ML20154D723

Informs That Encl Documents Titled, Status of Satellite 6 Friction Testing, & MOV Motor & Gearbox Performance Under Design Basis Loads, & MOV Actuator & Gearbox Testing, Being Submitted to PDR

ML20154D723
Person / Time
Issue date:	08/20/1998
From:	Weidenhamer G NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES)
To:	Mccolloch D NRC OFFICE OF ADMINISTRATION & RESOURCES MANAGEMENT (ARM)
References
NUDOCS 9810070310
Download: ML20154D723 (69)

Text

. _ _ . . _ - . . _ . - - . . - - .

fp9tc > -

y' g g UNITED STATES NUCLEAR REGULATORY COMMISSION

[Nk(h0/D g # WASHINGTON, D.C. 20MH1001 4 ,o 9***** August 20, 1998 MEMORANDUM TO: Donale A. McColloch Public Document Room u - ,

FROM: Gerald H. We n Electrical, Matt; rials, and Mechanical Engineering Branch Division of Engine ering, RES

SUBJECT:

SUBMITTAL OF DOCUMENTS PRESENTED AT PROFESSIONAL MEETINGS in accordance with NRC procedures, the attached documents titled, " Status of Stellite 6 Friction Testing," and "MOV Motor and Gearbox Performance Under Design Basis Loads / that were presented at the Fifth NRC/ASME Symposium on Valve and Pump Testing, held on July 21-23, 1998, in Washington, DC, are being submitted to the PDR. Another document titled, "MOV Actuator and Gearbox Testing," that was presented at the Joint AOV/MOV User's Group Meeting, on December 2-4,1997, in Clearwater, Florida, is also attached.

Copies of the papers have also been sent to the Document Control Desk end copies of the completed NRC Forms 390A have been sent to Publishing Services.

If you have any questions regarding this memorandum or the attached documents, please call me on (301) 415-6015. ,

Attachments: As stated l

I i

i I

.,jJp( ,p) 9810070310 980820 PDR ORG NREB PDR fG5(0-6

Fifth NRC/A9E Sytposian on Valve and Pump Testing MOV Motor and Gearbox Performance Under Design Basis Loads Kevin G. DeWall and John C. Watkms Idaho National F=p=~mng and Environmental Laboratory u -

~

Gerald H. WN-U.S. Nuclear Regulatory Can=ia= ion ABSTRACT This paper describes the results of valve testing sponsored by the U.S. Nuclear Regulatory C== ala= ion, Office ofNuclear Rp=*~y Research and conducted at the Idaho National Fap=-ing and Envuonmental Laboratory. The research objective was to evaluate the capabilities of specific actuator motor and gearbox assemblies under various design basis loadmg conditions.1he testing was performed using the motor-operated valve load simulator, a test fixture that simulates the stem load profiles a valve actuator would experience when closing a valve against flow and pressure loadsgs. We tested five typical motors (four ac motors and one de motor) with three gearbox assemblies at conditions a motor might exponence in a power plant, including such off-normal conditions as operation at high temperature and reduced voltage. We also determinad the efficiency of the actuator gearbox. The testing produced the following signiGaat results:

All five motors operated at or above their rated torque during tests at full voltage and ambient temperature.

l

• I For all five motors (de as well as ac), the actual torque loss due to voltage degradation was greater than the torque loss predicted using common methods Startup torques in locked rotor tests compared well with stall torques in dynamometer-type tests.

'Ihe methods F = dy used to predict torque losses due to elevated operating temperatures sometimes bounded the actual losses, but not in all cases; the greatest discrepancy involved the prediction for the de motor.

• Running efficiencies published by the manufacturer for actuator gearboxes were higher than the actual efficiencies ' determined from testing. In some instances, the published pullout efficiencies -

were also higher than the actual values.

Operation of the gearbox at elevated temperature did not affect the operating efficiency.

BACKGROUND Durmg the past several years, the U.S. Nuclear Regulatory Canmiecion (NRC), Office of Nuclear Regulatory Research has supported research at the Idaho National Engmeering and Environmental Laboratory (INEEL) addressing the perfonnance ofmotor-operated valves (MOVs). The research included tests and analyses to determine the capability of safety-related MOVs to perform their intended functions when subjected to their design-basis conditions. For some of these valves, the design-basis conditions

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

l l

include high flow and pressure loads, elevated temperatures (which can reduce th: output of the electric motor), and operation of the electnc motors at reduced voltage.

This paper presents the results of tests perfonned to address factors that affect the performance of MOV electric motor and actuator gearbox assemblies. Specifict.lly, the testing addressed the following questions:

~

How does the actual, measured output torque of the actuator motor compare with'the torque' -

characteristics published by the manufacturer?

=

How much does the output torque of the motor hs at various reductions in the voltage  !

supplied to the motor? How do these measured values of torque rMAw compare with the I rdWe calculated by typical analytical predictions?

How much does the output torque of the motor hs as the motor's operat'ng temperdure increases? How do these measured values of torque reduction compare with the reductions calculated by typical analytical predictions?

. What is the actual efficiency of the actuator gearbox, especially at high loadmgs and elevated twnores? How does the actual efficiency compare with the manufacturer's published efficiency values?

l TEST EQUIPMENT i

We tested six combinations of actuator gearboxes and electric motors, using five motors and three i gearboxes. Table 1 summarizes the information provided by the motor manufacturers and the actuator ,

manufacturer about the six actuator configurations. Note that the SMB-1 and the SMB-0 actuators were each

)

tested under two configurations.

The tests were perfccmed on the MOV load simulator (MOVLS), a test fixture owned by the NRC and built by the INEEL for testing valve actuators The MOVLS, shown in Figure 1, uses actuators, valve yokes, and valve stems the same as they are assembled on the valves. The MOVLS simulates valve loads by using a hydraulic cylinder / piston assembly and an accumulator that contains a gas ovo.y-..c. The specific valve load profile is controlled by the initial water level and gas pressure in the ac:umulator. This configuration allows us to impose a steaddy increasing load on the stem, very similar to what an actuator would experience when closing a valve against an actual flow load. The valve seating load is simulated when the piston bottoms out in the cylinder.

The MOVLS is instrumented to take all the measurements that are important for diagnosing valve actuator perfonnance Motor speed was measured directly, and motor temperature was measured using a combination of thermocouples and an infrared sensor, allowing us to amnitor the actual internal rotor temperature. The output torque of the electric motor was measured by a torque cell mounted between the motor and the gearbox. The output torque of the gearbox was measured by a calibrated torque arm attached to the valve stem. With direct measurement of both the output torque of the motor and the output torque of the gearbox, we were able to continuously monitor the efficiency of the gearbox.

I 1 l

Most actuator and motor testing is typically performed by applying a sudden torque load. This is accomplished by applying the brake on a dynamometer, hard-seating a valve without a flow load, or with a l locked stem (similar to a locked rotor test). In contrast, we used the MOVLS to produce dynamometer-type

tests that imposed a stem thrust load that gradually increased until it caused the motor to stall. Stall occurred l before the piston bottomed out in the cylinder. These tests allowed us to determme the actual output torque of the motor and of the gearbox over the entire operating range cf the motor. We also conducted locked rotor

! startup tests, measuring torque, current, and other parameters with no motor shaft rotation.

For each motor / gearbox combination, baseline tests were conducted with the assembly at normal conditions, then tests were umAM at various stages of reduced voltage, at various levels of elevated ,

openating temperature, and with selected combinations of the two A three-phase,60-amp-per-leg auto transformer was used to perform the degraded voltage tests. Tids same auto transformer was used as the supply to the de power source to permit degraded voltage tests for the de motor. In the elevated temperature tests, we wrapped each motor with heat tape and insulation to create a custom oven. Environmentally qualified motors were heated to 300*F, while the other motors were heated only to 250*F. All testing at ambient temperature was conducted with an internal motcr temperature between 70 and 80*F.

l AC MOTOR RESULTS Our research addresses most of the terms in the typical formula for predicting the output torque of a valve actuator. The formulais:

1

( i n

T -T. p r m F, F Eff, OAR (1) where T, = output torque of the valve actuator T. = rated starting torque of the electric motor V, = actual voltage supplied to the motor V. = the motor's rated voltage n = 2 for ac motors, I for de motors F, = factor to account for losses due to motor heating F, = application factor Effg= gearbox efliciency

, OAR = overall actuator ratio (gear ratio) l l The first three terms in the equation apply to the motor torque and account for the effects of reduced voltage and elevated temperature. The results concernmg the voltage- and temperature-related adjustments are meluded in this section for the ac motors and in the following section for the de motor. The application

m. _ ._.-._._ _ _ _ _.- _ _ _ _.___._._._.__..__ ___._.

y I

factor (F,) was not evaluated in this research. We used an application factor equal to 1.0 in all of the calculations discussed in this paper. The gearbox efliciency section of this paper discusses the final two terms in Equation 1.

Perfonnance Curves for the ac Motors (T )

1 Figure 2 presents the manufacturer's torque curves for the 25 ft-lb ac motor. The asterisk on the current trace marks the end of the manufacturer's curve, and the two "X"s are test points from earlier field - ~

testing. The remamder of the curve is extrapolated through the published locked rotor current. Figure 2 also shows the actual torque curves for the motor. The two speed / torque curves are quite similar in shape; both indicate that at about 1200 rpm the motor begins to stall. Very little additional torque is produced after the motor speed drops below 1200 to 1000 rpm. (This part of the trace is called the knee of the curve.)

Although the two speed / torque curves shown in Figure 2 are similar in shape, they show a difference in available torque. The test data show about 30 ft-lb torque available in this rpm range (1000 to 1200), while

- the manufacturer's curve shows about 25 ft-lb. Another difference is at very low rpm (less than 200). The manufacturer's curve shows a significant increase in torque as the motor approaches stall, while the test data shows a moderate increase followed by a rapid diww.w. Both curves show about the same stall torque.

Compt.risons for the 5 ft-lb and 40 ft-lb motors are similar.

Figure 3 presents the manufacturer's torque curves for the 60 ft-lb ac motor. Again, we extrapolated the current from the asterisk through the published locked rotor stall current. Figure 3 also shows the actual curves derived from our test data. The two speed / torque curves are different over the working range, but the absolute torque values are similar. However, the curve from our test data indicates a knee at about 1200 rpm i and 58 ft-lb, while the knee of the manufacturer's curve is at about 1300 rpm and 52 ft-lb. The actual peak I torque of about 64 A-lb occurs somewhere between 500 and 1000 rpm. The manufacturer's curve shows a peak torque of about 64 ft lb at stall. This motor's output torque is close to its 60 ft-lb rating, and the manufacturer's curve slightly overestimates the actual performance at stall.

i Overall, the test data show that the load threshold at which the motor will drop to a stall (the knee of the curve) does not always occur at the motor speed or at the load threshold indicated by the published data; however, the actual torque output at the knee of the curve (1200 to 1000 rpm for the three 1800 rpm motors, and 2400 to 2000 rpm for the high-speed 3600 rpm motor) is consistently higher than iadiceM by the manufacturer's curves. There is some variation between the published rated torque and the torque we measured at high loadings and at stall. All four ac motors exceeded their rated output torque, some by larger margins than others Also, the torque at the knee of the curve was greater than or equal to the stall torque for all four of these ac motors.

Degraded Voltage Testing of the ac Motors V

t m Equation 1, as applied by the manufacturer, uses the rated starting torque of the electric motor, however, the amount by which the actual torque exceeds the rated torque varies as shown above. Recently, there has been interest in substituting measured values ofmotor torque for the rated starting torque in the

equation. We therefore focused our analysis on the voltage squared term. For our reduced voltage evaluation we compared the measured motor torqw output in the baseline (100% voltage) test with the measured motor torque at each reduced voltage condition, as shown in the following equation.

_ _ _ __m--~ __ _ _. _ _ _ _ . _ _ _ _ _ _ . ._ _ _ __ _ _-.__.___ _ _ _

4

'V'*

(2)

T . T. x [=s s.

- T, =

actual torque at reduced voltage u -

~

=

T.i actual torque at 100% voltage V, = actual voltage V. i

= 100% voltage.

i Figure 4 shows motor speed / torque curves for the 40 A-lb motor at degraded voltages down to 60%

of the mmal 460 vac. The results of a voltage squared calculatian to predict a single value of the running torque (near the knee of the curve, at 2000 rpm) and the stall torque at degraded voltage are also shown. For the high speed 40-A-lb ac motor, the voltage squared e-lad +im overestimatac the motor torque by about 1.0 to 2.8 A-lb(3 to 7%) at 2000 rpm.

To further evaluate the voltage squared method, we ie . ged the formula shown in Equation (2) and used the data from all four ac motors as input to solve for the ey=t Figurc 5 shows the results of this evaluation. Each data point shown in Figure 5 i+w the cycr.c.ra in Equation 2 that pr=bres a predicted torque that matches the actual torque measured in the correspondag test. The results suggest that for these four ac motors, an exponent of 2.5 instead of 2 in the voltage squared calculation would produce i predictions of torque loss that consistently bound the actual torque losses l Locked RotorTesting of the ac Motors The small delta symbols at the bottom of the plot in Figure 4 are the locked rotor startup tortaes ,

These individual data points are produced by energizing the motor with a large load imaa-A on the motor that prevents the rotor from turmng. The speed curves in Figure 4 end (at motor stall) very near the locked rotor startup torques, indir=*ing that the motor torque at stall and the startup motor torque with the rotor l locked are about the same.

For all four of the ac motors we tested, the locked rotor startup torques were always lower than the peak running torque. We found the locked rotor startup torque to be a useful indication for compariron with the rating of a specific motor.

Elevated Temperature Testing of the ac Motors (F.,,,)

Figure 6 shows the actual motor torque measured at elevated tegore for 100% voltage and for 80% voltage for the four ac motors we tested. The figures also show the predictions for these motors, based on the manufacturtr's data (Reference 1). The predictions of torque loss due to elevated terrwnure bounded the actual losses for three of the four ac motors we tested. However, the results shown in Figure 6 for the 25 A-lb motor indicate that the predictions might not be appropriate for this motor. Note, however, that this motor is not environmentally qualified, a fact that might affect its output at elevated temperature.

e- -.v. _ -- , - -

_ _ __ _ _ _ __ . _ _ . . _ ___ __. ~_ _._ _

l l

l 4 i DC MOTOR RESULTS  !

The performance of a de motor is somewhat different than that of an ac motor. An ac motor tends to stall quickly when the load reaches a certain threshold. This is because above that threshold, very little additional torque is available to handle an increase in the load. In contrast, a de motor responds to a load meresse by continuing to produce additional torque, albeit at lower rpm, until the motor finally stalls at a peak torque value. There are also differences in the responses to degraded voltage conditions and elevated tsr64are conditions. " ~

~*

The 4,0 ft lb de motor used in this test program was the same one that had been used in earlier testing to evaluate the effects of carthquakes on valve operation and on piping system and piping support system integrity. We procured the valve in the mid-1980s from the .h- Asioned Shippingport nuclear power station and subjected it to the setsnue tests. The history of the motor and the results of the seismic tests are reported in Reference 2.

Performance Curves for the de Motor (T.)

As with the ac motors, we developed motor pedormar= curves for the de motor (motor speed versus motor torque, and motor current versus torque) for comparison with the motor data supplied by the actuator i manufacturer. Figure 7 presents the manufacturer's speed / torque and current / torque curves for this 40 ft lb de motor. The figure also shows the torque curves derived from our tests. 'Ihc actual torque output is lower than predicted by the manufacturer's data.

Degraded Voltage Testing of the de Motor Vy i,V In analytical evaluations of MOV capability, the actuator manufacturer issu.=

• a formula that is identical to the voltage squared method for ac motors, except that the exponent is 1 instead of 2.

Figure 8 shows the speed versus torque curves for the degraded voltage tests. Values representing estimates based on the method described above are also shown, identified by "x". The estimates are based on the results of the 100% voltage test at 1000 rpm,500 rpm, and stall. Thesc iesults show that the conventional method for predicting torque loss due to operation of de motors at reduced voltage does not bound actualtorque losses.

To further evaluate t'he method for pridicting de motor output at reduced voltage conditions, we used data from the tests as input so we could solve for the exponent. Similar to our effort to evaluate the voltage

, squad method for the ac motors, the purpose of this effort was to determine whether an exponent other than 1 might be consistent with the test results for the de motor. The results are shown in Figure 9. The results suggest that for locked rotor (stall) conditions, an exponent of 1.3 would bound the data; for a motor speed of 500 rpm, the results suggest an exponent of 1.8 in the formula to bound the torque output of this de motor.

Locked Rotor Testing of the de Motor Figure 8 also shows the locked rotor startup torques for the de motor. The locked rotor results show fair agreement with the stall torques indicated at the ends of(or extrapolated from the ends of) the traces representing the runmng tests. (Note that some of tie traces for the running tests do not reach zero rpm; we shut the motor off while it was still turning very slowly, to prevent overheating of the motor.)

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

l l

Elevated Temperature Testing of the de Motor (FQ As the ambient temperature increases, the de motor's maximum output torque decreases. The manufacturer's de actuator qualification requires the actuator to perform at 340*F. For de motors operating

! at this temperature, the manufacturer provides a table (Reference 3) rM+mc-=Ang adjustmems to the rated torque value when sizing a nuclear qualified actuator. According to this table, a de motor with a rated torque of 40 ft-lb would be sized as though it were a 39 ft lb motor. (The tw+2< =M adjustment is greater for larger motors; for example, a motor rated at 60 ft-lb would become a 54 ft-lb motor at 340 F.M -

Figure 10 contains the elevated temperature plots for the 40 ft lb de motor for 100% voltage. Our results show that the torque losses experienced by this motor at elevated temperature are greater than the losses indicated by the data provided by the manufacturer. & increase from room temperature to 250*F reduced the output torque by 8 to 10 ft-lb at high loads.

l i

GEARBOX EFFICIENCY Gearbox efficiency, along with the overall eher ratio, define the relationship between the input torque and the output torque of an actuator gearbox. W output torque can be represented by T, -T Eff, OAR (3) where '

T, = output torque Tw = input torque (motor torque after adjustments described earlier in this report)

Eff%= the efliciency of the gearbox.

OAR = overall actuator ratio (gear ratio) h input torque consists of the torque delivered by the electric motor to the input side of the gearbox, and the output torque consists of the torque delivered to the stem nut. W gearbox efficiencies identified by tle manufacturer are called the pullout efficiency, stall efficiency, and runmng efficiency. h pullout efficiency is the lowest of the three. This value applies when a motor is lugging at very low speed under a load or starting up against a load. h stall efficiency is higher than the others because it includes consideration of motor inr:tia during a sudden stall; it is typically used in evaluations of possible overload problems. The running efficiency applies when the gearbox is operating at normal motor speed and normal loads.

Figure 11 includes two simple data plots showing the actuator torque (output torque) and the motor torque (input torque) measured during the 100 percent voltage test of the SMB-1 actuator with the 60 ft-lb ac motor. Figure 11 also includes an x-y plot of actuator torque versus motor torque, which represents the

gearbox OAR times the actual gearbox efficiency. b two straight lines in the Figure 11 x-y plot are l calculated using the OAR times the (a) published runnmg efficiency and (b) he published pullout efficiency.

Figure 12 shows the motor torque and valve stem torque data for the SMB-1-60 reduced voltage

, tests, suggesting a relationship between efliciency and the speed of the SMB-1-60 ac valve actuator. In each J

I

j of the reduced voltage tests, the measured efliciency is near the published runnmg efficiency when the motor is near its normal speed, but drops toward the pullout efficierry as the motor experiences higher loads and approaches stall. For this actuator, the published pullout efficiency bounds the actual gearbox efficiency for all tests at torque loads up to the motor's rated torque.

Similar data for the other ac valve actuators are presented in Figures 13 through 15. For the SMB-00-5 (Figure 13), the actual gearbox efficiency is well below the published pullout efYiciency. For this actuator, the motor torque required to rotate the gear train without producing output torque (sometimes called ~

the hotel load) is a significant percentage of the total motor toique. A mon: meamngful comparison of gearbox efficiency can be made by subtracting a hotel load of 0.44 ft-lb motor torque from the SMB-00-5 data. Figure 16 includes this 0.44 ft-lb offset, based on no-load motor torque measurements, to account for the hotel load. This provides a more meamngful comparison; however, the performance is still lower than the pullout efficiency. The data show the importance of considering the hotel load when deternumng the actuator capability for smaller motors Figure 17 presents the data from testing of the SMB-1-40 actuator when powered by the de motor The shape of the curves is slightly different than in Figures 12 through 16 (the difference is due to the speed versus torque relationship of de motors as compared to ac motors), but the general trends are similar to those seen with the ac motors. For this SMB-1-40 actuator powered by a de motor, the sctual gearbox efficiency was consistently lower than the published running efficiency. As the motor speed drops under high load, the actual efficiency drops to values well below the published pullout efficiency.

Finally, Figure 18 shows the results of testing of the SMB-0-25 ac actuator to determme if gearbox l efficiency is affected by elevated temperature. In these tests, the second gear ratio shown in Table I was used l (a different gear set than in the tests described earlier). Three tests were performed 'Ihe first test was a baseline test to show gearbox efficiency at room twwsure. The second and third tests were perfonned l with the gearbox heated to 350*F. The results are about the same for all three tests, indicating that the gearbox efficiency was not affected by elevated temperature.

Figure 18 also shows that the measured efliciency was slightly higher than the published pullout l efficiency. By comparing this figure to Figure 14, we get an indication of the variation that can result from

! using different gear sets in the same actuator. Tests using a gear set with a lower gear ratio (higher output speed, lower output torque) produced a higher gearbox efficiency.

CONCLUSIONS

, Performance curves. Analysis of the test results included a comparison of the published motor performance data with the actual current / torque and speed / torque data. There were some minor differences in the shape of the ac motor curves near the knee of the curve, indicating that the load threshold at which the motor drops off to a stall occurs at a different rpm, or at a different torque load, than indicated by the

published data. For all five motors, the stall torque, the runmng torque before stall, and the locked rotor startup torque exceeded the rated torque. Some of the motors had more margin between the actual and rated values than others.

l Degraded voltage. With the measured ac motor torque at 100% voltage used as the basis for the i

calculation, the actual torque losses due to voltage degradation were greater than the losses estimated by the voltage squared calculation. This was true for all four ac motors at all the various reduced voltages we tested.

The results suggest that for these four ac motors, an exponent of 2.5 instead of 2 in the voltage squared calculation would bound actual torque losses due to operation at reduced voltage.

Using the rated torque in the voltage squared calculation (instead of the actual torque) will not provide a bounding estimate of torque losses unless the actual torque at normal voltage is significantly higher than the rated torque. For three of the ac motors we tested, the actual torque was at least 25% higher than the rated torque. For the other ac motor (the 60-ft-lb motor), the actual torque was very near the rated torque.

For this motor, the voltage squared calculation does not provide a boundmg estimate of torquelosses, for - -

cither the rated or actual torque value used as the basis for the calculation.

b de motor results were similar to the ac motor results; actual torque losses due to voltage degradation were greater than the losses estimated by the typical linear method used for predicting such losses. W results suggest that a formula similar to the voltage squared method, but with an exponent of 1.3, would provide a boundmg estimate of the actual torque losses for this de motor at locked rotor conditions. At a motor speed of 500 rpm, an exponent of about 1.8 would produce boundmg estimates of torque losses.

Elevated temperature. For three of the ac motors, the actual motor output torques measured at elevated temperatu c were equal to or higher than the industry predictions of motor torque at those conditions.

For the 25 fi-lb ac motor (not environmentally qualified), the actual torques were lower than the predictions, by 3 to 8%. W actual motor cunents measured during the tests were lower than the predictions. For the de motor, a torque prediction based on the actuator manufacturer's data fce de motor perfom ance at elevated temperature overestimated the actual torque measured in the tests, and by a significant margin.

Gearbor eficiency. For most motor / gearbox combinations, the actual efficiencies were lower than the runmng emeiencies specified by the actuator manufacturer. In no case did the published runmng efficiency provide a lower bound for the actual emciency of the gearbox when operating against high loads typical of design basis loads. Generally, the published pull-out efficiency bounded the actual gearbox emeiency at moderate loads, but in some instances it did not bound the actual gearbox emciencies at higher loads at or near motor stall. It is important to consider the hotel load when deternunmg the capabilities of actuators with smaller motors Operation of the gearbox at elevated temperature did not affect the operating efficiency of the gearbox.

Gearbox dYiciency tended to be lower with operation at lower speeds. This is particularly true for actuators powered by de motors, because de motors approach their highest output torque at low rpm. Higher gearbox efficiency corresponds with operation at higher speeds. This finding was indicated in all the test resuhs and confirmed by msults from testing of the same ac-powered actuator with two different helical gear sets. The gear set with the lower gear ratio (lower output torque, higher output speed) operated with higher emciency.

REFERENCES

1. Limitorque Technical Update #93-03, September 1993.

2. NUREGICR-4977, SHAG Test Series-Seismic Research on an Aged United States Gate Valve and on a Piping System in the DecommissionedHeissdampfreaktor (HDR), R. Steele, Jr.,

J. G. Arendts, Idaho National Engineering Laboratory, EGG-2505,1989.

3. Limitorque SEL-5, November 9,1988.

Table 1. Test hardware.

SMB-00-5ac SMB-0-25ac SMB-0-25ac SMB-1-60ac SB-1-40ac SMB-1-40dc Motorrated torque (ft-Ib) 5 25 25 60 40 40 Specified stall torque (ft-lb) 6.5 29.5 29.5 66.0 49.0 63.0 Motor rated speed (rpm) 1700 1700 1700 1700 3400 1900 -

Motor rated voltage 460 vac 34 460 vac 34 460 vac 34 460 vac 34 460 vac 3$ 125 vdc j Overallratio 87.8 69.56 34.96 42 50 32.13 42.50  ;

Worm gearratio 45 to 1 37 to 1 37 to 1 34 to 1 34 to 1 34 to 1  !

Helical gear set 22/43 25/47 37/35 32/40 37/35 32/40 L

Runnmg efficiency 0.50 0.50 0.55 0.50 0.60 0.50 ,

Pullout cuiciency 0.40 0.40 0.40 0.40 0.45 0.40 Stallefficiency 0.50 0.50 0.55 0.50 0.60 0.50 Application facter 0.90 0.90 0.90 0.90 0.90 0.90 Motor frame K56 PS6 PS6 FE56 184R2 D202G ,

1 I

l- i

a . - . - . . . ..___...-.n a_- ._n .u - . _ _ . . > . , . ~ ~ . _ . - . . _ . . . . _ . -- ~_.,._ ,~- _.--- ___-- _ _

W e

i a

i i

1 l

4 I

l 1

1 I

A. s.. l: ,.D '

'3. '

'4 .,,,p"' ,' s%

M ' ',. 4 4 u 1

YN HEN SVfr/ .s My4.v.s25%ips$ii@j'($:,,$,( $ " j$n .

7' s

>y .4

-( s' i

s%vx +

qg;g+'

. ,, ,.,,4 1>

ph:..ys:. ... : - Nw::y:j 3.g 3 ..... > ~.,1> gs * - .. < 't,'t',s ,.s s %

s os -

.s s

.,.& \,

.s ,  ; << ,  :' ,s xs ,;s ; t ,s : > q .

s;s ,  % ,, s 5

, b g +,  ?; ( .'s ;D$ + */s'^'; , " '.\

s,' % '^s 4

jm: , .M 4,.. 'f,

-lxss~u - -3

<+d ., st < ss,.,v.,' u, s. << <

a .

..v+

,,s< .* s A w p ,, (w '--s r y < ,-s a , -3'5 ' $ vmsI'

{ + ,t$( C'< <su, ' e#' ^< 2 '. / -

'[ <f's3 $ a 'd,' jvas' D'*4 q>,2 T,,'4SI $ ^ds .- ^%,3 i s 1

- 2:

( s

,,,4,Y,s.,-^s 1 e$

Y' s 's

<,psm ,a

' u *'# > 1 ;,.y, . ,s W, v%,

s s '

s,s sss (, ss '\ ' v ' <

+ . > .u

) jkv^'"Q*' 9 W"e 3 a > 's

~s

' , w-1 , n,i, s g sf',3rp ',.u"',' i ' {lsy,$ ' ;,% o ^ ' , Y, ,

{ [ Iv0 ,t, i ' ^) ^A ,, s. U, j

s m +v s g * ' <. s 4.< . ssse: s , -- .'ss%v g. s + > > s s s's' s s ,

s; %'s +d s-s ;g^-:y * 's n'g; s nj i: ,'; ^

~4,., fs l

'fft

, ^ ' % t s

{

g Q'Q ', ' s

, .i (_ ' >

,^p> a< d, e '

. 4 'egs s.

e. s ,..x'z. .

3' q,n ,'>' , n ,>) '4$>

+ t, s < - s q

, A 4 ys  ?'>'; e,.u [" A t / '; j 4 s)1 '? deg sf? s* ,, ,

j

+

- ,9 .,y W'(s q# fs. ; "' .t '; >'

, rg ' ' ,y , sn(  ;, ,, ;A 1  % 'h

' .];,fs'q d '^ .$' . qdr '.sO'(.sj ,.:4 3 3>'l [ s*/ 's b .

'l l*g ^, /l ['oSOY b " h 3 l * + " <

't',

>'s); #^ 'Us? ' '3 4 % <

• t , ,:,,

,., 9, q sg * ' l > '<m mc, j ',^ , { ', 9-/ ^ > 0 1 ,

d # x l'# [*

• s N,a s a, >n~ sd ,'J t.s .: 'ss9's siss >s> w srm e es

- > , . -us, ,m 9-m +A y s.'. s 4s s , x>c , s p

1 q .s  : ,^ .,1 ,>"ss: .)swqvs qA

'>s s

s .sg,es9 s

,'[ l <"OrY.s ' e s <

z' t s- - s- -

gM e)> 3 f '} ' &}{s ,<am 4

De < 3,, e<M'! t,z <v> +<>s.

,.e s <

N sf ,[ , 'G. > l^ , y.y '.,>sa, s <-

• j ,5 s s

_Y e- t ,s' , , h <>

q>

, e s+ 3,13, rt >+. m's ^' 3, s

e ,Nsyss<+

'Je u4 'c ;% <j S'(s f lg, >, ss( y'  : xIE 1 s p s's sw '

ss e  ; a 7 '<Q,,.

W' N' 4

>< +

, f, m y ,. , s i ,,s,9.h< s 9** >, e '

dl

, ' e jsyt sw y x.

n

< r. w, .,:

d

'g A s '" g+ s, +g, ,

>< > A' '

's

,/ A ,

^^ ,~ g '# z

)( ,

a,, ss'^'ss

,5 ~. %: s' :,,n, y,.; s

,,+W,s'

' ^'

v

$i%vq' x

, ...:..:2$f  : f.kM-~:f .. . .

g.:.[ .3 Ek;M INN

. :-s;p.x " ' -issi; q $jA$:$.:sx;. 4. y f G.j % h:ce

, : s:

i $:Y,fE e .st:8O.

-g: v,_.N:-

s.v^

y$;$$.&

>.m..,,

-;..w 4 = &, $kf5.-:: : j$i?*lifi'

. . . .: 5 NiNN ..N7

{ Nh yW; c: y .MS).v g NWF.<.ii+,

yY"$ ' W::p,>4htm&:v.,??WW&.;gby:

.I 1

x:E p;

We:'.

i?:p. :9.:

i>X:Y S n g: p:+I.

3Q.

%jp:g:y:i'f?.iM;y; j

m+::gy.> :xm<Mdp*.W;<5 e.vit:VSs;q:;j:d.'j j h'. ' "

P j:4ij 3.q.7,7,85't;y{ygg

,-. .w

%[3am',5,5 ,,

3 a

Gw;s^: ts!:is S%5. *W ';.)9.

+:

n ss't,P

?

w%v.

5< .

$ N$9.:3,U,gfM5%F8:4:

Y;8' vi .E ..$S@iki ,.yf<[;/45

%.x' .9,y:3h.<:VUSRA;kh":'  :-

5kNb l

?,"%55^ .i.

"'+'" '

) if i

d 3rdbI!Ms d l

~ WNd5MID..S:f.ES 5f 4  !

4 ,'+-

. .;A s I.' .U;kE*$NS$$

0$ili $:

,i@'iijjd.Nb[ .

^#w)#x57-+ gmp:pa"y^/fd

x*.+qq".vg:.s'W<'4:x+f :g'p+:

4 -

';> s M;:"' :e. . ' p ,m d

di' :j!;  ;$w.-x . ..* .::(4%'<'F2SW4 1 /,sn y.

$& >yg.Ly:.

I?'C ~'. M%L 9v,

'^ .>*. -P'

. . XYM I p, ...: ,2

^

' " ": y),i@ > y , ,, ,xi.$;is hjl - . - - h,kh{p3A?9>4j$:g:w: h f:h;; !5 "

ug i:p, c3

' ' f' , / ; ;, 3, %;^,, ,. ,

'mc

^

'ty ."  ::gyn:vgggg;g: j ng;9;ng n .ng:n S

<.psu ,

f v72 ,, ' ;, . > a~ ' < '-

y g 9 p.

[ $. k.n'( > ;[ .N, i' f 'j, l'; y .5> 7"' N h 9,^ ,tN((hhh h N N h?p.y dNtc gj,3,;s97fgj.;Sg }

. '.<>rq

- s .

A M.j >>. <e;" ' y: u,'. p."y+,s, 9 r.s. e "nn sp, M h,y.,.

< x. ' i ..

s+' ..

-

• g ,.g

,o, , p . "

, v ~ > -

> - , ^'.

. , ,:ag' p p,;o-  ;;, g;g s --,

,~..;, ~~,, . g"5$;._ , ;;;sg;:qqs({:q: _f,g:;;;:)g;&;,.g,

+g::;

p , ;;g j y,,. 7,,  ;..<1' .,.,%- 4 s

- ,,s., <.,e ' .

.s 4:pj;ps..g .,. .s..- ,y,4

.<*.p

.f".,

g,,(m,..

s. _. ..

s.

. :s::

-s

> s q

. , s s s.  ?:v:0- 3 , , g .jg., g.j:p:

ft:Cir:snW hwi$;Y

' 8: s $ y^

hWE.

• E $3:j[GD.,,,: .

4 i stf3bM O, < ~b:

f .q $.# tw-x::J f

1 - < ..

t + -

g$$si?5;$ess:$.sg a gsw M:spEter.s:.ep<4;4.ze:sessevav3:t: xsw+

w& ,

- , s x > - < < N.g m 1

i Figure 1. Photograph of the motor-operated valve load sinnlator (MOVLS).

J 4

4 1

i

2,800 . . . . . . . . 28 25 ft lb ac motor ,-

2,400

• End of manufacturers curve '

j 24 X Data points from cartior tosung f

--- Extrapolated /

2,000 -

,/ 20 g g 1

~- E

~ 1,600 -

,/

16 Test data m 1,200 - Manufacturer's data e' -

12 0 2 s' o

/ \ 5 800 -

8 2 Current 400 j 4

0 ' ' '

~

O O 4 8- 12 16 20 24 28 32 36 l Motor torque (ft-lb) , , , , ,

i Figure 2. The 25 ft-Ib nWor produced more torque at the knee (1,000 rpm than predicted by the manufacturer.

2,000 . . . . . ., 50 60 ft-Ib ac motor a /

/ l 1,500 E

/

he \

-Manufacturer's data Test data /

30 ;

1,000 / S e t-S -

./ 8 F -

20

$' ,/ 9'

s a 300 -

U

• End of manufacturo/s curvo

--- Extrapolated 0

O O 10 20 30 40 50 60 70 Motor torque (ft-lb) , , , , ,

Figure 3. The 60 t-Ib ac motor's speed / torque curves are different at lower loads but the torque at the knee (1,000 Rpm) is about the same.

4,000 . . . . . .

High speed 40 ft-lb ac motor, ambient temperature 3,500 -

A Locked rotor starting test -

O Calculated value 3,000 -

E " '

g 2,500 -

• 2,000 -

o o o o -

I b=

l S 1,500 -

o 2

1,000 -

500 - -

60% volt 70% volt 80% volt 90% voit 100% volt l 0 'i - i ' -i '" '

i ' '

O 10 20 30 40 50 60 70 Motor torque (ft-lb) . ..m Figure 4. For the 40 ft-lb ac motor, the voltage squared calculation overestimates the actual torque at the knee (2,000 rpm).

3.5 .

Motors at " knee" of the curve  :

3.0 -

g . .

g 2.5 -

O n ~

o o U E m E  : A A E  :

• O i

O. O O -

8 2.0 - -

g . .

e -

y  :  :

y 1.5 -- -

~

O 5 ft-Ib ac O 25 Alb ac -

l 1.0 -

A 60 ft-Ib ac o 40 ft-lb ac .

0.5 40 50 60 70 80 90 100 Voltage ratio (percent of nominal)

Figure 5. An exponent of 2.5 is a better bound of the INEEL data.

I 7 45 5 ne enmar

• e Aduw torque at 100%vennes -

a Amuusierque s sr%vennee

- Pro &uon

• 40 -

[

6 .

ku k e -

y _

g5

• Amust tonpo m 100% venues '

a Amuss w m eouveneso E

3 - ewdk: son 3 3 30 - -

4 ,

25 - -

i A -

• i .

• 4o na enesor ,

3 ' '

20 '

i 50 100 150 200 250 300 350 50 100 150 200 250 300 '

Temperature (*F) Temperature (*F) 35 e5 i i

e Amusiony.mioo%vanue.

• Amuw to que m ioo% ve==os : t

• Amuwsorgu.woosvon se e0 -

a Aauw i ,u. m eosvanese 3a ,

- er.mean . - er.eenen 55 -

l 5 *

• $ i E 25 . - E 80  :

48 '

3 20 W

)

& , 4a 15 -

a -

asna enmar. . so n-e enator '

25 10 50 100 150 200 250 300 50 100 150 200 6, 250 300 Temperature (*F) Temperature (*F) ep 19474 Figure 6. The predicted torque losses due to elevated ^- m.h e were equal to or greater than the actual losses for three o(the four motors

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

l \

l i

4,000 . . . . . 150 3,500 -

a 3,000 ^

n o.

E  : 100 E -

e. 2,500 -

3 j sw ,

current g, 2,000 m

a

{o N1 ,500 - Test Manufacturer's

$ l 50 @ l 1,000 -

500 -

0 O

O 10 20 30 40 50 60 Motor torque (ft-lb) .,

l

)

Figure 7. During testing of *he 40 R-lb de motor, the actual torque output was lower than predicted by the mutufacturer's curves.

l 2,500 , .- , , , , , , , ,

I 40 ft-lb dc rnotor, amblent tomporaturo A Locked rotor starting test 2,000 '100 %

- x calculated value .

p 1,750 - -

9 80 %

- 1,500 -

sos -

3 70 %

g 1,250 -

7d' m

60 % 90 %

B 1,000 -

x x x 50% -

E 50% 7 2 750 -

-aos -

m 500 -

x x -

250 - -

O'~ ' ' '

't '^ ^ :- ^ ^ ^ ^ = ^: * ' ' ' * ' - - '

l 0 5 10 15 20 25 30 35 40 45 50 55 l Motor torque (ft-Ib) w ...

, Figure 8. The conventional metixxl for predicting torque loss at reduced voltage underestimated

! the actuallosses.

4 _ _ _ _ _ _ _ . _ . . - _ . . . _ . _ . _ . _ . _ . . _ . _ _ _. - . _ . _ . _ . _ - . . _

3.5 ....,..... ... .... .... . .

40 ft-Ib de motor O 1000 RPM '

. O 500 RPM 3.0 --

A' Locked rotor-y .

O O -

g 2.5 -

0 0 .

~

g  : O  :

e  :  :

8y 2.0 -

O o o O

o 1.5 -

A A A A A 1.0 -

0.5 40 50 60 70 80 90 100 Voltage ratio (percent of nominal) l Figure 9. Fwts of 1.3 for kcked rotor and 1.8 for 500 rpm would provide a better bound <

of theINEEL data.

l 2,500 . . . . . .- . . . . 1 40 ft-lb de motor, high temperature Ambient 2,250 -

100*F -

. . .. . . . . . . . 150 p 2,000 -

\  %

- - - 200*F -

250*F g 1,750 -

x

g. A

- 1,500 -

3* 1,250 -

.. I

a. -., -

e -

'.N.'.

B 1,000 -

i5 ,'.~...

llE 750 -

.'-J..

8 ft-lb loss at 300 rpm 500 -

' C s's - -

' V'-

t m.

250 K, s

' ' ' ' ' ~~ " ' ' -

0 ' ' '

O 5 10 15 20 25 30 35 40 45 50 55 Motor torque (ft Ib) w aan Figure 10. The temperature increase from ambient to 250*F reduced output torque by 8 ft-lb at high loads.

1500 1,500 SMB-160 ac " * " " ' " nennibe emeanney omiculesan PAout erHency calculenan

_ 1.000 -

e 1,000 E I d' E 500 E

S -

g 5e 5 a < -

I 500 k

~

O N

A Ster

-500 0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 0 10 20 30 40 50 60 70 Time (s) Ww tonte M-M 150 SMB-1-60 ac 100 -

I f5 sten -

I  :

g 50 - .

B .

b 0 -N t

y , , , , , , ..a. , , , ,

1.0 2.0 3.0 4.0 SOI 6.0 7.0

• Time (s) cc ms ,

Figure 11. Gearbox emeiency was calculated from actual operator torque (valve stem torque) and motor torque measurements.

4

I 1,500 . . . . . .

SMB-b60 ac motor Running efficiency calculation Pullout officioney calculation 1,000 -

~

t -

i 8

e-  :

8 \

90 %

{' volt 3o 500 -

1 4 aos volt I 70%

volt 60 %

volt A Stall 0 ' ' '

O 10 20 30 40 50 60 70 Motor torque (ft-Ib) ==

Figure 12. For each of the SMB-I 60 ac actuator reduced voltage tests, gearbox efliciency drops as loads increase and motor speed drops.

250 . . . . . .

i SMB-00-5 ac l

, Running officioney calculation ,

g Pullout officioney calculation M%

} volt 5:s 150 -

E 90 %

S volt 6

@ 100 -

~

80 %

3 volt 70 %

50 _

volt 60 %

volt A Stall 0

O 1 2 3 4 5 6 7 8 Motor torque (ft-Ib) ==

Figure 13. For each of the SMB-00-5 ac actuator reduced voltage tests, gearbox efficiency was below pullout efficiency.

.. . - . - . ~ . - . . . -~ - - . - - . . -. . . -

1,000 . . .

SMB-0-25 ac OAR = 69.56 Running eHWncy calculath

_ 750 _

E Pullout officiency cakuiation 100a/,

6 --

volt u - .

e 3 90 % i 500 voit f . .

5 80 %

g voit t 70%

< 250 -

60 %

volt A Stall 0 '

O 10 20 30 40 Motor torque (ft-lb) c. .

Figure 14. For each of the SMB-0-25 ac actuator reduced voltage tests, gearbox emeiency drops  ;

as loads increase and motor speed drops.  ;

1,500 . . . . . .

SB-1-40 ac Running efficiency calculation

_ Pullout efficiency calculation e

e 1,000 -

e 100 %

3 - ~ volt 90%

V II

$ ff R 500 -

1 80% -

Y ,

, 70%

volt 80 %

volt A Stall 0 ' ' ' '

O 10 20 30 40 50 60 70 Motor torque (ft-lb) w a.:

Figure 15. For each of the high speed SB-1-40 ac actuator reduced voltage tests, gearbox emciency drops as loads increase and motor speed drops.

250 , , , , , ,

SMB-00-5 ac (0.44 ft-lb offset)

Running eMNcy calculation 200 ,

Pullout officiency calculation y volt

~

e 150 -

~

m 90 %

volt '

100 -

g3 E volt Calculation based on

< measured motor torque minus 0.44 ft-lb

_ 70%

volt 60 %

volt A Stall 0 ' '

O 1 2 3 4 5 6 7 8 Motor torque (ft-Ib) = "a '

Figure 16. Hotelload should be considered when determinmg the actuator capability for smaller motors.

900 , , , , , , , , , ,

SMB-140 dc 800 -

Running officiency calculation 700 -

Pullout efficiency calculation -

600 - 5 ]ogg .

g 8 500 -

90 % _

g 80 %

{ 400 -

7og -

300 - .

10% _

< 50%

200 -

100 -

A stall 0 ' ' ' ' ' '

0 5 10 15 20 25 30 35 40 45 50 55 Motor torque (ft-lb) = "a s Figure 17. Actuators with de motors are particularly sensitive to reductions in efficiency at high loads.

l l

l

500 . . . . . .

SMB-0-25 ac - high temperature test [

OAR = 34.06 0*F 400 -

Running efficiency calculation -

E E Pullout efficiency calculation e 300 -

~

s E

_O ha 200 -

15 100 -

A stali 0 ' ' ' '

O 5 10 15 20 25 30 35 Motor torque (ft-Ib)

Figure 18. Gearbox efliciency is not affected by elevated temp atum, faster gear sets improve actuator gearbox efficiency.

Fifth NRC/ASFE Sypositm on Valve and Ptsp Testing Status of Stellite 6 Friction Testing L John C. Watkins and Kevin G. DeWall l Idaho National Engineering and Environmental Laboratory l

Gerald H. Weidenhamer l U.S. Nuclear Regulatory Commission a -

ABSTRACT l l

For the past several years, researchers at the Idaho National Engineering and Environmental Laboratory, under the sponsorship of the U.S. Nuclear Regulatory Commission, Office of Nuclear Regulatory Research, have been investigating the performance of motor-operated valves subjected to design basis flow l and pressure loads. Part of this research addresses the friction that occurs at the interface between the valve disc and the valve body seats during operation of a gate valve. In most gate valves, these surfaces are hardfaced with Stellite 6, a cobalt-based alloy.

Analytical methods exist for predicting the thrust needed to operate these valves at specific pressure conditions. To produce accurate valve thrust predictions, the analyst must have a reasonably accurate, though conservative, estimate of the coefficient of friction at the dise-to-seat interface. One of the questions that remains to be answered is whether, and to what extent, aging of the disc and seat surfaces affects the disc-to-seat coefficient of friction. Specifically, does the environment in a nuclear plant piping system cause the accumulation of an oxide filtn on these surfaces that increases the coefficient of friction; and if so, how great is the increase?

This paper presents results of specimen tests addressing this issue, with emphasis on the following:

The characteristics and thickness of the oxide film that develops on Stellite 6 as it ages The change in the friction coefficient of Stellite 6 as it ages, including the question of whether the friction coefTicient eventually reaches a plateau The effect in-service cycling has on the characteristics and thickness of the oxide film and on the friction coeflicient.

INTRODUCTION The Idaho National Engineering and Environmental Laboratory (INEEL) has been investigating the ability of motor-operated valves (MOVs) to function when subjected to design basis loads. Methods exist to analytically predict the thrust needed to operate these valves at specific fluid conditions; however, the analyst must have a reasonably accurate, though conservative, estimate of the coefficient of friction at the disc-to-seat interface (see Figure 1). In most gate valves, these surfaces are hardfaced with Stellite 6, a cobalt-based alloy. One of the questions that has not been addressed is whether, and to what extent, aging of the disc and seat surfaces affects the disc-to-seat coeflicient of friction. Specifically, does the accumulation of an oxide film on these surfaces during long term operation in harsh environments increase the coefficient of friction; and if so, how much?

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

l This paper presents the latest results from an ongoing INEFL research project addressing this issue.

l The purpose of this project is to determine how aging degradation mechanisms can affect the performance requirements of MOVs over the long term. The results of this project will provide information on the coefficient of friction expected of a fully aged valve, including the effect of periodic valve cycling typical of in-service tests. The project consisted of subjecting Stellite 6 specimens to an environment simulating boiling water reactor (BWR) coolant conditions, representing the conditions a typical reactor water clean-up (RWCU) isolation valve would experience, and inducing the accumulation of an oxide film. The project

~

included analysis of the resulting oxide film and testing of the specimens to deternune the coeffieie'nt of ~

l friction. The Battelle Memorial Institute in Columbus, Ohio performed the testing under a contract with the INEEL.

l -

TEST PROGRAM This paper describes the results of two types of aging tests, both dealing with Stellite 6. The first investigated natural aging at simulated BWR conditions. For these tests, we determined the oxide film i

thickness and composition following 2.,10 ,20 ,25 ,40 ,50 , and 78-day exposure periods to BWR conditions simulated in a corrosion autoclave. We performed friction testing'on specimens after 2,10,20, 40, and 78-days exposure. The second test type also investigated natural aging but included the effect of periodic cycling on the valve oxide films. The purpose of these later tests was to determine whether the film thickness, composition, and friction coeflicient are influenced by the periodic disc wedging encountered I

during in-service testing (IST). For these tests, we determined the oxide film tinckness and composition both before and after the simulated valve wedging after 25 days and again after 50 days in the corrosion autoclave.

Friction testing was performed after 78 days in the corrosion autoclave.

The specimens were naturally aged at simulated BWR coolant conditions in a corrosion autoclave.

To simulate BWR coolant conditions, the autoclave was attached to a reservoir of water in which the oxygen was controlled in the range of 100 to 200 ppb. Water from the reservoir was continuously supplied to the autoclave, with the temperature in the autoclave maintained at 288'C [550*F], and the pressure at 7.24 MPa

[1050 psi], slightly above the saturated steam pressure, such that the water was slightly subcooled.

We used a friction autoclave to perform the friction tests. The friction autoclave, like the corrosion autoclave, was attached to a reservoir of water in which the oxygen was controlled in the range of 100 to 200 ppb. Testing was performed with the autoclave heated to 288'C [550*F] and pressurized to about 7.24 MPa [1050 psi]. During each test, a specimen assembly consisting of two smaller 6-mm. x 28-mm.

[0.25-in. x 1.10-in.] outer specimens and two larger 13-mm. x 76-mm. [0.5-in. x 3.0-in.] inner specimens was

! tested (see Figure 2). The outer specimens weie held in a stationary fixture, and the inner specimens were i attached to a carrier bar connected to a movable pull rod. Actuation of the pull rod caused the inner specimens to slide along the outer specimens at a relative velocity of approximately 400 mmimin

[16 inimin], a rate within the range expected for typical gate valve operation.

l The friction autoclave is equipped with a bellows that can exert a force on one side of the specimen assembly. Pressurizing the bellows imposes a normal force on the specimens to produce the specified nominal contact stress of 69 MPa [10 ksi]. The normal force required to achieve a 69-MPa [10-ksi] nominal

, contact stress on the 6-mm. x 28-mm. [0.25-in. x 1.10-in.] contact zone is 1250 Kg [2750 lb]. This value was selected to approximate the stress level occurring during operation of typical RWCU system valves.

(Assuming uniform load distribution, the calculated contact stresses at the seats for typical 100 and 150-mm.

. [4 and 6-in.] valves under a difTerential pressure of 7.24 MPa [1,050 psi] are 54 and 87 MPa [7.8 and 12.6 ksi], respectively.)

. . , . _ _ , y

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

0 For the simulated valve wedging representing IST cycling, we used an in-service testing simulation rig. Unlike conditions in both the corrosion autoclave and the friction autoclave, conditions in the in-service testing simulation rig' consisted of a water bath at room temperature and atmospheric pressure rather than i simulated BWR coolant conditions. To subject the entire surface ofeach specimen to a simulated valve '

wedging, we placed like sized specimens with their Stellite 6 surfaces face to face, applied a normal load of 133 MPa [20 ksi], and moved the specimens I mm [0.040 in.] relative to each other. We selected a stress level of 138 MPa [20 ksi] to approximate the bearing stress occurnng during c valve wedging cycle of a typical RWCU system valve. We considered using the high temperature friction autoclave for thcJsimulated * ~

cycling, but it could not achieve the required contact stress on the lars 13-mm. x 76-mm. [0.5-in. x 3.0-in.]

specimens. A normal load of 13,600 Kg [30,000 lb] was reqmred, well in excess of the capabilities of the friction autoclave.

l FILM CHARACTERIZATION The chemical composition of the oxide films was analyzed by Auger electron spectroscopy (AES) for i some specimens and by X-ray photoelectron spectroscopy (XPS) for others (and both methods for a few l specimens). With AES and XPS, the film is incrementally sputtered away, and the elemental compositions of l planes in the oxide film are measured. The results of these analyses are then plotted versus time (or depth, assuming a sputtering rate) to provide a depth profile so that the relative elemental concentrations can bc

evaluated. Figure 3 shows a typical AES depth profile for Stellite 6 after exposure to natural aging conditions. As can be seen from the plot, the chromium concentration is almost constant through the oxide film, whereas the cobalt is lean at the surface and rapidly rises between 1000 and 2000 A. [One Angstrom l (A) equals one ten-millionth of a millimeter.]

l The thickness of the oxide film was determined from XPS depth profiles and is defined as the point where the cobalt concentration, which is present in the base Stellite 6 material but greatly depleted in the oxide film, and the oxygen concentration, which is present in the oxide film but much lower in the base Stellite 6 material, are equal. Figure 4 is a plot of the oxide film thickness for the naturally aged specimens i versus time, along with a best fit of the data, indicating that the oxide film growth rate follows a parabolic

relationship.

We performed additional tests to determine whether the oxide film thickness changes following

simulated valve wedging, and if so, how additional exposure to BWR conditions affects the growth of the oxide film. Figure 5 presents the results of this testing and shows the effect a simulated valve wedging cycle has on the long-term growth characteristics of the oxide film. The oxide film thickness before an IST wedging cycle (boxes) is conipared to the oxids film thickness after an IST wedging cycle (inverted triangles); included for comparison is the data fit of the naturally aged oxide film thickness. The resulting ,

film thicknesses after 25 days and after 50 days appear to follow the general thickness trend (within the  !

observed data scatter) observed for specimens that were allowed to age without being disturbed.

FRICTION TESTING Five sets of naturally aged specimens underwent friction tests. The aging times for these specimen

l. sets were 2,10,20,40, and 78 days. This testing showed that the coeflicient of friction continually increased

! as the specimens aged and as the film thickness increased. In fact, Figure 6 indicates that the friction not only

increases as the specimens age, but that it does not appear to reach a plateau. These results question whether j the friction coefficient and the film thickness will reach stable values as the specimens continue to age. This l

d information is important, because as the friction increases, the thrust demands of a valve will increase and influence the available operating margin of the MOV. The friction tests were performed with the specimens in an autoclave at 288'C [550*F] and 7.24 MPa [1050 psi]. As such, the continuously increasing friction is the result of the oxide thickness increasing and is not due to the preconditioning phenomenon that is encountered when friction testing is performed on specimens in an ambient environment.

Figure 7 presents the same friction information in a difTerent format to show how the coeflicient of friction responds to continued stroking of the same specimen. The results show that the frictiordiehighest ~

during the first stroke and decreases with each subsequent stroke. Even though additional stroking of the specimens decreases the friction, the friction generally increased as the specimen aged. This continual increase in friction over time indicates the importance of trending the valve friction over time. The data also shows that the highest friction occurs during the first stroke and decreases with each subsequent stroke. The first stroke a valve experiences after it has been allowed to age and establish an oxide film will result in the highest coeflicient of friction and therefore require the highest stem thrust to overcome these friction forces.

The increase in friction as specimens age is due to the mechanical properties of the films (oxides and hydroxides) developed during aging. As the film thickness increases, the friction coefficient would be expected to approach that which might be measured using specimens of the same composition as the oxide film,i c., solid oxide specimens. Oxides typically have higher friction coefficients than their metal counterparts. However, for the natural aging cases investigated thus far, the oxide films were extremely thin.

It is most likely that under the relatively high contact stresses of the experiments (69 MPa [10 ksi]), these oxide films were immediately ruptured upon the onset of sliding and subsequently were mixed into the substrate surface. This is typical behavior of a thin hard coating (such as the oxide fihit of Stellite 6) on a relatively softer substrate. As such, the friction coefficient represents a mixture of the friction of the bulk Stellite 6 material and the friction of the oxide film.

One set of specimens subject to natural aging conditions and simulated IST valve wedging cycles underwent friction tests. The aging time for these specimens was 78 days, with simulated valve wedging cycles after 25 days and again after 50 days. Figure 8 presents the results of this testing and shows that specimens that were subjected to simulated valve wedging cycles had a lower coeflicient of friction, although it is not clear whether this trend will continue as the specimens age. During subsequent strokes, the effect of the simulated valve wedging on the resulting friction coefficient was either negligible or varied from stroke to stroke. This frictional behavior is influenced by changes in the condition of the surface due to previous strokes. As such, only the first stroke would be strongly infbenced by the simulated valve wedging.

CONCLUSIONS This paper presents the latest results of an ongoing INEEL research project addressing the aging of Stellite 6 specimens at simulated BWR conditions. This project identified a number of trends in the frictional characteristics of Stellite 6 that may influence the available ope;ator margin of a MOV as it ages. These trends are summarized below.

A data fit of the oxide film thickness versus exposure time for naturally aged Stellite 6 is parabolic.

The testing also demonstrates that the oxide film is altered during the simulated valve wedging, but appears to follow the general thickness trend observed for specimens that were allowed to age without being disturbed.

4 R

e- - - - - . -

The friction of the naturally aged specimens shows a continual increase as the aging time increases, with no evidence of reaching a plateau after 78 days of aging. This information is important because as the friction increases, the thrust demands of a valve will also increase and influence the available operating margin of the MOV.

The friction is highest during the first stroke and decreases with additional stroking. This continual increase in friction over time indicates the importance of trending the valve friction over time. The data also show that the highest friction occurs during the first stroke and decreases with each subsequent steoke. The - 4 l- first stroke a valve experiences after it has been allowed to age and establish an oxide film will result in the highest coefficient of friction and therefore will require the highest stem thrust to successfully operate the valve.

The single data point for the periodic valve wedging test suggests that periodic disc wedging will decrease the expected friction compared to a valve that is less frequently cycled. Although there is only a single data point and the effects of the IST cycle may be small for this case, the data suggest that increasing the time between IST cycling may allow the friction to increase enough to impact the available operating margin of an MOV.

1

~

g

[

Electric s. l p motor ,

_ 1 umitorque operator

/lh .l_q -

, i _ .

l Yoke I -

uve4oaded

• Seat ring packing system ,

""y _ / -7 ~' av s l g

3'i hilj gg l0

/ .

) \ o\

)

vriivo em - c ,

"'N j 0 #

_- $ E .,.

Disc Guido Figure 1 Diagram of a typical motor-operated gate valve showing the main components.

. w.* e,=*,

w TersucMe Thermeamsde Carumair

_ N F..

• W , m.

r

,c ,

a.m gg .

] q:

%,g;;;-  :

7 s c n,.

%- o *=

- ~; -,."*,,,,,

M- _. _ w Q '

ll%;' -- 1 .x N. a,

---

.A \.NEl' \%, O..

".t**"".

l p

\g w

/{

~,

s '- - %s~

l - L/ -

i Figure 2 Diagram of the friction autoclave.

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

l l

l '

l 1 00 . . . . ,

100 m x 100pm area -

80 -

s,,

a -

f#

i cea cet U

40-

, et cra o.y p~ " ~- "'~ ~ # ~ w n w .s n a 2a:- .

Q 01 .

es et 0 10'90 2000 3600 4600 ' b600 6d00 70'0 0 ' 8000 Depth, Angstroms ..

Figure 3 Typical AES elemental concentration versus depth for Stellite 6 after exposure to natural aging conditions.

2000 D

1500

$ o o

@1000 o x

.9 500 o 0

0 10 20 30 40 50 60 70 80 Timo (d) c., ,,3, 3 Figure 4 Film thickness versus time showing the data from the natural aging tests and the parabolic data fit.

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

l l

2000 1

1500 IST

/  : -

1000 '

f IST 500 0

0 10 20 30 40 50 60 70 80 Time (d) c., ,,, ,

Figure 5 Film thickness versus time showing the data from the simulated valve wedging tests and the parabclic data fit of the naturally aged tests.

0.50 0.45 -

0.40 -

Yi c 0.35 '

,g E

0.30 e Stroke 1 0.25 a Stroke 2

• Stroke 3

+ Stroke 4

+ Stroke 5 0.20 0 10 20 30 40 50 60 70 80 Time (d) c. .

Figure 6 Coefficient of friction versus time for naturally aged specimens l

l

l l

I i

1 t 0.6 l x 2 day

+ 10 day

+ 20 day

'0.5 -

a 40 day n a 78 day l

n a a f"0.4 a a a -

a .

g 1 4 a "

E 0.3

.g

+

i

• a 3 o y x g g * * *

+

+

U 0.2 -

x i

l l

0.1 0

0 1 2 3 4 5 6 7 8 9 ib Stroke number c..

Figure 7 Coefficient of frictie versus stroke for naturally aged specimens.

0.6 0 78 day w/o IST x 78 dayw/IST 0.5 -

a a

o 0.4 -

a E

• x

• E

= ,

o x c 0.3 -

, E o

.e o

E 0.2 -

l

\

l 0.1 l

l Q _ ..

l 0 1 2 3 4 5 6 7 8 9 TD Stroke number c1 Figure 8 Coefficient of friction versus stroke for naturally aged specimens and specimens subjected to a simulated IST.

l

J l

I. ~

!, MOV Actuator Motor l.

and Gearbox Testing i

Kevin G. DeWall

! John C. Watkins l USNRC Technical Monitor

Dr. G. H. Weidenhamer i

Presented at l Joint AOV/MOV

User's Group Meeting i Sheraton Sand Key

! Clearwater Beach, FL I N F_ F_ I _

"^""^""^"~~'"*"'"~""'"'^"^~^~'

December 4,1997 I

=

CSF 100s 3

l i

I Introduction l

l l . During the past several years, the NRC has supported j research at the INEEL addressing the performance of l MOVs. The research included tests and analyses to l determine the capability of MOVs to close or open at l their design basis conditions.

1 l . This presentation shows the results of tests evaluating i the operation of the MOV motors and gearboxes at i

typical design basis conditions, including degraded .

l; voltage anc elevated temaerature.

4 c., m. :

j 4

1 i

! l Outline

. Scope 1

l

. Test Equipment

! . Results - ac Motor Tests .

1 l . Results - dc Motor Tests

. Results - Gearbox Efficiency i

co ma -

i__________..________.__________.- _ . _ _ _ _ _ . - _ _ _ _ _ . _ _ _ _ _ _ _ _ _ . _ _ _ _ _ _ _ _ . _ _ _ _ _

l s . .

l  :

Scope J .

h 1 The testing focused on the capability of the electric motor and the efficiency of the actuator gearbox. It specifically addressed the l

following questions:

i

How does actual output torque of the motors compare with the j published torque?

)

How much does the motor output torque decrease at reduced l

voltage, and how does this compare with predictions?

How much does the motor output torque decrease at elevated temperature, and how does this compare with predictions?

! What is the actual efficiency of the actuator gearbox, especially l at high loadings and elevated temperatures? How does this compare with published values?

l

? . . . . .

Test Hardware

'l SMB-00-5ac SMB-0-25ac SMB-0-25ac SMB-1-60ac SB-1-40ac SMB-1-40dc Rated starting torque (ft-Ib) 5 25 25 60 40 40 Specified stall torque (ft-Ib) 6.5 29.5 29.5 66.0 49.0 63.0 Rated speed (rpm) 1700 1700 1700 1700 3400 1900 Rated voltage 460 vac 34 460 vac 34 460 vac 34 460 vac 3$ 460 vac 34 125 vdc Overall ratio 87.8 69.56 34.96 42.50 32.13 42.50 Worm gear ratio 45 to 1 37 to 1 37 to 1 34 to .1 34 to 1 34 to 1 Helical gear set 22/43 25/47 37/35 32/40 37/35 32/40 I

Running efficiency 0.50 0.50 0.55 0.50 0.60 0.50 Pullout efficiency 0.40 0.40 0.40 0.40 0.45 0.40 Stall efficiency 0.50 0.50 0.55 0.50 0.60 0.50 l

Application factor 0.90 0.90 0.90 0.90 0.90 0.90 Motor frame K56 P56 P56 FE56 184R2 D202G e.. . ;

Test Conditions

)

{

! . Most actuator and motor testing is typically performed by applying a sudden torque load or with a locked stem.

. . The MOVLS provided a dynamometer-type test that imposed a gradually increasing valve stem load until it caused the motor to stall.

- Simulates closing the valve against a high flow load.

. Tests performed on each motor:

, Baseline tests - 100% voltage,70 F l Reduced voltage tests - 60,70,80, and 90% voltage Locked rotor tests - 60, 70, 80, 90, and 100% voltage Elevated temperature tests - 80 and 100% voltage 400,150,200,250,300 F

. One gearbox was tested at 70 and 300 F.

CSF n042 2 '

l l Test Results

! 1 l . This research addresses the-terms in the typical formula l for predicting output torque for a valve actuator.

i output motor temp app gearbox l

i where .

Too,u, = output torque of the valve actuator l Tm oto, = rated starting torque of the electric motor V,oc

= actual voltage supplied to the motor V,,, = the motor's rated voltage i n = 2 for ac motors,1 for dc motor l F,,m, = factor to account for losses due to motor heating F,,, = application factor .

Eff, . x = gearbox efficiency OAR. = overall gear ratio

r Results - ac Motor, Tests l

! The 25 ft-Ib ac motor produced more torque at the

. knee (1)000 rpm) than predicted by the manufacturer.

2,800 . . . . . . . . 28 25 ft-Ib ac motor ,

2,400 -

• End of manufacturer's curve j I -

24 X Data points from earlier testing /

l Extrapolated /

_ 2,000 -

/ -

20 g E / E

$ n v. ~ . Speed / 5

! o 1,600 _ _u '~ / -

16 -

! e e

7 '

u l 1,200 - Manufacturer's data est data / -

12

' o

! 2o /

,4 , 15 2 800 -

8 2 Current

/

400 -

4 i

! 0 0 i 0 4 8 12 16 20 24 28 32 36 Motor torque (ft-Ib) o ,,,,,

i

Results - ac Motor. Tests The 60 ft-lb ac motor's speed / torque curves are different at lower l

loads but the torque at the knee (1000 rpm) is about the same.

2,000 . . . . . ><-

50 l 60 ft-lb ac motor 40 m A

' E E / E a x D -Manufacturer's data Test data /

30 y

V / c l 1,000 -

e $

) a -

' a

' 8 Jf' -

20 E y '

/ 2 2 2 '

h i 500 -

II

• End of manufacturer's curve 6

--- Extrapolated O 0 0 -

10 20 30 40 50 60 70 Motor torque (ft-lb)

CS7 0169 1

j r j Results - ac Motor Tests

~

I l Degra)ded Voltage Testing of ac Motors i

i 2

i Ve l act rat V,,1 i

! where

!. T,c1 = actual torque l T,,, = rated starting torque

. Voc, = actual voltage

! V,,, = rated voltage i

c., , ,

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

i j r l Results - ac Motor. Tests i

, For the 40 ft-Ib ac motor, the voltage squared calculation overbstimates the actual torque at the knee (2000 rpm).

4,000 '. . . . . .

4 High speed 40 ft-Ib ac motor, ambient temperature 3,500 -

~~

-, A Calculated value

! 3,000 -

j l E i

e 2,500 -

+

o 2,000 -

A A A A -

h 1,500 -

1,000 - -

500 - -

60% volt 70% volt / 80% voit 90% volt 100% volt 0 '

a 'A -

'A' a '

A - ' ' -

, 0 10 20 30 40 50 60 70 Motor torque (ft-lb) c......

l 1

1 i ,

1 Results - ac Motor Tests I

An exponent of 2.5 is a better I

I

' bound of the INEEL data.

i i

j 0.5 . . . .

i.... ....

....i....

i ,_ Motors at " knee" of the curve  :

I I 3.0 -

j _ _

! C

+

o 2.5 --

o -

8 o 9 E e  : A A @  :

O O O O 4

= 2.0 - -

, E -

. e -

i cn . .

y 1.5 - -

!  : O 5 ft-Ib ac  :

O 25 ft-Ib ac -

1.0 -

A 60 ft-lb ac o 40 ft-Ib ac -

0.5 40 50 60 70 80 90 100

! Voltage ratio (percent of nominal) 4 Oi W 4 :

, - - , . ~ , . , . ~ , .,,,-,,-,w, . . - , . + ,,,v_,~ ,m.4.,-, .. ,,_-,~_, . . - . . . , , , , _ _ , , - , , , _ . , , , . _ , - - , , + - -

-w*- -- - - . . _ . . .

i i

l ,

, Results - ac Motor Tests

! For the.25 ft-Ib ac motor, locked rotor startup torques were the same as dynamo 3ter-type stall torques.

2,000 . . . . . .

I 25 ft-Ib ac motor, am' Mnt temperature A Locked rotor starting test -

i 1,750 -

1,500 -

g j 1,250 -

4 E -

g 1,000 -

m i

i ' 100% volt 2 750 -

y 60% volt 70% volt 80% v 90% volt 500 -

i -

I 250 -

0 0' 5 10 15 20 25 30 35 Motor torque (ft-lb) ,,,,,,

i

Results - ac Motor Tests The combined effects of low voltage and elevated temperature amourits to about half of the 5 ft-lb ac motor's capability.

2,000 . . . . . .

5 ne ac motor _

1,800 -

1 1,600 1,400

~

80% vol 80 g 1,200 \\\\

3 g,

$ 150*F 1,000

\\ \

C 200*F 100 4 voltage

[ 800 - 250*F _

80* F  :-

300'F >

l h 100* F >

2 600 - ) 150*F >-

J 200*F s 400 250*F >

200 -

300*F + _

0 O 1 2 3 4 5 6 7 Motor torque (ft-Ib) . . . , , ,

?

l Results - ac Motor Tests The predicted torque losses due to elevated temperature were equal to or greater than the actual losses for 3 of 4 motors.

7 45 5 fte trupo' e W wp at im vduge .

• a Atual torque at 80% votage

- Preh .

49 .

6 .

W I 6 35

. g . usuai =9ue ano0% vaiage a5 25 E

2

• "'"*8 8 'que at a v aage

- Pseechan 5

"o 5N O

2 2 4

4 25 a g _

A a a e 40 ft ta rnotor 3 20 50 100 150 200 250 300 350 50 100 150 200 250 300 Temperature (*F) Temperature ('F) 35 . . 65 e klualIorque at 100% voltage e Atual torque at 100% voltage a Atual torqis at 80% voRage 60 -

a kaual torque.at 80% vonage

~ -

30 .

.i 55 E

• E i 0 -

25 o 3 $

E E 45 9 2 g 20 3 40 y *-

. j 15 a ,

30

> 25 ste snosos so ne anoior 10 25 50 100 150 200 250 300 50 100 150 200 250 300 Temperature (*F)

Te.rs::rature ("F) C97 0175

Results - ac Motor Tests 1

i s

Conclusions - ac Motors 1

Performance curves i

. The load threshold at which the motor drops to stall

! (knee of the curve) does not always occur at the speed l or load indicated in the published data.

. Actual torque output at the knee of the curve is consistently higher than indicated by manufacturer's curves.

! . All four motors exceeded their rated starting torque. .

. The torque at the knee of the curve was greater than or equal to the stall torque for all four motors.

a Dearaded voltaae testina

! . The voltage squared calculation consistently underestimates motor torque losses at degraded voltage conditions. -

. The use of a different exponent (2.5 instead of 2) to calculate torque loss is a better bound of the INEEL data.

CSF SES2 3 .

I Results - ac Motor Tests 1

I Conclusions - ac Motors Locked rotor testina i . The lockec rotor startup torques were lower than the peak l running torques for all four motors.

, . The ocked rotor startup torque orovides a useful indication

. of w7ere a specific motor is with respect to its rating.

1 Elevated temperature testina i . The predicted torque losses due to elevated temperature l were equal to or greater than the actual losses for three of l t7e four motors.

i

. The predictions might not be appropriate for the fourth (25 ft-Ib) motor. This motor is not qualified for nuclear service.

I __.

i 7

Results - dc Motor Tests i During testing of the 40 ft-lb de motor, the actual torque

output wa.s lower than predicted by the manufacturer's curves.

)

4,000 . . . . . 150 s

l 3,500 - -

/

/

i n

l 3,000 n

q 00 g 2,500 _

Speed !C D

2,000 ~ current @

4 km -

a a

U i

> S 1,500 - Test Manufacturer's o data data j j -

50 h 1,000 -

500 -

/ V O 0

0 10 20 30 40 50 60 l Motor torque (ft-lb) car unw :

~

Results - dc Motol Tests l Degra'ded Voltage Testing of dc Motors

v,'

ec act rat V,,,

where l Toc, = actual torque

T,,1 = rated starting torque l Voci = actual voltage j V,,, = rated voltage l

car is.2 s

4 I.

l

} Results - dc Motor Tests 4 .

For the 40 ft-Ib de motor, locked rotor startup torques

! were1 essentially.the same as the stall torques.

l 2,500 . . . . . -

' 40 ft-lb de motor, ambient temperature 2,250 -

A Locked rotor starting test

'UU' 2,000 ~

90%

g 1,750 -

! o. 80 %

! 6 1,500 -

~

i 3 70 %

g 1,250 -

m 60 %

B 1,000 - -

5 50 %

2 750 - -

1 500 - ~

250 -

~

0 'i i' i

• i' -

0 5 10 15 20 25 30 35- 40 45 50 55 Motor torque (ft-lb) c.,,,,,

4

k. .

'  ?

i Results - dc Motor Tests

~

The conventional method for predicting torque loss at

! reduce'd voltage underestimated the actual losses.

i 2,500 . . . . . . . . -. .

40 ft-Ib de motor, ambient temperature i 2,250 -

100% x calculated value 2,000 _

90%

! g 1,750 -

a. 80%

i 6 1,500 -

50 % -

3 70 % $/"

g 1,250 -

80 %

90%

i u) 60 % .

1,000 -

x x x 50% -

B 60 %

o 50 %

70 %

2 750 -

80%

90 %

500 -

x x -

N 250:- -

0 x' x' -x x' -x ' ' .

0 5 10 15 20 25 30 35 40 45 50 55 Motor torque (ft-lb) ,,,,,,

I l . .

i '

Results - dc Motor Tests l

l Exponeqts of 1.3 for locked rotor and 1.9 for 500 rpm l would provide a better bound of the INEEL data.

l 3.5 ,- - . . i -.. .

ii..ii_

1 . 40 ft-Ib de motor O 1000 RPM -

l  : O 500 RPM  :

i 3.0 -

A Locked rotor--

y  : O O i e 2.5 -

O O -

8 O J

e e

Sm 2.0 -

o O

)  ;

a O o o

5 - -

! g 1.5 - -

5 A A A A 'A 1.0 -

0.5 40 50 60 70 80 90 100 -

Voltage ratio (percent of nominal) i CSF Seda e

I l

.i

) Results - dc Motor Tests 4

i Elevated Temperature Testing of de Motors i

i 1

. The manufacturer's dc actuator qualification requires l; the actuator to perform at 340 F.

. For this tem aerature the manufacturer recommends a 1 ft-lb decrease to the ratec. torque of a 40 ft-Ib motor. ,

I . The decrease is greater for larger motors.

4

1 j ,

i Results - dc Motor Tests l Temperature increase from ambient to 250 F reduced

! output torque by 8 to 10 ft-Ib at high loads.

I 2,500 . . . . . . . . . .

40 ft-lb de motor, high temperature Ambient l' 2,250 - ------

100*F g

150'F 2,000 -

\ - 200*F -

)  %, -

- 250* F 1,750 N

^ -

x:,

4 i E

' a.

b 1,500 -

u '::~:..

o m 1,250 -

a.  :.

m ~.:. '

B 1,000 -

' 'ci .

I o \'i ,

2 750 -

~

-) . ~ 8 ft-lb loss at 300 rpm 500 -

t.

s

~ rN':n ,

u .,

y 250 -

t 0 0 5 10 15 20 25 30 35 40 45 50 55 Motor torque (ft Ib) c,, ,,,

Results - dc Motor Tests -

> Conclusions - de Motors Performance curves

. The actual torque output for our dc motor was lower than that predicted by the manufacturer's curves.

4 i

Dearaded voltaae testina

. The linear method for preclicting torque loss at degraded voltage underestimates the actual torque losses.

4 i . T1e use of a different exponent (1.3 instead of 1) to calculate torc ue loss would provide a better bound for the INEEL data.

4 cae mu s

.-, - ~,

Results - dc Motor Tests 1 Conclusions - dc Motors Locked rotor testina

. The locked rotor startup torques were essentially the same as the stall torques.

Elevated temperature testina

. The adjustment used to account for torque losses due to elevated temperature under-

estimated the actual losses for our dc motor.

- Predicted was 1 ft-lb at 340 F

- Actual was 8 't-lbs at 250 F

)

Results - Gearbox' Efficiency 9

Actuator Gearbox Efficiency J

. The actuator output torque is typically calculated by:

Tout,oi =Tinya1 ( E ffgee ,3cx MN where Too,go, = output torque Tingo, = input torque (motor torque after adjustments described earlier)

Effse ==x = gearbox efficiency OAR = overall gear ratio i

C37 243 ?,

Results - Gearbox Efficiency 1

Actuat6r Gearbox Efficiency (continued) l Typical gearbox efficiencies are referred to as

! Running Efficiency - the efficiency of the gearbox

at normal motor speed and normal loads.

Pullout Efficiency - the efficiency of the gearbox

when the motor is Iugging at very low speeds or starting up against a load.

1 1

Results - Gearbox Efficiency i

1500 . . . . . . . . . . . . . 1.500 . . . . . .

man ac not ' Fiunrdng efficiency calculation SMB-1-60 y Pullout effg:iency calculation 1,000 -

8""

l h 1,000 - -

j 100 %

! @ $ voit a- a j j 500 -

5 i N N s -

e 500 - -

< 0 t

A Stall 1.0 2.0 3.0 4.0 5.0 6.0 7.0 0 10 20 30 40 50 60 70 Time (s) Motor torque (ft-Ib) j 150 .

1 SMB-1-60 ac l Gearbox efficiency 100 -

was calculated i Stau from actual operator i 50 -

torque (valve stem ,i

torque) and motor 0 +

l' torque measurements. -

-50 ' ' ' ' ' ' ' ' ' ' ' ' -

1.0 2.0 3.0 4.0 5.0 6.0 7.0 Time (s) y v - --w-- y ..  %,y e.- - - . - , - ., - .. . , + - w- -

g e--

Results - Gearbox Efficiency For each of the SMB-1-60 ac actuator reduced voltage tests, l gearbox efficiency drops as loads increase and motor speed drops.

1,500 . . . . -

SMB-1-60 ac motor Running efficiency calculation

. Pullout efficiency calculation n

O 1,000 -

g, T t 2' 90 %

volt j 500 -

80 %

~

< volt 70 %

volt 60 %

4 volt A Stall 0 -

0 10 20 30 40 50 60 70 Motor torque (ft-lb) ,,,,,,

?

Results - Gearbox Efficiency For each of.the SMB-00-5 ac actuator reduced voltage tests, gehrbox efficiency was below pullout.-efficiency.

250 . . . . .

y SMB-00-5 ac l

0

""""i"9 *#'I i*" Y ' "' 'i " N Pullout efficiency calculation 100 %

g volt g -

e 150 -

o 90 %

volt 3u ~

j 100 -

80 %

volt

?

y

<C _

70%

50 -

volt 60 % A Stall volt 8

0 3 4 5 6 7 O 1 2 - . . -

Motor torque (ft-Ib)

t Results - Gearbox Efficiency

For eac.h of the SMB-0-25 ac actuator reduced voltage tests, gearbox efficiency drops as loads increase and motor speed drops.

1,000 . . .

SMB-0-25 ac OAR = 69.56 Running efficiency calculation --

750 -

E Pullout efficiency calculation 100%

7

- volt e

3 90 %

7 v it _

B 500 -

E 80 %

_o volt g

i g 70 %

< volt ~

250 -

60 %

volt A Stali 0

0 10 20 30 40 Motor torque (ft-lb) ,,,,,,

Results - Gearbox. Efficiency For each of the high speed SB-1-40 ac actuator reduced voltage tests,

gearbox efficiency drops as loads increase and motor speed drops.

1,500 . . . . . .

SB-1-40 ac e Running efficiency calculation

Pullout efficiency calculation

_a Z 1,000 -

=

c) 100%

volt o

E 90 %

volt i

O cc 3 500 -

, 80 % -

y ^

volt 70 %

volt

/, 60 %

volt A Stall 0 -

0 10 20 30 40 50 60 70

, Motor torque (ft-lb) . .,,,

l l

Results - Gearbox Efficiency Hotel load should be considered when determining the

~1 actuator capability for smaller motors.

250 . . . . . .

SMB-00-5 ac (0.44 ft-lb offset)

Running efHelency calculation %

200 l

Pullout efficiency calculation g 100%

Z volt O

150 e -

90%

S volt u

1 y 100 -

80 %

~

I volt 3 Calculation based on

<C measured motor torque minus 0.44 ft-lb 50 -

volt 60 %

volt A Stall 0

O 1 2 3 4 5 6 7 8 Motor torque (ft-lb) - > =

r Results - Gearbox Efficiency Actuators with de motors are particularly sensitive l to reductions in efficiency at high loads.

900 . . . . . . .. . . .

SMB-1-40 dc 800

~

Running efficiency calculation f

700 Pullout efficiency calculation 0 600

{ t- -

g -

=

100%

e N 90 %

g 500 -

A 80 %

g

^

{% 400 -

70 %

60 %

3O 300 _

<C 50 %

200 - -

100 -

A stall 0

0 5 10 15 20 25 30 35 40 45 50 55 Motor torque (ft-lb) , , , , , ,

1 Results - Gearbox Efficiency Gearbox efficiency is not affected by elevated temperature.

Fastler gear sets improve actuator gearbox efficiency.

500 . . . . . .,

3 SMB-0-25 ac - high temperature test iv 3 ohg-0* F OAR = 34.96 400 -

Running efficiency calculation E

Pullout efficiency calculation

_Z 5 300 -

.9 m '

j 200 -

a

'O 100 -

A Stall 0

0 5 10 15 20 25 30 35 Motor torque (ft-Ib) ="

. r.

Results - Gearbox Efficiency

' Actuator Gearbox Efficiency 3 Conclusions

. The published running efficiency was generally not adequate for predicting gearbox performance, especially at higher loads.

. The published pullout efficiency was adequate for moderate loads, but some actual efficiency data fell below the pullout efficiency.

. Different gear sets operate at different efficiencies, even in the same gearbox.

. Gearbox efficiency is affected by motor speed as well as torque load. Lower motor speed and higher motor torque correspond with lower gearbox efficiency.

. The actuator no-load motor torque (hotel load) can be significant for small motors.

. Elevated gearbox temperature did not affect the actual efficiency of the gearbox.

. . , , , , , , _ - _ . _ _ , . . , _ . _ , , ___,._,_,,,,,,,.._,,..,,.__.___.,,._,.__...,,...-..,,----..m.~- - ~ - . _ _ _

O 4 T

~1 The opinions presented here today are those of the authors and not~necessarily endorsed by our sponsor, the USNRC.

e C331-WHT 197410