Search

The self-attention mechanism has advantages in analyzing the internal characteristics of data and capturing local information. Therefore, it is applied to fault detection. Since complex industrial processes usually contain a mixture of properties, such as nonlinearity, dynamics, and non-Gaussianity, this places higher demands on fault detection methods. To further improve the performance of the self-attention mechanism for fault detection in complex industrial processes, an innovative self-attention mechanism integrating the adaptive double subspace (ISA-ADS) method is proposed for fault detection. First, the query matrix, the key matrix, and the value matrix of the self-attention mechanism are designed so that they can better focus on local information through the adaptive sample weight allocation strategy. At the same time, this design reduces the interference of nonlinear and noisy information and amplifies the impact of important fault information in the sample. Second, considering the problem of Gaussian and non-Gaussian distributions of the data and the different degrees of influence of different subspaces on the final process monitoring results, an adaptive Gaussian and non-Gaussian double subspace fault detection model based on the self-attention output matrix is established, which adaptively assigns different weights to different subspaces by determining their importance. The statistics of different subspaces are fused using the Bayesian inference and a new adaptive monitoring statistic is constructed to obtain a more accurate monitoring of the process state. Finally, the ISA-ADS is validated using the Tennessee Eastman (TE) process and the actual polyethylene production process. The results show that ISA-ADS has a low false alarm rate and a high fault detection rate, verifying the effectiveness of its fault detection performance.

This brief introduces a novel adaptive optimal output-feedback controller for permanent magnet synchronous motor (PMSM) systems, eliminating the need for prior knowledge of system dynamics, numerous integral window functions, or unmeasurable states and load torque. Initially, we design an adaptive optimal output-feedback controller by constructing internal states. Then, a policy iteration algorithm based on adaptive dynamic programming approximates the optimal output-feedback gain using only input and trajectory tracking error information. Notably, this method does not require the minimal polynomial of an exosystem or the solution of regulator equations, facilitating the overall design of the feedforward-feedback controller. The effectiveness of the proposed learning algorithm is validated on a PMSM system.

Control valves are important actuators in industrial processes. However, valve models identified by traditional system identification methods are difficult to work in actual industrial scenarios due to the influence of dynamic working conditions and interference, which also limits in-depth research on monitoring, diagnosis, control, etc. of control valves. To solve this issue, this article proposes a flow control valve identification method based on the idea of model-data-fusion. The proposed method can maintain good flow estimation performance by determining the relationship between the sensor data and the flow control valve model, even in case of changes in operating conditions. Meanwhile, a window down sampling (WDS) method is proposed to assist the modeling and testing process. On the basis of WDS, multiple flow characteristic models are established, and the performance of different models is evaluated based on the principle of feature selection. Furthermore, the convergence of the proposed parameter optimization method is proved in this study. The proposed flow estimation method is verified by the actual industrial data of a chemical plant.