Determination of Unidentified Leakage Using a Kalman Smoother

Since the safety significance of leaks from the RCS can widely vary depending on the source of the leak as well as the leak rate, the detection of the leakage is an important issue. The leakage is classified into 1) identified leakage which is defined as leakage into closed systems such as pump seal or valve packing leaks that can be captured, and 2) unidentified leakage which is all other leakage. The unidentified leakage is typically determined by the RCS inventory balance method which is based on NUREG-1107. Since the accuracy of leak rate calculation is dependent of the plant operating condition, the change in the RCS temperature, inventory, and the transient operating condition should be avoided during the measurement period. Nevertheless, the operation of the makeup of the borated water into the RCS and the diversion of the inventory to the outside of the RCS boundary makes it difficult to maintain the plant stable over an hour. Due to the large fluctuation of the calculated leak rate, it is sometimes hard to know the trend of the leakage as well as the instantaneous leak rate. Any fluctuation of operating conditions can results in unreliable leak rate. This study proposes a new way of determining the unidentified leak rate using a Kalman filter and smoother technique. The proposed algorithm enhances the accuracy of the leak rate calculation not only for the steady state operations but also for transients in a well timed manner