Research on Signal Reconstruction Technology for Nuclear Power Plant Based on Self-Associative Kernel Regression

The instrument control system of the nuclear power plant is an important part of ensuring its safe and reliable operation. The sensor is the core component of the instrument control system to monitor the operation status of the nuclear power plant and trigger the operation of the instrument. Therefore, the availability of the sensor plays an important role of the safe operation of the nuclear power plant. The virtual sensor technology based on self-associative kernel regression can accurately determine the fault of the sensor signal, such as instrument drift, missing instrument values, etc. And it can replace the fault sensor with reliable and accurate values in a short period of time, ensuring normal operation of the system in a short period of time. In addition, the output value of the virtual sensor can also be used to calculate the measured value of the unmounted sensor part. This paper introduces a method to replace sensor values with right values when measured sensor values has something unusual. This paper only concerns about replacing sensor values and the paper assumes that it can find sensor fault correctly and timely. The paper selects the main steam system of the pressurized water reactor of the nuclear power plant as an example to verify the feasibility and reliability of the virtual sensor technology based on the self-associative kernel regression. The research results show that it is basically feasible to implement the virtual sensor technology by using the related technical solutions proposed in this paper.