.h U.g
' j i.]'
-f
. f.a - + }_. h
__i.
.J 4 -l. -
.l..
_. l _._
+.
- g..._'y 4_
+
. ;. _j _ '.1,...
L_!
. _;_,-.___ _y__4_
=l q_
. -..;. _4 -._.. J _.
i,-t -2E 1
k"1L
.i rf-
-F. i--- I--
- f --
- l- - - -i q_ f_jy.._.
q_- -t-
. i.
_..._t f
3
+_
t r
n
.w
.4_.
_. j __
._ p._.__
4 4_
. _ _ _.- 7_ _g__.
- .j
..j __[._. a_,,
l 4
_)
4 _;
TT
.i..A.J -.a
!- _i _.!'+.
t I_ - t._...
1~ ;_;- a
..[2...1_ } _L. L e
r
'i 2._e c.;a.
7.,,_ _ _, ; l r,... i,.,__. t q.
3
.y i
Lii
-D r
-l 7
..r -
t-..
i-
.-- -n
- i.
4 T
_p
..p.
7 g. q.
g
~~
7.
_._} y _j_ __;.g..jj'
.1. h..L _p i
,._ l_.. j._.
4_. _p.,j.-.4 _p 4.:{_
_ _ _}_y_4_}.4_._.:4
- 2. _ _{.,
u.n l-wo
.m
... w e,_. a
. m _.4, oo..
m
. no m,.. _.m-.._m,. m
. m_._;.ue_..m._ _ m m_.-.uos.,o, n
.F i
h I i: 'f1 =
ilt1 1:.
T.... } 1
,g_._
. 1
._1....1_ _L
- _.4
.E
_.x_: ;
.. 2; 2._
. J
.m__
4 4.. 4.1.t. _
. _;_.._:_ i i
i i
i i
-n r wsregna.r.
r FIGURE 4.7.1.
Variation of Free-End Crankshaft Vibration with Engine RPM for Sum of 24 Orders, Predicted with COMHOL2
+.
11
-o
'{
~ j
~ =
M'. y 4' f
}- i{
p p ]
~
%,5 j f' }- % p
' -N
_ p._ l _' i
.y l..-
. _L.
i.__.._.'_i_..e. y_._. _._9 y_
._L 4
__.M.___..__.
- 4i r A
y.:.__.y. j_
.L..J
._,esL
_J
+k.
-.1 1..
. L.
4 L.
_.i _ I' '
._ _ y _. 4
._.._t.
.. _ i 1_ _
_L _
i
_._4
< _d.;._ _i_.L.y _4. { 2_., 4 __
.._l_._.._Q _].
1_._ j _ j.._i. __L. r.. J_i_ L. L. l a. i... LLl_ j _d.;_j.1 j24_4 4
{__ _. l _ l __ i.. i _ 4_a_}
~ ~ ' T01 V-Mdanaf egne O-='M
.._4_'.._. _... l Alasras/searwiryshussee M'd' estue' __'_
_AFooo
..q.. _i L__ '. j T
.e e
a,c,w nea,whein.ugher,#ffe,seo
.. h.
4_ y..
-. - f -.. j g _ t l_... f... '
Aparped, &E*/rf i
L.
I
.3._
_'~1_
_ sooo.
1
__wmaroc mapwne n.47-7
.__ _ L..'_.,_ L _5_._ f,. j__.t.. '
. _... _... a==+ se,* m=se. de vam,* ameweea (4/me1 L...__t__
__ L;.4 r
..;_p_
6 4
i i
r i i a
l l d areh-t
?
i..
f l
l-
' - ~
-f-i
._gper_
_ _. f. -.
h... - ! l!.J
.. [._ I...
l_ 4_.i 1
1 I
Y I
t
?
.]
.S 4 !_
.b
_L i
._4._._
..:.. L._... _ _ j __.
_ _ _ _. l j _. j.
L__ !.
].=
f,_,i
_ Lt. i_
- t..
.?
.g.
J.
a_.i _L. 1_ __._ L f_.;_.
2No' _
I 1 _.i,_j.__{
. 6.... !. :. 4_._.. _..a w
. 44 l
i 4
.. :.r_.
.w _._L
_. L.
i t
i -
e sf *.
uw
< g,ew. -
i=
+
j__4_ 4' 1_..__
.._;...+.
t a
.J,.
_ f__j _. j..
.p7 4_ L _
I'
..h n
m
.A. _ _
I
_.._ a _ p _'__ 7_3_..__ t..;, g..
g y
p._[
- ;y 4__
__;.3 4 4
+ 4_._
.., _.. _ -_,_.. _ _.. J p
!._ _m _.. _ ___j.
4 p
j i
_ ;_..__a 1 E
g
---i
--t -!-t-t-t+- - --t-
- --i-4f- --F -e-7s-3 j p;__ n_ _g2_p'__.2 1
L
+.
--_. g 4
4.p'4 a_.,_w.._. _.._._4:_. _:.
_2_
L g_o_m.
a=
v 1
s u,eT_
=-
- s. !
s.
- s 4
=
9.,
_,4._
e g
t
. g.
_: p y
-I. A 9
q_
y
-.79, u
4
.n m.
.2
+.
u
.}_..
- .y
._p_._h
{
p_
.p.f. _4..._f_
. _L /.T-~~_ _4_.!_ Larmen..L _L. :{ _.i
- p. :..1 l
_}
I 2_ _ Ls
/.
1 p
+
'/_.2__
__1.__
_ _ i L' __.a
._ a.
o X u.a.w 1;_.._
azooo.;.
4
,',','1i
+
.x-M M -+ -* $W Q-}t
~
i\\
-r w -,-f - r
--e ---t --t
,f, rv
[
a 1-i
[. ' i-V lh ' '
4 b
- +f*7--
-/f -j'l-T-
<i+ th+ __,, ~'
"---de L M sF
=,s Lt j- + -t-Pf-t =inj }-j }-lfi'D { ', &b}} if+W{-r' Mpy-}-F
+2 -
h t-Fir-- h 3pe w & t'6
,a _
m
_ mum _. _m _._no-m m.
_ _ sro._.uo.__no _._m_._mw n_,
ap oo[ em
_._,_.m w_.ao,._; as _ 7 6
-6
- . 4t t
-$d.ikik 4 :-
4 4;. v.
.u...
i.._..;_
I ';
}
I
_.2 _.-.,'
. _.: _.;a J
!i.
1
-- 1.
c Ar w n f su,ser s i
s
.n.
FIGURE 4.7.2.
Vibratory Torsional Crankshaf t Stresses for Single Orders, Predicted with COMHOL2 i
l 1
/
0 J
j --
[]i f
]
'E ].];i.-p
^
2
.i._h _j_._L._L 'Q_.._ji3_:.
l{. ' -j.
y.1_.
j_..1..p. _..U Il'r
_.:.p
.I 4
. _L.L _1_ 4 4
._n_.__.A
_p 4_L ig
.._k ___r i.
'4_ 4
. _ _ _._m _:_
._p.. _.
A
_r_
] J' " ~ ~ ~~ P'- ' 't ~
-'*---'-''_I~~-k-'f-
-L'~i~'-
~'i'-
~~---'- ~~ i- '"'
" - ' ' ~
' ' - ~ " - - * -
, T T- ~~ '
. _ TAIF#rmstunuff..
.o
..t
. Ahswear meramery Arumanf sensass d*als shut....
Jarr a/Narmerw.//5 bage Arusped. Armed e/drweare. AmtrasylufAmir adussse
'f.
__ p. p. t _ v.. _y.. p_ p -g.__4___9. _ t. 4 _ i _ '.__ } a. Anan,c muupnl/w- #* Jf.stuale see renamise,_ p +
_A.
5 a
L4_ Aj '
9 4marvie caneterfsy ansmeed serm)
_ pei_
_l
.a
.L' i.
t 8.
MN0 g __ g.
l_g.
.._..._ __.,. _ L.__ _. _ p g
._r _7 _ __
.._4
. 4
_ sow
._ ] L.._._ i._. ),_i E
_j.. (
l
~.I i.. ). _1. _ ;
-..tspe.
_ _l, 4_.
_ _ _. _... - _. _ __ _ '..__ _ s.s a or - -
_t__42q
_._L_
._L l._.
i J.u.}._ 'i L.
i i
i.
I l. !_
. p _j.
J. _ j _ __
, _L...,.. [.. !. 4 _ ! _ __J i.
j t
2,.._&, _.. u. a.. &,
e
._7. p.
. _ j L_. _p.__; p};__ p_,_ p_
.;_j_._.
yny
,p w.
L _f._..
_.p-y '
! _j., [ j r
_p ;
j
}__.p_.._
.._. !_._ p _ ps.* _.,. _. g}
_,4 n._.
gig
.._ _ _ _ q. _1 r -"~ r- ~!'-~~l-t-'
' i- -~hi
'b-
'I - l l
._j_ g_ +
. p.
~,
. p i
[_ J.
_7_ _g
.'1 '
,g :g j g_'j g s'-
_. p.q.
__1.
4 7__g j
_4_.g f_.,.., _
7 e,*.
_2_ ____
_ _.. p. }.
_,_ J i _.
__.. _. '. _.4_y_.
.4_q
_g_ _j.._p
- g g g_
H
~_ 42.J
.( ;gg
_j
_1 g
3 j
4
.._m p
m
} __
7.g.____.
L.._q.4y_p
_gg
- 4. p __
_ _}_
44.g_
_2_
_p_p 4 4
}. __
-l_.
._ j.1
.t.
_p. l 4
_Lt
- p. '4_ A
_._T
_. g. 4 4 L2_
g
..y.{
4 j
l i
_ _ }_j j
._j_ 4. _ _.._4_._p. Q._
_ _ _. _ _ [ _..
__.y... _, _.
i
._ f
.,_ 4 p 4 j ya_
7
.y 4 g
- __.y_.p_4 q. r_
a.
2.m_
u _t _4_
1.{ i _b. _;_.i
- liJ_.
4,.
j_.
1
=
',._ ' _.L _i,.
4_6
. _L. c
_._i
.Q.
L.,
J_
4...
h.._
.ZLii J?.E z
I.p m
i i
- 1
_t_a_. ;. J i
E~
2.
i
_.. ~. _ ~.
~
j 4
. _4... L_.._
_..f 7_.p.a_ 4 4
_.a. _
4
, j g _g..j m_
_ 4 jJ_4_._
_g. 4.__.
_ g_
q_
=---+-
-- -l-b -+-4 i
-+-
-+-
1-
- - --4:-e-FV EZ TF~-B 4 r
4
'--Tt-H?i4H1'i+i?H-1?rh?n Yt-H"rhYrHT-t-i?rt,TrhYrh?rh?rH?ctWM",t.PA%W FIGURE 4.7.3.
Vibratory Torsional Crankshaft Stresses for the Sum of 24 Orders, Predicted with COMHOL2 i
Prof. Sarsten made the following observations on the results: summarized in Figures 4.7.1 through 4.7.3:
e The resonant peaks at speeds corresponding to orders 4.5, 5.5, 6.5, and 8 are not closely spaced, and therefore would not be expected to augment each other during engine startup. The vibratory conditions associated with these resonances are acceptable for engine accelera-tion through this speed range No significant resonances of higher y
modes were detected in the speed range investigated.
e No responses _are shown for the 4th and 12th orders, because these orders vanish under the assumptions used in the analysis. The excitations of these orders from one bank of the V-engine cancel the excitations from the other bank, assuming the cylinders are evenly balanced and disregarding the effect of the articulated connecting rods on piston motion and timing. However, the 4th order in the actual engine is important because it occurs within.-5% of the rated engine speed. Any lack of uniformity in the pressure diagrams of individual cylinders will excite this o'rder, as will the slightly different piston motions and timing associated with the articulated connecting rods.
E PNL notes from FaAA's torsiograph-tests of the engines at Perry and Comanche Peak (Sections 4.7.5 and 4.7.7) that the maximum transient responses measured during startups and coastdowns produce nominal stress amplitudes well below the maximum resonant stress predicted in Prof. Sarsten's analysis (i.e.,
the 8th-order response of 8900' psi at 217 rpm, plotted in Figure 4.7.3).
- Thus, the torsiograph test results confirm that peak stresses under normal transient i
conditions are lower than the peak resonant stress predicted from Prof.
Sarsten's steady-state analysis.
4.7.8.1.2 Torsional Stresses at Rated Speed. Following the untimely death of Prof. Sarsten, Professors Engja and Valland used the computer program COMHOL2 to analyze the torsional system of the TDI engines at Comanche Peak.
They computed crankshaft torsional stresses for operation at full power
~ (225 BMEP) and rated speed, with and without misfiring, for single orders and for the sum of the first 24 orders of vibration. Misfiring was assumed to i
18
occur at cylinder No. 5 left bank (cylinder SL), for consistency with FaAA's investigation 'of misfiring (Section 4.7.5 of this report). To explore the effect of damping, the stresses were computed for two different values of the damping magnification factor, M.
The results are as follows:
Maximum Nominal Torsional Stresses, in psi, at 225 BMEP (7000 kW) and 450 rpm M = 22.4 M = 50 Jtormal Misfi ring Normal Misfiring Tiring Cyl. SL Firing Cyl. SL Single order (3.5) 2265 2910 2281 2982 Combined response 5347 7188 5377 7364 (24 orders)
Free-end amplitude, 0.503 0.888 0.507 0.880 degrees The above results for the combined response under normal firing conditions bracket the maximum nominal shear stress amplitude of 5367 psi from FaAA's dynamic torsional analysis of the crankshaft for the Grand Gulf engines (Sec-tion 4.7.4.1.2 of this report). FaAA's analysis accounted for the same number of orders, and was performed for a similar torsional system operating at the same power and speed. The independent analyses using COMHOL2 and FaAA's pro-gram also agree on the location of the maximum nominal stress, which is pre-dicted to occur between cylinders No. 5 and 6.
These comparisons further reinforce the observation in PNL's review of the crankshafts for 8-cylinder TDI engines (Section 4.6) that the two computer programs yield consistent results.
The stresses predicted with COMHOL2 for normal firing are higher by about 800 psi than the stresses estimated by FaAA from the results of the torsiograph test of a Comanche Peak engine (Section 4.7.7).
For the Grand Gulf engines there is a similar difference between the results of FaAA's torsional analysis (Section 4.7.4.1.2) and the stresses estimated by FaAA from torsiograph data (Section 4.7.4.1.3).
These comparisons suggest that the results produced by the two computer programs are conservative for the torsional systems modeled.
19
For misfiring of cylinder SL, the stresses predicted with COMHOL2 for the Comanche Peak engines are also 280 to 450 psi higher than the stresses esti-mated by FaAA from torsiograph data taken during imbalance tests of the Perry engines (Section 4.7.5).
The difference varies with the damping assumed in the analysis.
4.7.8.2 Review of Torsiograph Test Results The torsiograph tests performed by FaAA on five engines at four nuclear power plants provide actual crankshaft responses for a wide variety of steady-state and transient operating conditions, including loads through 110% rated load, startup and coastdown transients, off-speed operation, and operation with imbalance in cylinder firing. From a review of these tests, PNL's observations are as follows:
The instrumentation used in the tests, and the procedures followed in e
performing the tests, are consistent with established measurement practice. Confidence in the test data is enhanced by the close correspondence between pre-test and post-test calibration measurements, e The nominal shear stresses estimated from the torsiograph data confirm that crankshaft torsional stresses at rated speed and at engine loads up to 110% of rated load are well below DEMA-recommended limits for both single orders and combined orders of vibration.
Maximum nominal stresses estimated from the torsiograph data taken e
during startup and coastdown tests of engines at Perry and Comanche Peak are lower than the peak resonant stress predicted in Prof.
Sarsten's steady-state analysis of resonant stress amplitudes below rated speed.
In addition, the maximum transient responses during coastdown tests were found to be, approximately one-half the maximum transient responses during fast starts, indicating that the lower BMEP during coastdown resulted in a significant reduction in excita-tion. Startup transients measured on the Perry engines were found to have very little sensitivity to initial crankshaft position. These results confirm Prof. Sarsten's observation that the resonant peaks 20
identified in his analysis (Figure 4.7.3) are acceptable for, engine acceleration to rated speed. Further, these results show that the transient responses are acceptable for either the slow starts (with lower BMEP) or the fast starts of the type conducted during the tests.
e Measured values of the 4th-order critical speed are within -55 of rated engine speed for all five engines tested, and range from 429 rpm at Catawba to 436 rpm at Perry. This range reflects the differences in the crankshaft torsional systems of the engines. The test results also show the contribution of the 4th-order response under various operating conditions. For example, the 4th order contributed 0.240 degrees to the 0.340-degree total amplitude of free-end crankshaft vibration measured at Perry during no-load operation of the U101 engine at 434 rpm. FaAA found that the increases in the 4th-order response during imbalance tests of two engines at Perry were not large enough to affect crankshaft adequacy.
- The test results confirm the computed values of the first-mode natural frequencies of the torsional systems of all five engines.
4.7.8.3 Conclusions and Recommendations Subject to implementation of the recommendations discussed later in this section, PNL concurs with the Owners' Group that the crankshafts for the TDI 16-cylinder, OSRV-4 series engines at the Catawba, Comanche Peak, Grand Gulf, and Perry nuclear power stations are adequate for their intended service at loads to full rated load of 7000 kW, and to 110% rated load (7700 kW) for the percentage of operating time at overload that is allowed by the manufacturer.
The torsional analyses and tha torsiograph tests discussed in Sections 4.7.4 through 4.7.8 of this report substantiate the adequacy of the crankshafts of these engines for nuclear ::::rvice. Furthermore, the results of these evalu-ations are supported by satisfactorf performance of the crankshafts in several operational tests (e.g., at Catawba) and by the absence of any evidence of fatigue crack initiation in the crankshafts during post-test inspections.
21
These conclusions may be extended to crankshafts of the same design in TDI engines at.other nuclear installations, provided that the crankshaft mechanical properties and the torsional crankshaft stresses are shown to be similar to those in the installations just mentioned. Recommendations for establishing this similarity are included in this section.
PNL's recommendations are as follows:
To avoid the effects of the 4th-order resonance, the engines should e
not be operated at constant speed below 440 rpm except as necessary to comply with TDI recommendations for initial operation of new engine parts. Such operation should be at speeds that avoid resonant peaks. Steady operation at speeds more than a few rpm below the rated speed of 450 rpm should be avoided, particularly for the engines at Perry. At 436 rpm, the measured value of the 4th-order critical speed of the Perry engines is the closest of the engines tested to the rated operating speed. The effect of the 4th order on crankshaft stresses varies with the extent of imbalance between individual cylinders, and extends to either side of the. critical speed as shown in FaAA's torsiograph test data for each engine tested.
e Because torsional analyses and torsiograph tests confirm that cylin-der imbalance may have a significant effect on crankshaft stresses, appropriate precautions should be taken to prevent sustained engine operation with this condition. Exhaust gas temperatures should be monitored during engine operation to verify that differences between individual cylinder temperatures and the average temperature for all cylinders remain within the range recommended by TDI.
In addition, cylinder firing pressures should be measured no less frequently than the interval recommended by TDI, It would also be prudent to analyze the trends of cylinder pressura and temperature measurements to detect changes that might indicate a need for maintenance of fuel injection equipment. Any abnormalities should be corrected expeditiously.
l 22
- During each 5-year engine disassembly and inspection, the oil holes and fillets of the three main bearing journals (Nos. 4, 6, aid 8) subject to the highest torsional stresses should be ' examined with' fluorescent liquid penetrant and, as appropriate, with eddy current.
The oil Mles and fillets in at least three of the crankpin journals No. 3 through 8 also should be examined in this manner. These inspections Ore recomended to verify the continued absence of fatigue cracks. PNL notes the comment of FaAA (P. Johnstion) in a meeting with ti.e TDI Diesel Generator Owners' Group (transcript dated October 22,1984) that the fillet is the location of highest stress in a crankpin of these V-engine crankshafts.
Further, PNL notes FaAA's observation in report FaAA-84-3-16 (May 1984) that the oil holes in the main journals represent a more critical stress concen-tration in torsion than do the c.rankpin fillets. Thus, the oil holes in the main journals are the areas of highest stress. Nevertheless, PNL views the fillets of the most highly-stressed main journals and the fillets and oil holes of the crankpins as critical surfaces that also warrant surveillance.
If cumulative results for several engines show that-these examinations reveal nothing of significance, the scope and/or frequency'of the examinations could be reconsidered.
I To verify that crankshaft alignment remains within manufacturer's e
recomendations, crankshaft deflection should be measured under both
" hot" and " cold" conditions at each refueling outage (as planned at the four nuclear power plants with 16-cylinder engines addressed in this review). PNL's recommendations on the conduct of these exami-nations have been documented previously in plant-specific reports (e.g., PNL-5211, August 1984).
It may become appropriate to recon-sider the frequency of these deflection measurements if substantial evidence develops that no significant changes are found from outage to outage. Any new interval between checks should, of course, be no less than the frequency recommended by the manufacturer.
e To establish whether or not the conclusions reached in PNL's review may be extended to crankshafts of the same design in TDI 16-cylinder 23
L engines at other nuclear power plants, the following action should'be taken for each installation. This action is in addition to'any tests performed by the manufacturer to confirm proper engine assembly, balance, and timing.
A torsiograph test should be performed on at least one engine of a multiple-engine installation. One test will be sufficient if the torsional systems of the engines are of the same design.
Each engine should be tes~ted if the torsional systems differ (e.g., if the generators are of different designs). The test should include variable-speed operation over the entire speed range from shutdown to +5% of rated speed, to determine the location of any potentially serious resonances (e.g., the 4th order).
In addition, the test should. include variaale-load operation to determine crankshaft responses up to at least full rated load. The test results and torsional stresses calculated from these results should be compared with similar data for other 16-cylinder engines at nuclear power plants. Depending on the outcome of these comparisons, further testing and analysis may be necessary. to establish crankshaft adequacy.
The materials certification' reports on the crankshafts should be reviewed to verify that the crankshaft mechanical properties (e.g., ultimate tensile strength) are within design specifications.
Preservice crankshaft inspections should include fluorescent liquid penetrant and, as appropriate, eddy-current examinations of the oil holes and fillets of the three main bearing journals (Nos. 4, 6, and 8) subject to the highest torsional stresses.
These inspections are recommended to verify the absence of rejectable machining irregularities and crack-like indications.
24
r=
4.7.9 References Diesel Engine Manufacturers Association (DEMA).
1972. Standard Practices for low and Medium Speed Stationary Diesel and Gas Engines. 6th ed.
Failure Analysis Associates (FaAA). May 22, 1984 Evaluation of Emergency Diesel Generator Crankshafts at Shoreham and Grand Gulf Nuclear Power Stations. FaAA-84-3-16, Palo Alto, California.
Failure Analysis Associates (FaAA). May 29,1984. Torsiograph Test of Emergency Diesel Generator IA at Catawba Nuclear Power Station.
FaAA-84-5-23, Palo Alto, California.
Failure Analysis Associates (FaAA). July 24,1984. Torsiograph Test of Emergency Diesel Generator Div. I at Grand Gulf Nuclear Generating Station.
FaAA-84-7-10,.Palo Alto, California.
Failure Analysis Associates (FaAA). February 1985. Torsiograph Test of Emergency Diesel Generator 1DG2 at Comanche Peak Steam Electric Station.
FaAA-85-1-8, Palo Alto, California.
Failure Analysis Associates (FaAA). May 1985.
Torsiograph Tests of Emergency Diesel Generators, Divisions 1 and 2, at Perry Nuclear Power Plant--Unit 1.
FaAA-85-4-1, Palo Alto, California.
Failure Analysis Associates '(FaAA). June 1985.
Addendum to Torsiograph Tests of Emergency Diesel Generators, Divisions 1 and 2, at Perry Nuclear Power Plant--Unit 1 on Cylinder Imbalance. Fa AA-PA-R-85-06-ll, Palo Alto, California.
O Grand Gulf Nuclear Station. February 1984 Com)rehensive Report on Standby Diesel Generators - Significant Activities to Enhance and Verify Reliability. Forwarded to NRC (H. R. Denton) as Enclosure 2 of Mississippi Power & Light Company (J. P. McGaughy) letter dated February 20, 1984 I t Pacific Northwest Laboratory. August 1984. Review and Evaluation of Transamerica Delaval, Inc., Diesel Engine Reliability and Operability -
Catawba Nuclear Station Unit 1.
PNL-5211, Richland, Washington.
25
.