Statistical drift control methods using plant specific as-found/as-left data

Abstract

The instrument setpoints are affected by many sources of uncertainty. ISA-RP67.04 (1994) provides the guideline for determining setpoint under several uncertainty terms, in particular, including instrument drift term. The nuclear industry has generally accepted the statistical approach suggested in this document for ensuring that setpoints for safety-related instrumentation are established and maintained within the technical specification limits (U.S. NRC Reg. Guide 1.105, 1999). However, the current methodologies for setpoint drift analysis may be insufficient to manage the setpoint drift of instrumentation devices, because they are basically designed to determine instrument setpoint using statistical prediction techniques. In this paper, we propose a new statistical analysis procedure composed of six steps for the management of the setpoint drift using the plant-specific as-found/as-left data of the instrumentation devices. The use of statistical graphic tools in the procedure can facilitate the process of the previous setpoint drift evaluation. The scope of the drift analysis is enlarged by the adoption of non-parametric statistical methods in the view point of the methodology. Several statistical process control techniques will provide the plant staffs with more efficiency in the management of the instrumentation device.