05000261/FIN-2010009-08
From kanterella
Jump to navigation
Jump to search
Finding | |
|---|---|
| Title | Deficiencies in Non Safety-Related Cable Installation |
| Description | To determine the circumstances surrounding the fault in the cable that led to the first electrical disturbance and subsequent reactor trip, the team performed the following activities: Determined details of the cable construction including conductor size, insulation thickness and material, and type of shielding using the manufacturers data sheet Compared the cable construction to system requirements and standard industry practice Reviewed relevant portions of the plant modification that installed the cable Viewed the site of the cable fault Reviewed cable records to determine where other similar cables are installed in the plant Interviewed engineering staff involved with the electrical distribution systems and components Reviewed the licensees causal analysis for the cable fault Evaluated the licensees proposed corrective actions The cable that faulted did not meet many of the specifications for the design change that installed the cable. This contributed to the cable failure. The cable, manufactured by the Rome Cable Corporation (Rome), was installed in 1986 when 4 kV Bus 5 was installed as an extension of Bus 4 per Plant Design Change Number DCN-851. The cable, identified as C21344A, served as the interconnection between 4 kV Buses 4 and 5 and was comprised of two conductors for each of three phases. The cable was installed in two steel conduits, with each conduit containing all three phases. As noted in Section 2.1, all 4 kV buses at Robinson are non safety-related. The Bill of Materials for DCN-851 indicated that the cable should be in accordance with Standard Specification L2-E-035 for 5,000 Volt Power Cable. However, the Bill of Materials did not indicate a purchase order number for the cable that faulted, as it did for other cables installed by the modification, such as 3/c No. 12 AWG cable. Records reviewed by the team indicated that the cable came from reel number HBR-13505. Differences between Standard Specification L2-E-035 and the actual installed Rome cable are listed below: L2-E-035 called for coated copper conductors. The installed cable had uncoated conductors. L2-E-035 called for all cables to be provided with an outer jacket. The installed cable did not have a jacket. L2-E-035 called for cable insulation and jacketing that was self-extinguishing and non-propagating with regard to fire as described in IEEE 383-1974, Type Test of Class 1E Electric Cables, Field Splices, and Connections for Nuclear Power Generating Stations. The Rome catalogue data made no claim as to fire propagation properties. The event demonstrated that the cable lacked fire propagation properties because 1) the cable ignited following the fault, 2) the cable did not self extinguish after the fault was denergized, and 3) flame was propagated along the cable. L2-E-035 called for 133 percent insulation level and insulation shielding if specified in the purchase order. The installed cable did not have either of these features. The cable consisted of single conductor 500 MCM uncoated copper with 130 mils of cross-linked polyethylene insulation rated for continuous operation at 5 kV and 90 degree Celsius. The insulation thickness was determined from the overall cable diameter and from the licensees measurement of conductor diameter. The manufacturers catalogue information (SPEC 7155 dated January 1, 1991) stated that an insulation thickness of 120 mils is suitable for applications requiring 100 percent insulation levels. However, due to the high-resistance grounding scheme used on the Robinson 4 kV electrical system, an insulation level of 133 percent or 173 percent was required, depending on how long a ground fault could remain on the system. The significance of not having adequate insulation thickness was that, should a single line to ground fault occur the voltage on the two un-faulted phases would exceed the rating of the insulation. The cable did not have a jacket. The significance of not having a jacket was that the cable insulation was more vulnerable to damage during installation. Also, the jacket, if installed, would have provided a buffer between the insulation and grounded metal parts, such as the conduit or bus enclosure. The cable did not have an insulation shield. When an insulation shield is not installed, the electric field will be partly in the insulation and partly in whatever lies between the insulation and ground. This situation could be conducive to corona if a thin layer of air lies between the insulator surface and ground, which can lead to insulation deterioration. IEEE 666-1991, Design Guide for Electric Power Service Systems for Generating Stations, Section 12.3.6 states: Power cables rated 5 kV and over should be equipped with insulation shield. The significance of not having a grounded insulation shield was that voltage stress on the insulation was not symmetrical and uniform around the circumference, but rather greater at points where the insulation contacted a grounded surface, such as a metal conduit, than at other points around the circumference. The following information indicated that shielded cable was originally intended for this cable: 1) Design Change Notice No. 6 to DCN-851 changed the termination detail from one depicting the grounding of shield wires to one with no shield wires, and 2) installation instruction 4.35 directed installation of a stress cone for cable C21344A, which would be needed only for a shielded cable. Cognizant licensee engineers stated that the Rome cable installed as part of the Bus 5 modification was different than other 4 kV cable installed at Robinson and was used only for the Bus 4 to Bus 5 connection and the feeder from Bus 5 to station service transformer 2E. During the inspection, the licensee did not present any documentation explaining or justifying why the installed cable for the Bus 5 modification was different than Standard Specification L2-E-035 and the typical cables installed in the plant. The team reviewed the 4 kV cables connected to Buses 1 through 5, and found this statement to be correct. All the 4 kV cables connected to Buses 1 through 5, except for the two cables mentioned above, met or exceeded standard specification L2-E-035, with at least 133 percent insulation and insulation shield. In addition to construction details of the faulted cable, the team reviewed various design considerations related to the cable. The ampacity of two 500 MCM, 90 degree Celsius, cables installed in conduit in free air is 954 amperes. The team estimated the maximum continuous load on Bus 5 as 493 amperes; 216 amperes for the 1750 HP Circulating Water Pump and 277 amperes for the 2000 kV station service transformer. The overcurrent relays were set at 1000 amperes. Therefore, the cables were not overloaded during normal operation. The conduits were the correct size for the cables installed within them. The number of bends in the conduits did not exceed the recommended maximum number of bends. Therefore, pulling tension limits should not have been exceeded during installation. This did not preclude the possibility that the three single conductors became twisted as the cable was pulled through three 90 degree bends. The licensees Event Review Team (ERT) visually examined the faulted cable and the station service transformer 2E feeder cable and determined the three single conductors were twisted. Twisting of one conductor around the other two conductors could result in jamming of the cables in the conduit since the combined diameter of the twisted cables would be greater than the inside diameter of the conduit. The twisting would have led to excessive pulling force being applied during cable installation. The required pulling force is proportional to the side wall pressure exerted on the cable at a bend. Because of the extensive damage resulting from the length of time the fault was energized, the failure mechanism could not be determined with absolute certainty. The licensees causal analysis determined with a fair degree of certainty that the initial fault occurred at a point where the conduits terminate at the top of Bus 5 switchgear. After consideration of the above facts and review of the licensees causal analysis, the team concluded that the failure mechanism probably involved one or more of the following factors: Degradation of the insulation at the surface of the cable due to corona Damage to the insulation due to inadvertent twisting of the three conductors during the pulling-in process resulting in excessive side-wall pressure at one or more of the three 90 degree bends in the conduit Rubbing of the cable against the conduit or switchgear top plate due to turbine building vibration A secondary fault at the Bus 4 cable compartment for circuit breaker 52/24 was caused by plasma gas migrating inside the conduit and through a hole in the conduit seal, along with terminations that were not taped. The ERT postulated that the hole was caused by pressure built up in the conduit as a result of the fault. The ERT further postulated that this secondary fault at circuit breaker 52/24 created permanent degradation of the insulation at that location. All of the cable within the compartment was completed destroyed when Bus 4 was reenergized about four hours after the initial fault was cleared. The licensee stated that corrective actions related to the cable failure would be to replace the Rome cable feeding station service transformer 2E before plant startup. The licensees Significant Adverse Condition Investigation Report for the event states that a search, using catalog identification numbers, was made across the Progress Energy fleet for this type of cable or similar cable and none was found. The licensee did not believe revisions were needed to the design control process because the process had been changed earlier to preclude the problems described herein, i.e. lack of proper control over purchasing and field changes. The team concluded that the apparent root cause of the initial cable failure and subsequent associated short-circuits was poor quality control over the non-safety-related modification process for installing the cable. A cable of lesser quality than other 5 kV cables installed throughout the plant was installed as a substitute during this modification. The cable terminations were not taped and the cable was not restrained to prevent rubbing. The consequence of the cable fault was a reactor trip. Because of the magnitude of the electrical fault, the reactor trip would have occurred regardless of whether bus tie circuit breaker 52/24 was fully functional. The resultant voltage transient decreased RCP speed which lowed RCS flow and initiated a reactor trip. This occurred faster than the time delay overcurrent protective relays associated with circuit breaker 52/24. Additional review by the NRC will be needed to determine whether the cable installation represents a performance deficiency. An Unresolved Item will be opened pending completion of this review. The issue will be identified as URI 05000261/2010009-08, Deficiencies in Non Safety-Related Cable Installation. |
| Site: | Robinson  |
|---|---|
| Report | IR 05000261/2010009 Section 4OA5 |
| Date counted | Jun 30, 2010 (2010Q2) |
| Type: | URI: |
| cornerstone | Mitigating Systems |
| Identified by: | NRC identified |
| Inspection Procedure: | IP 93800 |
| Inspectors (proximate) | J Hanna J Hickey L Miller L Wert P Fillion P Pieringer R Monk S Currie F Ehrhardta Nielsenc Kontz D Bollock D Mills G Laska G Skinner J Beavers J Brady J Hickey J Worosilo M Bates P Braxton R Musserr Monkf Ehrhardt J Hanna L Miller J Hickey P Fillion L Wert P Pieringer S Currie |
| INPO aspect | |
| ' | |
Finding - Robinson - IR 05000261/2010009 | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Finding List (Robinson) @ 2010Q2
Self-Identified List (Robinson)
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||