ML20081J082

Ro:On 831003,during Full Power,Site Electrical Power Lost & Emergency Generator Operated in Degraded State.Cause Probably Due to Addition of STP Gas Treatment on 830923. Gas Pumped from Generator Tank.Sys Refilled

ML20081J082
Person / Time
Site:	University of Missouri-Columbia
Issue date:	11/01/1983
From:	Alger D, Mckibber J MISSOURI, UNIV. OF, COLUMBIA, MO, MISSOURI, UNIV. OF, ROLLA, MO
To:	Thomas C Office of Nuclear Reactor Regulation
References
NUDOCS 8311080292
Download: ML20081J082 (3)

Text

.

Research Reactor Facility UNIVERSITY OF MISSOURI Research Park Columoia. Missouri 65211 Te!ephone (314) 882-4211 November 1, 1983 Director of Nuclear Reactor Regulation U. S. Nuclear Regulatory Commission Washington, D. C. 20555 Attention: Mr. Cecil 0. Thomas, Chief Standardization & Special Projects Branch

Reference:

Docket 50-186 University of Missouri Research Reactor:

License R-103

Subject:

Difficulty of Emergency Generator to Assume All Emergency Electrical Loads

Description:

At 0154 on October 3,1983, with the reactor operating at full power, site electrical power was lost and the emergency generator started and assumed the full emergency electrical loads but operated at 50 Hz instead of the normal 60 Hz fre-quency. At 0207 the building exhaust fan, whose electrical power is supplied via the emergency electrical buss, was secured to reduce the electrical load. This permitted the motor to increase to normal operating speed and supply electrical power for the remaining emergency electrical loads. Therefore the emergency gen-erator was operable but in a degraded state, since it could not supply all emer-gency electrical loads at the normal 60 Hz frequency.

Analysis:

The emergency generator is a 45 KVA generator powered by a Ford 292 cubic inch V8 industrial gasoline engine with a one barrel carburetor fuel systen, a 0.5 gallon gasoline day tank, and a 120 gallon underground gasoline storage tank.

8311080292 831101 PDR ADOCK 05000186 g S PDR W 4;m mi <

/ oto p COLUMBIA KANSAS CITY ROLLA ST. LOUIS Il an equal opportunity institution

Director of Nuclear Reactor Regulation November 1, 1983 Page 2 After normal electrical power was restored a mechanic from the University services support group was called in to investigate the problem. He checked the spark plugs, and verified the timing and dwell were set properly. The emergency generator is load tested by opening the breaker supplying normal power to the emergency electrical buss in the surveillance procedure labeled CP-17. This test was performed for the mechanic to observe the generator's performance under load and again the generator assumed the loads but operated at 50 Hz and was pinging badly as it had done at 0154 when under full load. The exhaust fan was secured and again the generator increased speed to 60 Hz and ran properly. The mechanic stated he thought the problem was the fuel was too low of an octane rating. The engine was shutdown, and the fuel pumped and drained from the system. The system was then flushed, redrained, and refilled with new regular gasoline after instal-ling new gas filters. Additionally the oil and oil filters were changed. The emergency generator load test was performed at 1225 and the generator operated satisfactory. The reactor was returned to operation at 1416. On the following maintenance day, October 10, 1983; the emergency generator was load tested again and operated satisfactory.

The problem with the fuel on the norning of October 3,1983 cannot be defi-nitely determined since none of it was saved to be tested; however, it was associ-ated with the addition of STP Gas Treatment on September 23, 1983. Due to the emergency generator's Ford engine being of 1959 design, it was designed for

" regular" gasoline of higher octane than is the minimum octane of " regular" gaso-line supplied today. Therefore in an effort to boost the performance three 12 fluid ounce cans of STP Gas Treatment composed of alcohol, mineral spirits and exylene were added to 67 gallons of gasoline that were in the storage tank on September 23, 1983. The mixture was recirculated using the storage tank pump for approximately 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br />; then the generator was run unloaded for 30 minutes to veri-fy operability. The emergency generator was operated twice on September 26, 1983 and again on October 2, 1983. Each time it was operated unloaded for at least 30 minutes with no abnormalities. The first time the generator operated fully loaded after adding the gas treatment was the loss of site power at 0154 on October 3, 1983. The previous load test was performed on September 19, 1983, when it operated properly, prior to having the gas treatment added.

M.F.A. Oil Company Laboratory personnel were consulted concerning the problem with the fuel. They could not give a definitive answer without testing the fuel, but felt the probable cause was the alcohol in the STP treatment combined with condensate in the storage system and was then carried to the fuel filters. They stated that the condensate could then foul the filters reducing the maximum flow rate through the filters to the carburetor. Therefore, the engine via the choke and governor was demanding more gasoline than the system could deliver under the full electrical load condition with the filters fouled, but would perform properly under reduced load. This hypothesis is supported by the Ford Owner's Manual which gives lean fuel mixture as possible cause for " pinging sound when engine is rapid-ly accelerated or heavily loaded".

The analysis for loss of electrical power with the reactor operating at 10MW and the emergency generator failing to start is covered in Hazards Summary Report (HSR), Addendum 5, Section 2.4.1, 2.4.2, and 2.5. In this dual failure mode, the reactor will shutdown if not already shutdown and decay heat removal will be

?

b o

Director of Nuclear Reactor Regulation November 1, 1983 Page 3 performed indefinitely by the in-pool heat exchanger. The control blades are re-leased on loss of electrical current to their electromagnets and drop to the full inserted position by gravity. All process systems are placed in the shutdown con-dition due to the failsafe design of these systems; i.e. the redundant (only one is required) primary in-pool heat exchanger valves 546A and B open by spring actuation placing the convection cooling in-pool heat exhanger in service. This failsafe design of the system permits shutdown decay heat removal with no electri-cal power (Appendix D of Addendum 4 to HSR). The containment building integrity could not be indefinitely guaranteed if facility electrical power and emergency generator were not available, but the reactor would be shutdown and contain-ment would not be required.

Corrective Action:

The possibly contaminated gasoline was pumped from the tank, the system flushed and refilled with new regular gasoline. Emergency Electrical Generator Load Test (CP-17) was performed satisfactory. The preventative maintenance system and the Standard Operating Procedures are being revised to require a load test after any work on the emergency generator except for routine maintenance that has been proven to not cause a problem such as engine oil and filter changes, governor oil checks, new fan belts, and gas tank sediment draining checks.

Sincerely, J. C. McKibben Reactor Manager Endorsement:

Reviewed and Approved

\. g Don M. Alger Associate Director JCMK:vs cc: V. S. Nuclear Regulatory Commission c/o Document Management Branch James Keppler, Regional Administrator Region Operations - Region III Reactor Advisory Committee Reactor Safety Committee John H. Tolan, Radiation Safety Officer