ML13330B253
| ML13330B253 | |
| Person / Time | |
|---|---|
| Site: | San Onofre  |
| Issue date: | 02/09/1987 |
| From: | IMO DELAVAL, INC. |
| To: | |
| Shared Package | |
| ML13330B252 | List: |
| References | |
| NUDOCS 8709280114 | |
| Download: ML13330B253 (50) | |
Text
Imo Delaval Inc.
Enterprise Engine Division Enter ris Hay W a m 85h An nue Uaklandl. (A 94621 41-5Y7-741A)
ANALYSIS OF VARIOUS MODIFICATIONS TO THE SAN ONOFRE DIESEL GENERATORS TO REDUCE THE POTENTIAL OF FATIGUE CRACKING IN CRANKSHAFT OIL HOLES DUE TO STARTUP AND COASTDOWN TRANSIENT STRESSES PREPARED BY IMO DELAVAL INC.
ENTERPRISE ENGINE DIVISION FOR SOUTHERN CALIFORNIA EDISON COMPANY SAN ONOFRE NUCLEAR GENERATING STATION UNIT 1 EMERGENCY DIESEL GENERATORS FEBRUARY 9, 1987 PDR ADOCK PDR P
IM0
- 1.
EXECUTIVE
SUMMARY
During an inspection in mid-1984, indications of cracks were discovered in the crankshaft oil transfer holes of the Delaval supplied San Onofre Emergency Diesel Gener ators.
Subsequent analysis by the owner's consultant attributed these cracks to fatigue induced by transient starting and coast-down stresses.
The crankshafts have been repaired and the units have since been returned to service.
The NRC's consultant, concerned about the planned continued operation of the units, recommended performing periodic inspections to confirm crankshaft suitability.
They also recommended investigating several potential alternative modifications to the diesel generators designed to reduce the transient stresses.
These recommendations include the following:
Altered mass elastic system Installation of pendulum damper(s)
Installation of a viscous damper Altered firing order Of these, only the installation of a viscous torsional damper appears technically feasible.
However, this option involves a moderate degree of risk.
It is also substantially more costly than other alternatives.
Continued operation of the system as is, with monthly slow start surveillance tests and prepositioned fast starting of the units only when fast starts are required is a second option.
Operation in this manner will significantly delay further crack initiation.
No cracking should occur for the life of the plant.
A third option is the installation of high strength replace ment crankshafts which would be far less susceptible to the fatigue phenomena associated with startup and coastdown transient stresses.
Page One
IM@
II.
INTRODUCTION The crankshaft oil transfer holes of the two San Onofre Emergency Diesel Generators (S/N 75041/42) were inspected from July through October, 1984 as a part of the DRQR effort (Ref.
1).
Eddy current and dye penetrant inspections revealed indications of cracks in the oil holes in journals 8, 9 and 10 of DG 1 and in the oil hole in journal no. 9 of DG 2 (Ref.
2).
After further analysis and testing, a consultant retained by the owner concluded those indications were fatigue cracks induced by transient torsional stresses exceeding the endurance limit and that these stresses occur only during startup and coastdown (Ref. 3).
Based on these findings, the NRC's consultant recommended the investigation of various potential alternative modifications to the diesel generator's torsional system to reduce crankshaft stresses (Ref. 4).
These include altering the mass elastic system, installation of pendulum or viscous dampers and alteration of the firing order.
In this report, we first establish the stress levels and predicted life of the current system.
Then we review the technical feasibility of the various modifications.
A detailed estimate of cost and time to implement the three most feasible modifications follows.
Page Two
IM)
Ill.
METHODS OF ENHANCING RESISTANCE TO FATIGUE CRACKING INDUCED BY-TRANSIENT STRESSES A.
Overview of Engine Torsional System Dynamics An engine crankshaft translates the reciprocating motion of power pistons into the rotary motion required by the prime mover.
The required crankthrows, flywheel, counter weights, et cetera, result in a non-uniform distribution of inertias and stiffnesses creating a number of torsional natural frequencies of this mass-elastic system.
When excitation is applied at or near these frequencies, large oscillatory torsional stresses may occur depending on the magnitude of the excitation (firing pressure and inertia) and the damping (frictional dissipation of vibratory energy) of the torsional system.
Such oscillating torsional stresses may exceed the endurance limit of the crankshaft material, inducing fatigue cracks in areas of stress concentration such as the fillets or oil holes.
Methods of enhancing resistance to such fatigue fall into three basic categories:
Modify the mass-elastic system to move the critical natural frequencies out of the engine speed range and/or increase system damping to dissipate vibrational energy.
Viscous dampers, pendulum dampers and changes in the rotating masses (inertias) as suggested by the NRC's consultant fall into this category.
Alter the excitation (in particular, firing pressure) to reduce or eliminate those components (called har monics) occurring at the natural frequencies.
Changes in the engine's firing order, disabling cylinders and modifying starting fall into this category.
Enhance the crankshaft's resistance to high cycle fatigue cracking.
Increasing material fatigue strength falls into this category.
A discussion of the technical feasibility and associated risks of each of these modifications for the San Onofre Diesel Generators will follow, after first reviewing the original system design history and current operating conditions.
Page Three
INS B.
The Current Torsional System
- 1.
Design History - The torsional system of the subject San Onofre engines was originally designed using domestic industry standard analytic techniques to meet contractual requirements (Ref. 5).
Subsequent experience indicates that a unique phemonena occurs in the DSRV-20-4 in nuclear standby service.
The fast starts, coupled with the system's unique torsional characteristics, result in starting transient crankshaft stresses never previously encountered in engines manufactured by Delaval.
A review of the torsional system using recent, more advanced techniques has been conducted and is described in the following section.
- 2.
Calculation of Nominal Transient Stresses - Recently, Delaval has developed and benchmarked several new computer simulations for calculating actual "combined" stress in each crankmember using both frequency domain (modal superposition and dynamic matrix) and direct integration techniques.
By maintaining relative phase information for each individual order, these techniques can predict the total combined response of each crankmember in contrast to the more traditional techniques (Ref. 6) used in the generation of Figure 1.
These latter techniques predicted the response of each crankmember*
to each individual order but did not retain the phase information needed to predict the combined response.
- Note:
Nominal torsional stresses are usually calculated for the smallest diameter crankmember, typically the crankpin.
In crankshafts with equal diameter crankpin and journal, as used in the engines in question, nominal stresses in each are equal.
Traditionally, the crankpin fillet was still considered the point of highest stress, hence the reference to crankpin stresses in Figure 1.
As will be seen in the following section, more recent investigation reveals the journal oil hole constitutes the point of maximum torsional stress for the units in question.
Therefore, the term "crankmember" encompassing both the crankpin and journal nominal stress will be used in this report in lieu of "crankpin" as used in Figure 1.
Page Four
IM0 We have used these simulations to conduct quasi-steady state analysis of fast start induced stress for the San Onofre diesel generators throughout the speed range of 100-475 rpm.
Calculated stresses at critical speeds proved very sensitive to crankmember relative damping.
Therefore, the proper damping value was obtained iteratively based on reported free end measurements (Ref. 3).
The results are displayed in Figure 2.
The nominal alternating torsional stresses reach a maximum of 20.4 ksi in member 8-9.*
Essentially equal stresses occur in crank members 9-10 and 10-11.
Of course, this quasi-steady state analysis represents a worst case condition assuming fully developed (steady state) vibratory response.
The experimental free-end data indicates this does not always occur but, rather, depends upon crankshaft position at the time of start cycle initiation.
Similar quasi-steady state analysis of slow start and coast down events indicates alternating torsional stresses peak at 11.1 ksi for these events at the same locations (9-10 and 10-11).
This represents over a 40% reduction in stress from that induced by a fast start.
As noted, experimental data indicates crankshaft transient torsional response during a fast start varies with initial crankshaft position at the time of start initiation.
Calculations performed by others (Ref. 3) based on this data indicate fast starting alternating stresses vary from a low of 9.17 ksi to stress levels as high as that induced by fully developed vibration.
These calculations indicate 30 - 40% of the possible starting positions will generate such fully developed level vibratory stresses with 20-40 high stress cycles occurring per start-up/coast down event.
- 3.
Oil Hole Stress Concentration - As described by the Owners' consultant (Ref.
2), the oil hole cracks were oriented at 45 degrees to the crankshaft centerline in perpendicular planes on opposite ends of the holes.
Such cracking is indicative of fatigue induced by reversed torsional shear stresses.
- Note:
x-y designates the two crank members adjacent to a journal.
For example, journal 10 is between cranks 9 and 10.
The corresponding stresses associated with journal 10 are denoted 9-10.
Page Five
IM' The alternating shear stress in this location ( t ) is obtained by multiplying the alternating torsional (shear) stress calculated above ( T nom) by an oil hole stress concentration factor K:
= KTnom In order to compare this shear stress with reversed bending strength endurance limit data (see below), an equivalent stress (Oeq ) is calculated using the Von Mises criteria.
For pure torsion, this becomes:
eq Combining these equations yields:
Deq TK Tnom Many researches combine the v and the K into a new stress concentration factor Kt as follows:
eq t nom A standard reference for stress concentration factors (Ref.
- 7) indicates a stress concentration of 3.65 is appropriate for holes ranging from 15/16 in. to 1-1/2 in.
diameter in a 13 inch diameter shaft.
(The oil holes in the SCE shafts range from 15/16 inch to 1-1/2 inches diameter).
A standard text on crankshaft design (Ref.
- 8) indicates Kt ranges from 1.7 to 3.5 for typical oil hole treatments while the Bicera Handbook (Ref.
- 9) indicates values of 1.6 to 1.9 are appropriate.
Ker Wilson notes that most crankshaft failures occur in fillets due to bending with only 7% occurring in oil holes (presumably in torsion).
In a review of failures experienced since World War II, Archer (Ref.
- 10) describes one series of torsional failures due to resonance in a crankshaft of equal crankpin and journal diameter.
Some times the failures began in the journal fillets, other times in the journal oil holes, indicating the two had essentially equal stress concentrations.
Page Six
IMS The preceding literature review suggests that journal oil holes -become a critical stress concentration when the journals and crankpins are of equal diameter (a rare occurrence in other than articulated V-engines).
In this instance, the oil hole stress concentration appears to equal or slightly exceed that of the journal and crankpin fillets.
Archer suggests fillet Kt = 2.8 typically and Delaval's strain gage work confirms journal fillet Kt is less than 3.0. Therefore, it appears reasonable to assume that oil hole stress concentration is not less than 3.0 and no more than 20 per cent greater than the fillet stress concentration or 3.6.
Based on these stress concentration values and the transient nominal alternating stresses noted in Section B.2 above, the range of alternating torsional stresses are predicted conservatively as follows:
Highest Stress Fast Start:
61-73 ksi Lowest Stress Fast Start:
28-33 ksi Slow Start/Coast Down:
33-40 ksi
- 4.
Crankshaft Endurance Strength - Based on industry design standards and field experience, high cycle fatigue design criteria appears to be the most reliable for evaluating crankshaft life.
(Life is defined as time to initiation of a detectable crack.)
Basically, this technique consists of the generation of a stress-life (S-N) relationship based on ultimate tensile strength and endurance strength.
For a given stress, the number of reversed cycles to failure is then calculated.
Various researchers have developed S-N relationships based on semi-log (i.e.,
S vs. log N) or Log-Log (i.e.,
log S vs. log N) plots of experimental data (References 12 - 15).
In the first case, a straight line is drawn from the tensile strength to the endurance strength on semilog paper, yielding an equation of the form:
Page Seven
IMO Sn = A log N + B with Sn = Fatigue strength at N cycles N = Number of reversed cycles A,B = Constants determined by the ultimate strength (Su) and endurance strength (Se).
In the other case, a relationship of the form S = 10c Nb is used in the so-called high-cycle fatigue region (i.e.,
between 103 and 106 cycles).
In this case, b and c are functions of the endurance strength and the fatigue strength at 10 3 cycles.
Tensile strength (Su) for the shafts in question were obtained from the certified material test reports, which indicated an average strength of 95 ksi.
Endurance strength (Se) is estimated by the equation:
Se = rSu where r is defined as the endurance ratio, an experimental value.
Fatigue testing conducted on crankshafts of similar diameter, chemistry and mechanical properties to the San Onofre crankshafts (Ref.
- 16) indicates a fatigue ratio of 0.37 is conservative and appropriate.
This results in an endurance strength of 35 ksi.
Regretably, the endurance strength at 103 cycles is not so clearly defined.
Shigley (Ref.
- 12) recommends calculating this value from the tensile strength (similar to the endurance limit) using a fatigue ratio of 0.80 to 0.65.
Presumably the lower, more conservative value includes some size, surface and notch effects along with data scatter.
Using both the semi-log and log-log (with 0.65 and 0.80 fatigue ratios) relationships noted above, plots (and equations) of S vs. N were generated (see Figure 3).
The semi-log plot fell between the two log-log plots following the 0.65 fatigue ratio plot below 104 cycles and the 0.80 fatigue ratio plot above 104 cycles.
Page Eight
IM4P Assuming 40 stress reversals per start or coast down event and the maximum stress concentration (3.6), the following lives were predicted for the two bounding log-log plots:
Fatigue Ratio at 103 0.80 0.65 Highest Stress Start 40 Starts 4 Starts Lowest Stress Start Infinite Infinite Slow Start/Coast Down 5000 Starts 3000 Starts or Stops or Stops At the time cracks were detected at San Onofre, DG 1 had accumulated 740 starts and DG 2 had accumulated 450 (Ref.3).
Cracks had just initiated in one oil hole of the latter unit.
As noted, 30 - 40% of the possible starting positions generate high stresses, so service experience indicates cracks initiate after approximately one hundred highest stress starts.
This empirical data clearly demonstrates the lower log-log plot (0.65 ratio) is far too conservative for this application, accounting for size effects, surface effects and notch effects which are not completely operative in this regime.
In addition, the built-in statistical reductions would not be applicable to highly polished, carefully inspected oil holes.
Even the upper log-log curve (corresponding to a 0.80 fatigue ratio) underpredicts life, by a factor of 2.5.
This indicates either the actual stress concentration is somewhat lower (about 3.3) or there are fewer reversed cycles per start (about 20).
If life is recalculated based on either a lower Kt or fewer stress reversals, the effect is the same, so the following lives to crack initiation based on empirical data can be reliably predicted for the crankshafts in their current condition:
Highest Stress Fast Starts Approximately 100 Starts Lowest Stress Fast Start Infinite Slow Start/Coast Down Over 10,000 Starts or Stops Having established this baseline for the current crank shaft, a review of the various potential modifications Page Nine
- O follows.
However, it is immediately clear that elimin-ating fast starts alone will sufficiently reduce transient stresses to permit continued use of the current crankshafts.
C.
Modified Mass Elastic System
- 1. Altering Rotating Masses - The resonant frequencies of a mass elastic system are determined by the system's stiffness and inertia.
For the simplest case of a torsional pendulum (see Figure 4a), the system can be modeled as a massless torsional spring and a rigid inertia (see Figure 4b).
The natural frequency, W, then is:
W= vK/
Where:
K Stiffness J = Inertia While calculation of the natural frequencies of an engine's torsional system is far more complex (there are n-1 natural frequencies for n inertias), the basic relationship between natural frequencies, stiffnesses and inertias still hold.
Once the system's natural frequencies are established, the excitation can be reviewed to determine if significant input occurs at or near these critical frequencies.
For engines, the excitation sources (i.e., combustion and inertia) are usually broken up into components or "orders" referenced to one complete revolution.
Events occurring once every other revolution are called 1/2 orders, once per revolution first order, et cetera.
Critical speeds" (N ) occur when a given order's energy at that speed is great enough to excite the natural frequency, i.e.,
N = W/x where x is the "high energy" order number.
(As shown in Figure 1, for example, the first natural frequency of the system in question occurs at 1,194 vib/min.
A V-20 has a relatively high fifth order inertial induced excitation Page Ten
IM0 so a critical speed occurs at 1194/5 = 238 rpm).
Typic-ally, only the first few natural frequencies or "modes" of vibration are considered, since there is usually insufficient energy to excite the higher natural frequencies.
When critical speeds do occur, they can be altered by changing the energy content of a given order (see below) or by changing the natural frequencies (changing either the system stiffnesses or inertias).
For engines in general, the system stiffness (i.e.,
crankpin, crankweb and journal geometry) is essentially fixed by bearing load and bending strength considerations; however, some adjustments can be made to the inertias, specifically the counterweights and flywheel.
Typically, such "tuning" can raise or lower critical speed 10%, or, in the case of the DSRV-20-4, up to 40 rpm.
However, the three criticals (Figure 1) of the torsional system in question occur at very low speeds.
Physical geometric considerations prohibit the very large change in inertias required to raise the criticals out of the speed range.
Therefore, significant reduction of stresses by these means is not possible.
- 2.
Pendulum Dampers - Originally introduced in the 1930's for radial aircraft engines (Ref.
17), pendulum dampers consist of one or more masses mounted eccentrically on a carrier disk.
When the disk rotates, the pendulum(s) generates a resistive torque proportional to the square of the disk's angular velocity.
By properly selecting the pendulum's location relative to the disk, a resistive torque can be generated for a given speed (harmonic) completely eliminating the critical vibratory torque.
Pendulum dampers essentially alter the torsional system's stiffness and are therefore actually "detuners", not "dampers".
The pendulum is mounted on the crankshaft free-end or on counterweights of selected crank-throws (see Figure 5).
As a disk is required to detune each critical harmonic, installation on in-line and V-type engines with multiple critical speeds is complex.
In addition, problems with wear and rattle at low speeds have historically limited the application of pendulums to higher speed engines.
Page Eleven
IM@0 With the successful development of viscous dampers (which damp all harmonics) in the late 1940's, pendulum dampers fell into disuse because of their greater cost, complexity of installation and unreliability.
Pendulum dampers have, in fact, never been utilized on Enterprise engines or to any significant extent on similar engines for these reasons.
While theoretically intriguing, the installation complexity of a pendulum damper, coupled with questions of reliability and the lack of industry experience and the greater cost, make it a less suitable alternative compared to viscous dampers; therefore they do not merit further consideration.
- 3.
Viscous Dampers - Viscous dampers, the torsional analog of linear dashpots (i.e.,
shock absorbers), consist of a large inertial mass rotating in a viscous fluid and housed in a casing coupled to the crankshaft (see Figure 6).
Relative motion between the mass and its housing shears the viscous fluid and generates a resistive torque proportional to the relative velocity.
To maximize effectiveness, such dampers are mounted at "anti-nodes" (i.e., the points of maximum deflection), most often at the crankshaft front end.
As the installation of a damper increases the torsional system's inertia, it usually lowers the natural frequencies 10 -
15%.
In some circumstances this moves a critical down into the rated speed range.
Installing a spring between the inertial mass and the crankshaft adds a second "mode" of vibration with one critical below the original natural frequency and one above, hence effectively splitting the original critical.
Such dampers are usually referred to as "tuned", in contrast to those lacking such a spring.
In addition to splitting natural frequencies, tuning usually reduces the damping requirements of a given torsional system thereby decreasing damper heat load and subsequent parasitic losses.
Damper selection proceeds as follows:
First, the inertia of the inertial mass is selected, typically the largest that fits in the given envelope.
- Then, stiffness (if the damper is tuned) is optimized to obtain best tuning (i.e., equal first and second mode Page Twelve
IM vibratory amplitudes); finally, damping is optimized to obtain the minimum overall front end displacememts throughout the operating speed range.
Delaval has successfully applied untuned dampers manufactured by Houdaille to hundreds of diesel and spark ignited variable speed engines.
These dampers, in diameters up to 43 inches, are filled with silicon fluid and sealed.
The method of sizing is somewhat crude and the manufacturer does not appear to have the damper parameters required for conducting a precise analysis of a crankshaft fitted with their damper.
Delaval began utilizing tuned dampers manufactured by Geislinger in 1978, due to their greater size and damping range.
In addition, the manufacturer has clearly quantified the damper parameters required for precise modeling techniques.
We have most recently modeled, applied and tested Geislinger dampers success fully on various Delaval engines.
This damper uses oil supplied from the engine lubricating system as the viscous fluid.
Damping is obtained by forcing the oil between chambers of fixed clearances.
Leaf springs between the primary to the secondary form the elastic element (see Figure 7).
Varying chamber clearance controls damp ing and varying spring size changes the stiffness.
Preliminary quasi-steady state calculations were performed using the Dynamic Matrix Technique for San Onofre's DSRV-20-4 engines.
While this technique, like modal superposition, utilizes a frequency domain transformation, it offers two advantages:
Iterates to the lowest natural frequency first (which is of greatest interest), minimizing numerical errors.
Can directly model a viscous front end damper.
As noted, we have used this technique to simulate torsional dynamics of Delaval engines, obtaining excellent agreement with experimental measurements and time-domain direct integration calculations.
Page Thirteen
IM)
The calculated nominal stresses for the San Onofre engines fitted with such a damper are shown in Figures 8a and 8b.
As noted above, the tuned damper splits the original natural frequencies, creating a new first mode at 767 vib/min and a second at 1396 vib/min.
Nominal stresses now peak at 170 rpm in the 8.0 inch diameter shaft between the gearset and No. 1 cylinder (see Figure 8a).
The keyway in this shaft creates a stress concentration (Kt = 3.0), resulting in a combined equivalent stress of 20,900 psi, well below the endurance limit of 35 ksi.
The maximum oil hole stress of 16.6 ksi (based on Kt = 3.3) occurs between cylinders 8 - 9 at 290 rpm (Figure 8b).
Again, this stress level is well below the endurance limit.
Based on the quasi-steady state calculation, it appears the viscous damper in question will reduce torsional oscillatory stresses sufficiently to prevent crack initiation for any starting event.
Use of such a damper introduces two practical problems.
First, although the crankshaft in the San Onofre units is designed to accept a shrunk-on torsional damper, the tapered shrink collar was not designed to react the damper torque encountered when operating with maximum fuel at a critical speed (i.e., fast start conditions).
As a result, some alternate form of fastening would have to be used to accomodate this large torque.
This introduces moderate risk, as we have no experience with such an assembly.
Second, the diesel generators in question are fitted with front end pumps and a front end auxiliary skid and will not accept a damper without removal of the pumps and pipes interfering with the damper.
While utilization of an extension shaft and outboard support would preserve most or all of the pumps, it would necessitate removal and repiping of some auxiliaries.
The estimated costs and time for such rework are detailed in Section IV-A.
Finally, installation of such dampers on the units in question involves some additional risk.
To our knowledge, viscous dampers have only been used to attenuate quasi steady state torsional vibration.
While in theory such devices could attenuate the first high displacement vibration cycle during start up acceleration, we have some doubt that the actual hardware will possess rapid Page Fourteen
IM1 enough time response.
If the damper fails to respond quickly enough, quasi-steady state calculations indicate nominal stresses could exceed 40 ksi (see Figure 8c).
The resultant oil hole stress would exceed the ultimate strength, initiating a crack in one cycle.
Experimental transient testing of an actual damper could resolve this question. However, even if satisfactory transient response is demonstrated, one start with an inactive damper (due to loss of fluid, loss of pressure, et cetera) would permanently damage the crankshaft.
D.
Altered Excitation
- 1.
Retiming - The timing, or relative firing order of the cylinders of an engine is selected in order to optimize trade-offs in engine balance, turbocharging and torsional vibration.
On the order of 200,000 possible firing orders exist for a 20-cylinder V-type engine.
Of these, 192, representing twelve different crank arrangements, give balanced external moments.
Regardless of the crank arrangement, the relatively long crankshaft in a 20 cylinder engine yields low natural frequencies.
For the engines in question the first mode frequency of 1194 vib/min places the 2-1/2 through the tenth order harmonics in the operating speed range.
No matter which firing order is chosen, the fifth and tenth order harmonics will have essentially the same magnitudes, since they are governed mostly by bank angle and inertial terms.
However, changing the firing orders will alter the relative level of other harmonics, especially the 1/2 orders.
As shown in Figures 1 and 2, the current firing order gives rather high 4-1/2 and 5-1/2 harmonics, creating critical speeds in the 200-270 rpm range.
Other possible firing orders for the given crankthrow geometry (i.e.,
1-8-5-7-2-10-3-6-4-9) eliminate 4-1/2 and 5-1/2 order excitation at the expense of greatly increased 3-1/2 and 2-1/2 orders (Figure 9).
This creates a critical at 341 rpm, generating higher stresses than the original 217 rpm (5-1/2 order) critical.
In addition, the greatly increased 2-1/2 order excitation (478 rpm) generates stresses exceeding the endurance limit at rated speed.
Page Fifteen
IM Such firing orders are clearly unacceptable.
As noted, the fifth order at 239 rpm remains unchanged.
The existing DSRV-20-4 firing order offers the best rated speed performance for this or any other crankthrow orientation.
We know of no other that will reduce low speed criticals without creating unacceptable rated speed criticals.
- 2.
Disabling - As the units in question were originally designed to accept a very large, single step load, they have a lower full power rating than most R-4 Series engines.
An alternative method of reducing stress is to reduce the relative excitation of various orders by disabling some cylinders.
In particular, disabling right and left cylinders 1 and 10 reduces peak stresses 12% at 239 rpm (fifth order) and greatly reduces the 5-1/2 and 4-1/2 order contributions (Figure 10).
However, the 3-1/2, 3 and 2-1/2 order contributions increase significantly, the latter generating stresses exceeding the endurance limit above 460 rpm, an unsatisfactory condition.
Disabling banks 1, 2, 9 and 10 reduces the fifth order stresses 25%, the 4-1/2 even more, and nearly completely eliminates the 5-1/2 (Figure 11).
The 6 and 3-1/2, however, increase, the latter to levels nearly equal to the fifth, which still exceeds the endurance limit.
The rated speed nominal stresses increase from 3040 psi to 4740 psi.
However, the oil hole stress of 15.6 ksi is still well below endurance limit.
Starting acceleration time would increase somewhat beyond the required ten seconds.
- 3.
Modified Start - Rather than reduce excitation of critical harmonics by disabling individual cylinders, critical order harmonics can also be reduced by reducing the excitation of all harmonics across the board; that is, reducing all peak firing pressures (Figure 12).
As noted, low fuel slow starts yielded 45% lower stresses (from 20.4 ksi to 11.1 ksi) than worst case full fuel fast starts.
As fast starting is not required for surveillance tests, consistent slow starts can be obtained by modifying the control system.
The repaired crankshafts should withstand on the order of 5,000 (slow start) surveillance tests without initiating a crack.
Assuming one surveillance test per month for the remaining life of the plant Page Sixteen
IM (fifteen years) gives an expected 180 such tests, or only 4 per cent of the expected slow start/coastdown crankshaft life.
As noted, prepositioning the flywheel will induce starting stresses below the endurance limit.
Work by others (Ref.
3), however, does indicate deviations in starting duration exceeding +/- 0.5 seconds can induce large stresses.
A review of the 300-start factory test data (Ref.
- 18) (Table 1) reveals starting duration deviates
-0.2/+0.4 seconds, about a 6.3 second mean with a standard deviation of 0.1 second.
This confirms the consistency of fast start duration.
Moreover, as seen in Figures 4-6 of Ref. 3, it may well prove possible to select a crankshaft starting position yielding satisfactorily low starting stresses for expected deviations in starting duration.
Finally, as noted earlier, approximately one hundred high stress starts are required to initiate a crack.
Fast test starts are expected to occur only once every eighteen months during SIS/LOP (Safety Injection Signal/Loss of Offsite Power) testing, for a total of ten more such starts in the expected plant life.
Even if all ten anticipated fast starts do unintentionally induce high stresses, it is very unlikely that this low number of starts will initiate cracking.
Considering the satisfactory stress reduction and ease of implementation, this slow start surveillance testing scheme seems the most practical of the altered excitation alternatives.
E.
Higher Strength Crankshaft All of the methods of enhancing crankshaft life discussed above focused on reducing stresses in order to retain the original crankshafts.
An equally effective way to enhance life against high cycle fatigue is to raise the material's fatigue strength.
Traditionally, Delaval has utilized low carbon steel crankshafts for the R-4 Series engines because of this material's damage tolerance in the unlikely event of a bearing failure. However, Delaval does have experience with higher strength, low-alloy steel materials.
Such materials Page Seventeen
IM have tensile strengths on the order of 125 ksi and fatigue ratios of 0.41 (Ref.
16 and 19),
resulting in fatigue strengths exceeding 50 ksi.
If such higher strength crankshafts were retrofitted to the San Onofre engines, they would provide infinite life (with a margin of 1.3 on stress) for slow start coast-down events and tolerate more than 1600 high-stress fast starts.
No changes in physical crankshaft dimensions would be required, so no other hardware changes or certification testing of the diesel generators would be required.
.F.
Discussion of Alternatives The alternatives noted above fall into three basic categories:
- 1.
Those offering no improvement or involving high risk and therefore not meriting further consideration.
- 2.
Those completely eliminating cracking potential.
- 3.
Those which will satisfactorily eliminate cracking potential.
Category 1 alternatives include modification of the mass elastic system and engine retiming, neither of which will offer any improvement.
Installation of pendulum dampers also falls into this category because of the risks associated with technical complexity, questionable reliability and lack of industry experience.
Installation of a viscous damper falls into category 2, although it is not risk-free and experimental testing should be conducted to confirm such dampers possess suitable transient response.
If they are not fully functional in the first ten revolutions of the engine, even higher stresses could result in permanent damage to the shaft.
In addition, actual installation of the hardware is complex and expensive (see Section IV.A).
Misassembly may occur, greatly increasing installation time and cost.
Cylinder disabling, modified starting and installation of a higher strength crankshaft all comprise category 3 alter natives. The first is least desirable, as it requires extensive hardware changes and may create excessive vibration due to the unbalanced forces and moments created, yet still yields higher nominal alternating stresses than the modified Page Eighteen
0 IM*
start (15.0 ksi vs. 11.1 ksi).
The modified start significantly enhances crankshaft life and requires simple, limited modifica tions to the engine starting controls.
The installation of a higher strength crankshaft is technically simple, requiring no modifications or special tests and incurs no significant risk.
Therefore, based on the previous review, Delaval believes the following three options deserve further consideration, as each can adequately resolve NRC concerns and meet functional requirements for standby service in nuclear power plants:
Installation of a viscous damper.
Retrofit with a higher strength crankshaft.
Modified start with periodic inspection.
Page Nineteen
IM IV.
IMPLEMENTATION As described above, Delaval believes three system modifications the torsional damper, higher strength crankshaft and the modified start - constitute viable solutions.
A description of the hardware modifications and estimated time and cost to implement each follows.
A.
Torsional Damper
- 1.
General - In order to maximize effectiveness, the torsional damper must be mounted at the front end.
As the units in question must maintain black start capability, the engine driven pumps mounted at the engine front end cannot be removed to accomodate the damper.
Therefore, the damper would have to be attached to an extension shaft and outboard self-aligning bearing to provide the required pump clearance.
In order to rigidly support the outboard bearing, the original sump tank would be replaced with a two section tank providing clearance to anchor an "A" frame support for the outboard bearing.
As noted in Section III.C.3, special assembly techniques would be required to obtain the proper interface fits between the crankshaft and extension shaft hub.
The crankshaft would first be measured and then a hub with the proper interference precision machined.
The hub would then be heated in a bath of oil to 350 -0/*20 degrees F, and slid on the crankshaft.
Moderate risk in assembly operation exists since the hub could seize part way up the shaft if not properly aligned.
If this occurs, the hub would have to be cut off and the operation repeated with a new component.
While we have not installed a damper hub on a crankshaft in this manner, we regularly use a very similar technique to assemble the coupling hub onto the compressor crankshaft in packaged engine-compressor units.
- 2.
Description of Scope of Work - All piping mounted on or connected to the sump tank would have to be removed and then the sump tank pulled out.
Foundation bolts would be installed and the outboard bearing support grouted and bolted in place.
A new, two-piece sump tank would then be installed.
Page Twenty
IM The extension shaft hub, held in place by an interference fit, would have to be heated to 350 degrees F and then smoothly and rapidly assembled on the crankshaft.
The gear case would have to be removed to accomodate the mandrel and fixtures required to slide the hub on.
The gearcase would then be machined to clear the hub and then reinstalled.
After mounting of the damper and installation of the outboard bearing, the piping and auxiliaries would be reinstalled with the exception of -the following piping and supports requiring new fabrication:
Jacket Water Pump Suction Engine Lube Oil Supply Auxiliary Lube Oil Pump Discharge Prelube Supply to the Engine Auxiliary Fuel Pump Supply and Return Miscellaneous Control and Instrument Wiring and Tubing After completion of the reassembly, the system would be tested to assure suitable damper operation.
Then any qualification tests mandated by the NRC could be conducted.
Table II contains a detailed estimate of time (122 days) and cost ($1,150,000) to complete this activity for each engine.
B. High Strength Crankshaft
- 1. General - As noted, installation of a new, high strength crankshaft requires no modifications; physically, all engine components would remain exactly the same.
To remove the old crankshaft, all heads, pistons and rods would be removed and the flywheel and generator disconnected.
All fluid piping (including that attached to turbochargers) would be disconnected at the engine and the intake manifolds, gearcase and idler gears removed.
The crankcase would then be unbolted, cribbed upward and slid laterally toward the near wall about four feet.
The crankshaft would then be rolled out over the open portion Page Twenty-One
IM of the base and removed through the louvered door.
(The starting air compressors and dryers may have to be removed to clear the crankshaft.)
The new crankshaft could then be installed and the engine reassembled.
As all engine componentry is the same, no qualification testing would be required other than a standard post maintenance operability test.
- 2.
Scope of Work - Table III details estimates of time (34 days) and cost ($825,000) to implement this change for each engine.
Delaval has successfully performed such work before.
C.
Modified Start
- 1.
General - Implementing automatic slow surveillance starts requires installation of a pneumatic cylinder which limits fuel rack travel during acceleration, and then retracts upon reaching synchronous speed.
Some minor control panel modifications are required.
All necessary components are already qualified.
We have extensive experience with this system.
It is standard equipment on commercial power plant engines.
- 2.
Scope of Work -
A quote for design time and hardware for a so-called "soft start" system was submitted under separate cover (Ref.
- 20) with a price of $90,290 for the lead engine.
It would take two men five days to install, at a cost of $6,000, and a Test Engineer and Service Representative four days (or less) to adjust and verify operation at a cost of $7,360.
(Note that actual hours may increase due to inspection requirements, facility limitations and other unforeseeable events.)
Total estimated cost to implement the modification is $103,650 for the first engine and $6,150 for the second, or $109,800 total.
Page Twenty-Two
REFERENCES
- 1.
"Evaluation of Emergency Diesel Generator Crankshafts at Midland and San Onofre Nuclear Generating Stations",
Failure Analysis Associates Report FaAA-84-6-54.
- June, 1984.
- 2.
"Eddy Current Examination, DG1 & DG2 Crankshafts, San Onofre Nuclear Generating Station", Failure Analysis Associates Report.
August, 1984.
- 3.
"Evaluation of Transient Conditions on Emergency Diesel Generator Crankshafts at SONGS 1", Failure Analysis Associates Report FaAA-84-12-14, Rev. 1.0.
April, 1985.
- 4.
"Review of Resolution of Known Problems in Engine Components for TDI Emergency Diesel Generators", Technical Evaluation Report PNL-5600, Prepared for USNRC, Pacific Northwest Laboratory.
1985.
- 5.
"Diesel-Driven Electric Generating Sets For San Onofre Nuclear Generating Station, Unit 1, Standby Power Addition", Specification E-73001 Rev. 6, Southern California Edison Company.
January, 1976.
- 6.
"Torsional & Lateral Critical Speed, Engine Numbers 75041/42, Delaval-Enterprise Engine Model DSRV-20-4, 6000 kW/8303 HP at 450 RPM",
Prepared by Delaval Engine & Compressor Division for Southern California Edison.
October, 1975.
- 7.
R. E. Peterson, "Stress Concentration Factors", John Wiley & Sons.
1974.
- 8.
W. Ker Wilson, "Practical Solution of Torsional Vibration Problems:
Strength Calculations", Vol.
3, Chapman & Hall Ltd.,
London.
1965.
- 9.
E. J. Nestorides, "A Handbook on Torsional Vibration", B.I.C.E.R.A.
Research Laboratory, Cambridge University Press.
1958.
- 10.
G. Archer, "Some Factors Influencing the Life of Marine Crankshafts",
Transactions of the Institute of Marine Engineers, Vol. 76, Pg. 73.
1964.
- 11.
"Rules for the Calculation of Crankshafts of Diesel Engines",
International Association of Classification Societies, Working Party On Engines.
- 12.
J. E. Shigley, et al, "Mechanical Engineering Design", McGraw-Hill, New York.
1983.
INS
- 13.
R. C. Juvinall, "Stress, Strain and Strength", McGraw-Hill, New York.
1967.
- 14.
"Fatigue Design Handbook", Advances in Engineering, Vol.
4, Society of Automotive Engineers.
1968.
- 15.
H. E. Boyer (Ed.),
"Atlas of Fatigue Curves", American Society for Metals, Metals Park, Ohio.
1986.
- 16.
M. Nishihara, et al, "Fatigue Properties of Full Scale Forged and Cast Steel Crankshafts", Transactions of the Institute of Marine Engineers.
January, 1978.
- 17.
W. Ker Wilson, "Practical Solution of Torsional Vibration Problems:
Devices for Controling Vibration", Volume 4, Chapman & Hall Ltd., London.
1965.
- 18.
"Qualification Test for Delaval Engine Generator Set, San Onofre Station #1, Southern California Edison Company", Rev. 6, Transamerica Delaval Inc. Test Report.
February 13, 1976.
- 19.
Y. Fukui, et al, "The Recent Improvement on the Fatigue Strength of Solid Type Forged Steel Crankshafts", Bulletin of Marine Engineering Society in Japan, Vol. 3, No. 5.
1975.
- 20.
"Soft Start Controls Modification - Proposal AM8688", Transamerica Delaval Inc. (S. Schumacher) Letter to Southern California Edison (C. Cates).
October 10, 1986.
IMO TABLE I
.DISTRIBUTION OF FAST START DURATION DURING 300 START FACTORY TEST SAN ONOFRE ENGINE S/N 75041 TIME (SEC.)
6.1 6.2 6.3 6.4 6.5 6.6 6.7 STARTS (NO.)
8 111 85
- 67.
12 9
3 AVERAGE TIME:
6.3 SEC.
STANDARD DEVIATION:
0.1 SEC.
IM TABLE II DETAIL OF COST AND TIME TO INSTALL VISCOUS DAMPER ENGINEERING MAN DAYS PRICE DESIGN AND DETAIL SPLIT SUMP TANK 50
$46,000 DESIGN AND DETAIL EXTENSION SHAFT AND 30
$28,000 BOARD SUPPORTS 120 DESIGN AND DETAIL REPLACEMENT PIPING
$110,000 CONDUCT DAMPER TRANSIENT PERFORMANCE TEST 60
$55,000 DESIGN FIXTURING FOR DAMPER INSTALLATION 5
$5,000 SEISMICALLY QUALIFY NEW COMPONENTS 45
$41,000 WRITE INSTALLATION PROCEDURE 30
$28,000 TOTAL 345
$313,000 HARDWARE LEAD TIME PRICE DAMPER 6 MONTHS
$100,000 SUMPTANK 6 MONTHS
$67,000 OUTBOARD SUPPORT AND EXTENSION SHAFT 3 MONTHS
$50,000 REPLACEMENT PIPING 6 MONTHS
$120,000 FIXTURES 3 MONTHS
$64,000 TOTAL
$401,000 ENGINE PREPARATION (2)
MAN DAYS ($)
DOWNTIME REMOVE AUXILIARY PIPING AND EQUIPMENT MOUNTED ON SUMPTANK 120 10 DAYS REMOVE GEAR CASE 15 5 DAYS REMOVE SUMPTANK 8
2 DAYS TOTAL 143 17 DAYS
$84,000
eiNS IM0 TABLE II, (Contnued)
Page Two DAMPER INSTALLATION (2)
MAN DAYS ($)
DOWNTIME LOCATE FIXTURING 1
1 DAY PREPARE HEATING TANK 1
1 DAY INSTALL STUB SHAFT AND DAMPER 70 7 DAYS ALIGN AND GROUT OUTBOARD SUPPORT 20 4 DAYS TOTAL 92 13 DAYS
$54,000 ENGINE REASSEMBLY (2)
MACHINE GEARCASE 3
INSTALL GEARCASE AND CONNECT ALL FITTINGS 54 5 DAYS AND WIRING TOTAl.
57 5 DAYS
$34,000
- Non-Critical Path SKID REASSEMBLY (2)
INSTALL SPLIT SUMPTANK 12 3 DAYS INSTALL ORIGINAL PIPING 56 14 DAYS FLUSH ALL LUBE OIL AND FUEL OIL PIPING 14 7 DAYS FIT UP, SEAL, WELD AND PICKLE NEW PIPE 550 55 DAYS TOTAL 632 79 DAYS
$373,000
IM40 TABLE II, (Continued)
Page Three PERFORMANCE TESTING MAN DAYS ($) DOWNTIME PREPARE AND BENCH MARK TRANSDUCERS 40 INSTALL AND CALIBRATE INSTRUMENTATION 6
2 DAYS CONDUCT TRANSIENT TEST 18 6 DAYS REDUCE DATA, WRITE REPORT 40 TOTAL 104 8 DAYS
$96,000 CERTIFICATION TESTING IMPOSSIBLE TO ACCURATELY ESTIMATE AT THIS TIME.
TOTAL ESTIMATED COST LEAD ENGINE ENGINEERING
$313,000 HARDWARE
$401,000 LABOR
$545,000 PERFORMANCE TESTING
$96,000 TOTAL
$1,355,000 DOWNTIME 122 DAYS TOTAL ESTIMATED COST OF SECOND ENGINE HARDWARE
$401,000 LABOR
$545,000 TOTAL
$946,000
IMO TABLE II (Continued)
Page Four OPERATION MAY BE PERFORMED CONCURRENTLY WITH OTHER ACTIVITIES.
(1) APPROXIMATE ESTIMATES PENDING FINAL DESIGN (2)
ACTUAL HOURS MAY INCREASE DUE TO INSPECTION REQUIREMENTS AND FACILITY LIMITATIONS.
INS1 TABLE III ESTIMATED COST AND TIME TO INSTALL HIGHER STRENGTH CRANKSHAFTS MATERIALS LEAD TIME PRICE CRANKSHAFT 6 MONTHS
$330,000 BULL GEAR ON HAND
$8,000 MAIN BEARINGS ON HAND
$35,000 CONNECTING ROD BEARINGS ON HAND
$33,000 PISTON RING SETS ON HAND
$20,000 HEAD GASKET SETS ON HAND
$5,000 MISCELLANEOUS GASKETS AND CONSUMABLES ON HAND
$20,000 COUNTERWEIGHT STUDS 5 WEEKS
$8,000 TOTAL
$459,000
INS TABLE III, (Continued)
Page Two PREPARATION PROCEDURES - DRAFT AND APPROVAL CYCLE 30 DAYS (1)
STRUCTURAL ANALYSiS - ALLOWABLE FLOOR LOADING (2)
APPROX. 50 TON REQUIRED ALLOWABLE ROOF LOADING (2)
APPROX.
20 TON REQUIRED TOOLING - RIGGING AND CRIBBING FIXTURES
.20 DAYS ENGINE TOOLING 20 DAYS TOTAL 60 DAYS
$55,000 DISASSEMBLY ESTIMATED MAN HOURS PREPARE STORAGE AND WORK AREA (2)
TARP GENERATOR (2)
REMOVE LOUVERED DOOR 148 ATTACH RIGGING POINTS 128 REMOVE AFT INTAKE AIR SILENCER -
AIR BRANCH 10 REMOVE EXHAUST EXPANSION JOINT 20 REMOVE FORWARD AIR INTAKE SILENCER 10 REMOVE STARTING AIR COMPRESSOR SKIDS 64 REMOVE STARTING AIR DRYERS 32 DRAIN ENGINE FLUIDS 16 REMOVE FORE/AFT INTERCOOLER PIPING 64
IMP TABLE III, (Continued).
Page Three DISSASSEMBLY (CONT.)
ESTIMATED MAN HOURS REMOVE FORE/AFT INTERCOOLERS AND PANS 192 REMOVE LB/RB INTAKE MANIFOLD 16 REMOVE HEADERS 8
REMOVE SUBCOVER ASSEMBLIES 40 REMOVE HEADS.
152 REMOVE PISTONS AND RODS 192 REMOVE PIPING OBSTRUCTIONS -
INTERCONNECT 240
- L.O. COOLER STARTING AIR FRONT TO REAR PIPING REMOVE CONTROL TUBING -
INTERCONNECT 32 BASE TO CRANKCASE
- GEARCASE REMOVE GEARCASE AND PUMPS 48 REMOVE PUMP DRIVE GEAR ASSEMBLIES 48 REMOVE IDLER GEAR ASSEMBLIES 24 REMOVE FLYWHEEL GUARD ASSEMBLY 16 REMOVE STATOR SHROUDS 8
REMOVE PEDESTAL BEARING 8
REMOVE FLYWHEEL BOLTS AND CRIB ROTOR AFT 96 REMOVE FLYWHEEL AND CRIB - AFT 16
TABLE III, (Continued)
Page Four DISASSEMBLY (CONT.)
ESTIMATED MAN HOURS REMOVE REAR SEAL AND INTERNAL TUBING 24 UNTORQUE CRANKCASE 48 CRIB UPPER ENGINE STRUCTURE UP TO CLEAR THROUGH BOLTS AND SIDEWAYS TO CLEAR BASE ASSEMBLY 240 REMOVE COUNTERWEIGHTS AND STUDS 40 REMOVE MAIN BEARING CAPS 20 RIG CRANKSHAFT OUT 16 SUBTOTAL 2,016 17 DAYS AT 6 MEN/SHIFT, 2 -
10 HOUR SHIFTS/DAY REASSEMBLY RIG CRANKSHAFT AND INSTALL 16 MAIN BEARINGS AND CAPS 20 COUNTERWEIGHTS AND STUDS AND TORQUE 60 INSTALL THRU BOLTS - CRIB ENGINE IN PLACE 384 TORQUE CRANKCASE 48 CONNECT INTERNAL TUBING 48 INSTALL REAR SEAL 8
INSTALL FLYWHEEL - GENERATOR ROTOR 160 REAM NEW BOLT HOLES -
TORQUE ASSEMBLE PEDESTAL BEARING 16 ALIGN ENGINE/GENERATOR PEDESTAL 64
TABLE III, (Continued)
Page Five REASSEMBLY ESTIMATED MAN HOURS TEST STATOR/INSTALL SHROUDS AND TARP 16 INSTALL FLYWHEEL-GUARD 16 REASSEMBLE GEAR TRAIN 144 INSTALL GEARCASE 72 INSTALL CONTROL TUBING 32 INSTALL PIPING 240 INSTALL PLATFORM ASSEMBLIES 32 INSTALL PISTONS AND RODS AND TORQUE 240 INSTALL HEADS AND TORQUE 160 INSTALL SUBCOVERS AND HEADERS 80 INSTALL INTAKE MANIFOLDS 16 INSTALL INTERCOOLERS AND PANS 192 INSTALL INTERCOOLER PIPING 64 INSTALL AIR DRYERS AND COMPRESSOR SKIDS 128 INSTALL INTAKE AIR SILENCERS AND EXHAUST 60 EXPANSION JOINT INSTALL REAR LOUVRED DOOR 144 CLEAN ENGINE AND FLUSH SYSTEMS 64 FILL FLUID SYSTEMS 16 TEST CONTROL SYSTEMS 16 TEST ENGINE GENERATOR SET 24 SUBTOTAL 2, 580 22 DAYS AT 6 MEN/SHIFT, 2 -
10 HOUR SHIFTS
IM TABLE III, (Continued)
Page Six TOTAL LABOR 4,596 MAN HOURS 39 DAYS
$339,000 TOTAL ESTIMATED COST LEAD ENGINE (3)
MATERIALS
$459,000 PREPARATION
$ 55,000 LABOR
$339,000 TOTAL
$853,000 TOTAL ESTIMATED COST SECOND ENGINE MATERIALS
$459,000 LABOR
$339,000 TOTAL
$798,000 NOTES:
(1)
REQUIRES PARTICIPATION BY SCE.
COST NOT SPECIFIED.
(2)
TO BE PERFORMED BY SCE.
COST NOT SPECIFIED.
(3)
ALL FIGURES FOR MATERIAL PRICE AND LABOR HOURS ARE ESTIMATED ONLY.
ACTUAL COSTS AND HOURS MAY INCREASE DUE TO INSPECTION REQUIREMENTS AND FACILITY LIMITATIONS.
/A'/V/q
-,;IVs I
A'A
.?
Lo FiueI Snl re Torioa Repne Soten aiori diojomay DSV2-
FIGURE 2 SCE CRANKMEMBER STRESS CRANKMEMBER 8-9 STANDARD CONDITIONS -
FULL BMEP CRANKMEMBER 9-10 1st MODE NATURAL FREQUENCY IS 25 F1194 vib/min (N) IS 1st MODE ORDER NUMBER 22.5 (5.0) (4.5) 20 17.5 15
-5 15 (5.5) 12.5 10 7.5 (10.0)
(3.5).
5 2.5 100 125 150 175 200 225 250 275 300 325 350 375 400 425 450 475 SPEED (rpm)
GAD 2-2-87 SCESTAN2.VAR
90 FIGURE 3 FATIGUE STRENGTH vs.
LIFE 80-LOW CARBON STEEL SCE CRANKSHAFTS 70 350 z
u c
Ln 40 -----
u ID
<30 LOG-LOG (0.80 FATIGUE RATIO)
LOG-LOG (0.65 FATIGUE RATIO)
SEMI-LOG 20 10 G.M.B.
12/2/8 0
3 4
5 6
10 10C) 10 1
Inertia = J FIGURE 4a
- SIMPLE TORSIONAL PENDULUM Stiffness = K Iertia J
FIGURE 4b -
IDEALIZATION OF TORSIONAL PENDULUM
FIGURE 5a -
MATHEMATI-CAL MODEL 0
OF SIMPLE PENDULUM DETUNER A
FIGURE 5b EXAMPLE OF TYPICAL FRONT END MOUNTED PENDULUM DETUNER B filar type Bifilar type Roller type Roller type Ring type Roller typc Composite (or ball typo)
(roller and rings) typ, FIGURE 5c - EXAMPLES OF CRANKWEB MOUNTED PENDULUM DETUNERS (FROM REFERENCE 8)
ANNULAR CHANGER CONTAINING:
HUB (MOUNTED TO CRANKSHAFT).
INERTIAL MASS FIGURE 6 -
CROSS SECTION OF A TYPICAL VISCOUS DAMPER
C b
Or I
FIGURE 7 - CUTAWAY OF A TYPICAL TUNED DAMPER (GEISLINGER TYPE)
FIGURE Ba SCE CRANKMEMBER STRESS DAMPER INSTALLED CRANKMEMBER GEAR-1 1st MODE NATURAL FREQUENCY IS 767 vib/min (N) IS 1st MODE ORDER NUMBER 8,
2nd MODE NATURAL FREQUENCY IS 1396 vib/min (4.5)
[N] IS 2nd MODE ORDER NUMBER 7
6
[5.0] [4.5 4
3/
2/
0 100 125 150 175 200 225.250 275 300 325 350 375 400 425 450 475 SPEED (rpm)
GAD 2-2-87 SCEDAM1 (5)
FIGURE Bb CRANKMEMBER 8-9 SCE CRANKMEMBER STRESS CRANKMEMBER 9-10 DAMPER INSTALLED 1st MODE NATURAL FREQUENCY IS 767 vib/min (N) IS st MODE ORDER NUMBER 2nd MODE NATURAL FREQUENCY IS 1396 vib/min
[N] IS 2nd MODE ORDER NUMBER 7
6 E5.0
[4.51 5
-(4.
- 5)
(5.0)
4 2
1 100 125 150 175 200 225 250 275 300 325 350 375 400 425 450 475 SPEED (rpm)
GAD 2-2-87 SCEDAM2 (5&6)
FIGURE Sc CRANKMEMBER DAMPER-GEAR SCE CRANKMEMBER STRESS CRANKMEMBER GEAR-1 DAMPER INACTIVE 1st MODE NATURAL FREQUENCY IS 767 vib/min (N) IS 1st MODE ORDER NUMBER 50 2nd MODE NATURAL FREQUENCY IS 1396 vib/min 45
[N] IS 2nd MODE ORDER NUMBER (4. 5) 40 W
35 (5.0) m 30 U) 25 Un (5. 5) 20, -
[5.0]
[4. 53 15/I)
[5. 53 10
-(3.5) 100 125 150 175 200 225 250 275 300 325 350 375 400 425 450 475 SPEED (rpm)
GAD2-2-87 SCEDAMG. OUT (4&5)
FIGURE 9 CRANKMEMBER 8-9 (2.5)
SCE CRANKMEMBER STRESS CRANKMEMBER 9-10 RETIMED 1st MODE NATURAL FREQUENCY IS 1194 vib/min (N) IS MODE ORDER NUMBER 25 22.5 20 (5.0) 17.5-I (3.5) 15 1-12.5 Lii 10 H
7.5 (10.0) 5 2.5 0
I 1~L~
L--
100 125 150 175 200 225 250 275 300 325 350 375 400 425 450 475 SPEED (rpm)
GAD 2-2-87 SCERTIMB.VAR (3&4)
FIGURE 10 CRANKMEMBER 10-11 SCE CRANKMEMBER STRESS lst MODE NATURAL FREQUENCY IS CYLINDER BANKS 1&10 DISABLED 1194 vib/min (N) IS 1st MODE ORDER NUMBER 25 22.5 20 (5.0)
W 17.5-(2.5)
()
15 U)
(4.5) 12.5-(3.5)
U)
(5.5) 10
~
(3. 0) 7.5 (10. 0)
(7. 0) 5 2.5 100 125 150 175 200 225 250 275 300 325 350 375 400 425 450 475 SPEED (rpm)
GAD 2-2-87 SCE110B.VAR (5)
FIGURE 11 CRANKMEMBER 9-10 SCE CRANKMEMBER STRESS CRANKMEMBER 10-11 CYLINDER BANKS 1-2-9-10 DISABLED 1st MODE NATURAL FREQUENCY IS 1194 vib/min 25 (N) IS 1st MODE ORDER NUMBER 22.5 20 17.5 (5.0) 15 (3.5) 12.5 (6.0) 10 F
(4.5) 7.5(3.0) 7.5 0
5 2.5 O~
A I
I I
I I
I I
I
~
I I
I I
I 100 125 150 175 200 225 250 275 300 325 350 375 400 425 450 475 SPEED (rpm)
GAD 2-2-87 SCE1291B.,VAR (4&5)
FIGURE 12
_CRANKMEMBER 9-10 SCE CRANKMEMBER STRESS STANDARD CONDITIONS -
REDUCED BMEP 1st MODE NATURAL FREQUENCY IS 1194 vib/min (N) IS lst MODE ORDER NUMBER 12 1 1 (5.0) (4.5) 101 10 9
00 7
(5.5) 6 5
40 3 3 (10. 0) 1 (3.5) 2 0
100 125 150 175 200 225 250 275 300 325 350 375 400 425 450 475 SPEED (rpm)
GAD 2-2-87 SCEND2 (4&5)