Development of a Novel Thermomechanical Modelling Framework for Predicting Actuator Performance Degradation in High-Temperature Industrial Mechatronic Systems in Nigerian Steel Plants

Electromechanical and hydraulic actuators embedded in the control systems of electric arc furnaces, ladles, and continuous casting machines in Nigerian steel plants experience severe thermomechanical loading from radiation, convective heat transfer, and periodic thermal cycling that accelerates material degradation and leads to premature failure mechanisms including thermal fatigue cracking, seal degradation, and insulation breakdown. The accurate prediction of actuator performance degradation over operational life in these environments is essential for reliability-based maintenance scheduling yet is inadequately addressed by existing actuator lifetime models that assume near-ambient temperature operation. This dissertation develops a novel thermomechanical modelling framework for actuator performance degradation prediction under high-temperature industrial cyclic loading, integrating coupled finite element thermal-structural simulation, physics-informed neural network-based constitutive modelling, and probabilistic damage accumulation theory. A key theoretical contribution is a multi-scale constitutive model for the thermo-viscoplastic behaviour of actuator structural materials at elevated temperatures, fitted using a physics-informed neural network architecture that satisfies thermodynamic consistency constraints by design. The framework is parameterized and validated using actuator degradation data collected from three Nigerian steel plants in Delta, Ogun, and Kwara States over 24 months of instrumented operation. Predicted actuator life distributions show 80 percent confidence interval coverage of 87 percent of observed failure events, demonstrating significantly higher predictive accuracy than existing time-temperature parameter methods. A maintenance scheduling optimization study demonstrates potential 26 percent reduction in actuator replacement cost under the proposed model-based maintenance policy. Keywords: thermomechanical modelling, actuator degradation, steel plant, physics-informed neural network, high temperature