Continuous online monitoring in pressurized water reactors during flexible operation using PLSR-based technique – Case study: Load following test

Abstract

Abstract This paper proposes the application of a multivariant-based technique, namely, Partial Least Square Regression (PLSR) to enhance the process of continuous online monitoring and control in pressurized water reactors (PWR) during flexible operation. The results of the investigations of this paper have shown that the proposed technique provided effective performance during flexible PWR operation even with a small size of available history operating data samples. More specific, the proposed model provides the reactor operator with a comprehensive and simple presentation of the reactor status via a single dynamic trajectory during flexible operation which helps detecting anomalies at an early stage. In addition, results also have shown that the proposed PLSR-based model minimized the limitation on the ex-core accuracy resulting from the operation in the flexible mode, provided reliable, fast with excellent monitoring efficiency. Moreover, the results have also shown that the proposed model provides a graphical control guide which is achieved via a prioritized, reduced, comprehensive and manageable set of the PWR physical parameters. The analysis of this graphical guide has shown that it provides the fine control tuning, which in many cases is left for the operator, thus it minimizes the human error by relieving operators from having to pay attention to additional tasks during transient contingencies. In the context of this paper, a load-following test is conducted to ascertain the efficiency of the proposed technique. The data of the test was observed at Unit-1, Tihange nuclear power station, Belgium (three-loop PWR, 2873\xa0MWth, 962\xa0\xa0MWe net). The PLSR-based model for continuous model for continuous monitoring as well as the graphical control guide are developed and validated using SIMCA software package.