20 October 2025 in Proceedings of the Institution of Mechanical Engineers, Part I: Journal of Systems and Control Engineering
Considering the occurrence of actuator faults in the wind turbine generator main drive system, a fault model of the wind turbine generator main drive system is established. First, an adaptive sliding mode observer is designed to estimate the actuator faults, which is able to identify and reconstruct the faults quickly and provides the necessary information for the implementation of fault-tolerant control. In order to improve the design process of the observer, the gain matrix solution problem is transformed into an optimization problem subject to linear matrix inequality constraints. Secondly, considering the existence of unmodeled dynamics and external noise disturbances in the main drive system, an adaptive feedback fault-tolerant controller is designed using the fault reconstruction information after an actuator failure occurs in the system to compensate for the system errors caused by actuator failures, unmodeled dynamics, and external noise disturbances, which ensures a stable operation of the system. Finally, a simulation study of a 5 MW wind turbine system shows that at an average wind speed of 7.5 m/s and a turbulence intensity of class A, the system is restored to a normal state after the addition of the fault-tolerant controller, and the generator electromagnetic torque error is only 1.89%, which verifies the validity of the proposed method.
8 December 2023 in Proceedings of the Institution of Mechanical Engineers, Part I: Journal of Systems and Control Engineering
Aiming at the shortcomings of the Golden Jackal optimization algorithm, such as low convergence accuracy and easy falling into the optimal local solution, an improved Golden Jackal optimization algorithm was proposed. First, sine and piecewise linear (SPM) chaotic mapping was introduced to increase the population number to achieve the purpose of initial population diversity. The self-adaptive weight and sine–cosine algorithm improved the position update formula of the Golden Jackal optimization algorithm, so the global search ability of the golden jackal algorithm is improved, and avoid the algorithm that fell into local optimality. Second, simulation experiments with eight standard test functions are performed to prove that the algorithm has excellent optimization ability. The improved Golden Jackal optimization algorithm was applied to optimize the kernel parameters of hybrid kernel principal component analysis. A fault diagnosis model is proposed to improve the golden jackal algorithm to optimize the kernel principal component analysis. Finally, the proposed method is used to fault diagnosis in the hot strip mill process. According to the study of simulation results, the faulty data can be identified effectively by this method, the accuracy is up to 100%, and the fault false alarm rate is greatly reduced.
20 March 2024 in Proceedings of the Institution of Mechanical Engineers, Part I: Journal of Systems and Control Engineering
A mathematical model of the main drive system of rolling mill is developed for the motor armature faults in the main drive system of rolling mill, taking into account the effects of nonlinear damping, friction damping between rolls and external perturbations on the system during operation. The unknown input observer is designed to process perturbations and errors as unknown inputs, and the [ ] performance index is designed to improve the robustness of the unknown input observer to fault reconfiguration. At the same time, in order to ensure the stability of the main drive system of rolling mill after the occurrence of faults, an active fault-tolerant controller is designed for the main drive system of rolling mill after obtaining the accurate fault reconfiguration value, which realizes the active fault-tolerant control of the main drive system of rolling mill. Through the simulation study on the main drive system of 2030 mm cold rolling mill stand F4, it is shown that the system is restored to the normal state after adding the fault-tolerant controller, and the motor angular speed error and roll angular speed error are 2.12% and 2.43%, respectively. Comparing with the fault-tolerant control method based on fault estimation, the root-mean-square errors of the motor angular speed and roll angular speed estimates of the designed fault-tolerant control method are reduced by 4.24% and 3.28%, respectively. Simulation verifies the effectiveness of the proposed method.
17 January 2026 in Transactions of the Institute of Measurement and Control
To address compound faults—including actuator faults, sensor faults, and unknown disturbances—in the main drive system (MDS) of a rolling mill, a fault diagnosis model is established based on the d-q model of a three-phase AC motor. To overcome the limitation of conventional model-based fault detection (MFD) in isolating faults under compound fault scenarios, a novel fault diagnosis framework is proposed by integrating model residuals with deep learning. An Unknown Input Observer (UIO) is designed to detect system faults, and its convergence is rigorously proven using Lyapunov theory and linear matrix inequalities (LMIs). The coupled residual signals generated by the observer are segmented into sequential subsequences and processed by a Convolutional Neural Network (CNN) for feature extraction and classification. To account for the temporal and dynamic nature of the residuals, a Tyrannosaurus Rex Optimization Algorithm (TROA) is adopted to optimize the CNN hyperparameters. Numerical simulations on a rolling mill demonstrate that the proposed UIO achieves superior state estimation performance, with a 16.48% reduction in the Root Mean Square Error (RMSE) of angular velocity difference of the motor compared to the Sliding Mode Observer (SMO). Furthermore, the proposed TROA-CNN outperforms Bayesian Optimization (Bayes)-CNN, Whale Optimization Algorithm (WOA)-CNN, and Grey Wolf Optimization (GWO)-CNN in terms of fault classification accuracy (99.76%) and noise robustness (96.27% under 5% noise). In scalability tests where the number of fault types or dataset size are doubled, the inference latency increases by approximately 10%, and the training time rises by about 40%. These results demonstrate that the UIO-TROA-CNN achieves high accuracy, strong robustness, and excellent scalability, making it well-suited for fault diagnosis in industrial environments with high noise and complex fault types.