Sensor fusion and fuzzy logic-based approach for maintenance in hydrogen production plants

Temperature sensors used in control systems of hydrogen production plants are critical components which are subject to various environmental conditions that impact on their reliability and accuracy of measurements. The control system depends on accurate temperature measurements to make decisions such as activation of the cooling unit or optimisation of the process. The systems can also be equipped with fault detection units to detect out-of-range temperatures. Problem arises when the temperature sensor measurements drift from the actual value. This can prevent the control system from responding correctly and result into problems such as membrane drying or overheating. This paper proposes a practical solution to detect drift faults in temperature sensors embedded in electrolysers using an intelligent concept based on the combination of sensor fusion and fuzzy logic. The proposed solution can be integrated into control systems to ensure that drifting sensors are detected, recalibrated or replaced early.