To eliminate unnecessary plant trips, the Wolf Creek Nuclear Generating Station in Kansas is taking a systematic approach to rooting out component vulnerability.
By: Ray Barth, Wolf Creek Nuclear Operating Corp., and Joe Spencer, Framatome ANP
Five plant trips within three years were too many for the Wolf Creek Nuclear Generating Station in Kansas. The 1163 MW (net) Westinghouse pressurized water reactor facility had amassed an excellent operating record over 20 years, at one time holding the industry record for consecutive days of operation, and consistently operating at a high capacity factor.
The causes of the trips – automatic plant shutdowns caused by malfunctioning equipment – were varied. However, guidance on equipment reliability from the Institute of Nuclear Power Operations (INPO) helped Wolf Creek put a name to the problem: Single Point Vulnerability (SPV). A SPV is, by Wolf Creek’s definition, any system or component in a plant whose failure can lead to a reactor trip, a turbine trip or an output power reduction greater than 5 percent.
The nuclear industry has always, and understandably, focused its attention on trips caused by malfunctions in nuclear plant safety systems, those associated with the reactor. These systems must have multiple backup layers to meet regulatory requirements, and therefore, are designed to be single failure proof. No such redundancy protects many electrical, mechanical, and instrumentation and control components in the balance of plant (BOP).
The Wolf Creek Nuclear Power Plant.
As plants age, however, the potential for sudden failure of BOP systems increases. Malfunctions may have minimal safety consequences, but in today’s competitive electricity marketplace, their economic significance has grown.
Programs to increase plant reliability by reducing SPVs have mainly focused on qualitative reviews of critical plant equipment, identifying those most prone to fail. Mitigation has frequently awaited scheduled replacement of obsolete equipment.
Wolf Creek management wanted to take a more proactive approach to identifying and evaluating SPVs. Wolf Creek’s Support Engineering Equipment Reliability personnel worked with a team from Framatome ANP to develop a comprehensive three-phase SPV review strategy and evaluation process.
Phase I identified potential SPV candidates. Phase II evaluated and risk-ranked the identified SPVs. Phase III, development and implementation of mitigation strategies, is under way.
The Phase I review considered information available from the Nuclear Regulatory Commission (NRC), the industry regulator, and from INPO, an industry-wide oversight body. Both provide information on equipment reliability issues to nuclear sites, and their reports provide insights into the industry history of reported SPV events. The review also considered information available from EPRI, the original equipment manufacturers (OEMs), industry working groups, owner’s groups, and information on what other nuclear sites have done with SPV events. Finally, Phase I leveraged information from sister plants, with equipment from the same manufacturer and in similar configurations.
The review began by determining which plant systems covered by the NRC’s Maintenance Rule – a rule focused on performing maintenance on the right structures, systems and components to ensure they are capable of fulfilling their intended functions – contained components that met the SPV definition. In addition, the Framatome ANP team obtained a listing of all Wolf Creek systems containing components with the Production Critical 1 (PC1) classification. Wolf Creek designates those components affecting power production as either PC1 or PC2. PC1 components, by definition, are those that could cause reactor or turbine trips, or power output reduction of more than 5 percent.
The Maintenance Rule and PC1 components represented the universe of systems that could include SPVs at Wolf Creek. The team asked a number of Wolf Creek operations and engineering personnel, along with Framatome ANP personnel knowledgeable about SPVs, for their input on systems that might contain SPVs. The team entered data for the resulting 130 systems and/or components into a spreadsheet for more detailed review in Phase II.
Also going into the spreadsheet was an array of data on all forced outages at Wolf Creek since 1985, plant performance metrics, and SPV-related information gathered from Wolf Creek’s sister plant, the Callaway Nuclear Plant in Missouri.
Other information entered into the spreadsheet included:
- Wolf Creek and Westinghouse entries on the INPO Equipment Performance Information Exchange System and Nuclear Plant Reliability Data System;
- Reports from the North American Electric Reliability Council;
- Data from EPRI and a number of individual site reports, as well as regulatory and oversight body data from the NRC and INPO, along with a review of a number of INPO topical reports.
The Phase I review identified 48 systems that could potentially contain SPVs. During Phase II, the team narrowed the scope to 28 systems with a total of 631 potential SPVs. These SPVs were considered candidates for further review during Phase III. Of the 631 potential SPVs, 449, or 71.2 percent, were classified as PC-1, with the remainder PC-2 or non-critical.
Using a failure modes and effects analysis (FMEA) risk-ranking methodology, the team systematically examined the failure consequences of the SPV candidates by decomposing them down to the component or subcomponent level. They identified the effects of failure in each case, gathering information through such avenues as interviews with system engineers and review of system drawings, descriptions, operating procedures, plant technical specifications and technical manuals.
The 631 SPVs were spread across 31 component types. Electronic cards were by far the most numerous, with 207, or nearly 33 percent. Switch instruments accounted for nearly 10 percent, and breakers nearly 9 percent. Expansion joints were fourth most numerous, with nearly 8 percent.
The team developed a risk matrix to rank each individual potential SPV. The risk matrix correlates the consequences of a particular failure effect with the likelihood of that failure occurring. The matrix contains risk bands, uses qualitative criteria and scales for failure likelihood and consequences, and serves as a risk-significance ranking method for each Phase II SPV identified.
The risk matrix uses four levels to classify the likelihood of occurrence, from high (failure with some regulatory impact) to extremely low (not expected in the plant’s operating lifetime). It also uses four measures of severity of consequences, from catastrophic to marginal. The SPVs are placed in five risk bands. Band 5, the highest, would call for plant design change or development of redundant systems to mitigate the expected consequences. Band 1, the lowest, represents minimal risk and does not require any special handling.
The Phase II review found that the systems with the highest number of SPVs were: feedwater, reactor protection, engineered safety features actuation system (ESFAS) and condensate. At the time, Wolf Creek believed it was too early to state with any confidence what the top three or four SPVs were.
After completing Phase II in December 2004, Wolf Creek assembled a three-person core team. The team included representatives from systems engineering and operations, maintenance, licensing, design engineering and executive management. With the list now at 631 possible SPVs, the team was prepared to hammer through them, according to the risk ranking.
In late June 2005, the Wolf Creek team held the first Mitigating Strategy Review meeting (Phase III), with a senior reactor operator, nuclear station operator, system engineer, and instrumentation and control technician augmenting the core group. The team intentionally did not start with the feedwater system, even though it contained the highest number of potential SPVs. Wolf Creek wanted to gain some experience from the process since it had never done this before.
Instead, the team started with 30 components of the main steam system, comprising four component groups: the main steam header pressure input signal to the main feedwater pump speed control (electronic card), main steam pressure loop (card), main steam isolation valves, and main steam code safety valves.
The team determined that the header pressure input signal card did not meet SPV criteria, since the operator could take action to prevent a trip in the event of card failure. The same held true for the pressure loop cards. However, the team noted that all printed circuit boards (PCBs) with high purity plating were susceptible to “tin whiskers” – needle-like crystalline structures of tin that commonly form on tin and silver electronic components if not coated. Wolf Creek developed a program during a recent refueling outage to remove the tin whiskers – which can cause short circuits – by dusting the boards with paintbrushes.
The team also found that Wolf Creek’s preventive maintenance program is adequate to manage any problems with the main steam code safety valves. Wolf Creek will study another plant’s preventive maintenance program for its similar main steam isolation valves, because that plant is satisfied with their performance and is not looking to replace them.
The team will work one week per month on the reviews, documentation and work packages for each of the SPVs, and expects to complete Phase III by mid-2007.
Ray Barth is the Equipment Reliability Engineer in the Support Engineering Department at Wolf Creek. He has also worked in the Maintenance and System Engineering Departments in his 15 years at Wolf Creek. He began his career more than 24 years ago as a Mechanical Field Engineer involved with constructing three other nuclear plants. Barth holds a bachelor of science degree from Hofstra University.
Joe Spencer is Product Manager of the Framatome ANP Reliability Improvement (RI) Group, where he oversees and conducts consulting services in plant equipment reliability improvement. Prior to this role, Spencer served as the point contact for the Framatome ANP Center for Plant Reliability. He holds a master’s degree in mechanical engineering from the University of Virginia and a bachelor’s degree in mechanical engineering from Virginia Polytechnic Institute and State University.