Optimal Sensor Placement in Nuclear Power Steam Water System Based on Signed Directed Graph

Abstract Nuclear power plant is expected to achieve the comprehensive fault monitoring and ensure the secure and stable operation. Enhancing the reliability of the digital measurement system will effectively mitigate the potential risks associated with compromised operation of the nuclear power system, confine the extent of fault impact, and bolster the operator’s situational awareness proficiency. This paper presents a comprehensive modeling strategy for obtaining the fault-sensor correlation matrix in the deaerator system of a nuclear power plant based on SDG theory. Based on the utilization of Advanced Process Simulation Software (APROS), we successfully developed a dynamic numerical simulation model for the deaeration system in nuclear power plants. By conducting a comprehensive analysis of the correlation between fault propagation path and sensor, we have successfully established a probability model that accurately assesses the performance of sensor fault detection. A multi-objective optimization model for sensor layout, considering the cost of sensor placement, detection ability, and fault correlation, was developed based on constraint relationships. Compared to conventional sensor placement schemes, the optimized sensor placement scheme exhibits superior performance in terms of cost-effectiveness in arranging sensors, reliability in fault detection, and correlation between faults and sensor points.