Evaluation of CANDU safety-system calibration accuracy through monitoring

CANDU operators are increasingly being required to demonstrate that estimated accuracies of the transmitters used to measure process variables in CANDU special safety systems are being achieved in operation. At present, this is done by recalibrating all transmitters periodically, typically every one to three years. Functional checks are performed more frequently. These include panel checks (i.e., visual monitoring and cross-comparison of redundant process measurements) and safety-system process-trip tests (i.e., valving-in test pressures to activate the alarm units). These functional tests provide assurance that transmitters and related circuits are working, but they are not always sufficiently sensitive to evaluate calibration accuracy. Continuous computerized monitoring of safety-system signals is being investigated as one approach to verifying safety-system calibration accuracy and transmitter functionality. Eventually, if successful, it may enable a reduction in the frequency of process trip testing (a labour-intensive, error-prone operation), or even the elimination of process trip tests and their associated hardware altogether. Using both Shutdown System (SDS) 1 and SDS2, it is possible to obtain at least three, and more often six, independent, redundant measurements of each of many process variables. If signals from the reactor regulating, containment and emergency coolant injection (ECI) systems are also used, then it is possible to obtain up to 13 independent, redundant measurements of many process variables. Assuming these measurements are truly independent, and that there are no common mode errors, calibration accuracy can be verified by intercomparing the nominally identical readings. A CANDU Owners Group (COG) R and D project was set up to evaluate this monitoring concept. This project has progressed through several phases: (1) data acquisition, (2) data analysis and algorithm development, and (3) on-line calibration analysis and operational experience. Phases 1 and 2 are complete and phase 3 has been in operation since 1995 August. This paper describes the progress made to date and the expectations for the future. (author)