The Sizewell B experience1 June 2010
US research institute EPRI has just finished a major review of the first commercial implementation of OLM.
In December 2009, the USA’s Electric Power Research Institute published a comprehensive three-volume report on data from the commercial application of OLM, at the UK’s Sizewell B (EPRI Technical Report 1019188). The 1200MWe PWR used the technique to extend the calibration intervals of all the pressure and differential pressure transmitters in the primary system and some in the secondary system. It also performs automated cross-calibration of resistance temperature detectors during cool-down and/or heat-up. Sizewell B’s long-term goal is to expand on-line calibration monitoring from the initial set of about 200 transmitters to nearly 2500 transmitters, including many in the secondary side of the plant.
EPRI senior project manager Rick Rusaw, who helped produce the report, recently retired from South Carolina’s V.C. Summer station, where he was principal instrumentation and control engineer. He was originally involved in introducing the first US OLM pilot project in 1988. Rusaw explains how Sizewell B got involved in OLM.
“EPRI had developed a users’ group dedicated to monitoring instrument calibration. For most US plants, OLM would provide a relatively low rate of return on investment because of the lower number of instruments that would be involved. The technique uses statistical averaging, so its value increases as the level of redundancy increases. What was appealing to Sizewell B was that it had a different design, with two to three times as many instruments as an average US power plant [based on UK regulators' requirements for a fully-redundant backup I&C system].
“Sizewell B saw an opportunity and made preparations for taking the technology to the regulators. Just setting up the sequence of events took time to establish to the regulators’ satisfaction. That gave them an opportunity for a phased approach that they could implement in small pieces. The report reflects the Sizewell B experience over three cycles of plant operation, following licence approval in 2003-4. The US NRC granted general approval for the technology in 2000.”
Q: Can we conclude from Sizewell B that OLM is a safe, precise and reliable form of sensor calibration?
“Yes. We took the technology to the U.S. regulators, and they presented us with a long list of issues and concerns that we addressed. While on-line monitoring can’t be used as a direct substitute for calibration, the regulator accepted these methods as a verification of the last field calibration. In introducing a technology such as this, you have to have a thorough understanding of the process and how it affects the current operating licence. The fundamental question is, what are the additional uncertainties of the process? Once you know what they are, you can quantify them to the regulator’s satisfaction and assess how they might interface with operational limits. If you can show they are not detrimental, and there are no increases in risk, then generally the process is acceptable [for the regulator].
“There are often tradeoffs in the development of new technologies. Conventional methods of calibration assessment are conducted on a periodic basis over extended periods of time, and have to rely on conservative assumptions regarding instrument behaviours. With on-line monitoring, equipment performance is monitored continuously, but the monitoring process introduces a minor amount of additional uncertainty. This uncertainty is well-understood and quantifiable, and the additional data collected through the continuous monitoring offsets the uncertainty in the measurement process.
“The additional uncertainty in the case of Sizewell B is extremely small because of the level of redundancy. If you have 10 measurements, and take the average, then you have high confidence that the resulting average is closer to the true value than any one measurement. There are limits, but the smaller the number, the higher the uncertainty.”
Q: What is the minimum number of sensors?
“With the Sizewell B techniques, which are based on redundancy, you need probably three, ideally four redundant sensors. The techniques for a pair of sensors are much more complicated; they have not been reviewed and approved by the regulator and are still at university research level.”
Q: How much interest have you seen among US plants to adopt OLM?
“Interest has picked back up in recent years for two reasons. The NSSS vendors who are building new designs of plants are interested in applying the technology as an option. And we are starting to get interest at high levels of utility management thanks to the savings achieved at Sizewell B. [EPRI estimates operator British Energy saves $5 million per cycle by saving five days of critical path outage time].
“Furthermore, several utilities are involved with NSSS vendors and EPRI to take another look at finding a unified approach to streamline the process.”
Related ArticlesSizewell B outage will extend into summer Listening in real time Listening in real time - the case of Diablo Canyon Sizewell B nears restart