Anomaly detection is the cornerstone of the health management of many large industrial mechanical equipment. Most machinery anomaly detection methods try to find a variable threshold to indicate whether machinery is functioning normally. However, due to the scarcity of anomalous data, how to drive anomaly detection with variable threshold only by normal data remains a significant challenge. This is because, in practice, mechanical equipment’s operating speed and load conditions are constantly changing. To solve this problem, an unsupervised anomaly detection approach based on autoencoders and temporal convolutional networks was proposed in this paper, which is named Variable Cumulative Error Anomaly Detection (VCEAD). The autoencoder will be used to compute the signal reconstruction error, and temporal convolutional network will be used to predict vibration signal. Finally, an algorithm based on the probability distribution of the signal reconstruction error will be proposed to calculate the so-called variable cumulative error. To validate the effectiveness of the proposed method in anomaly detection, multi-group lifetime datasets collected from the laboratory were studied and analyzed. The comparative results show that the method presented in this paper achieves superior performance.

Iterative feedback tuning (IFT) enables the data-based optimization of feedback controller parameters without the plant model and disturbance model. However, the traditional IFT suffers from a generally slow convergence rate and requires multiple experiments in each iteration, which severely limit its applicability for precision motion industry. In this paper, a new framework for IFT with its focus on the practical applicability is synthesized. Specifically, in order to improve the convergence rate, a novel two-loop iterative algorithm of IFT is proposed by introducing a weighted gradient of the performance criterion. This algorithm seeks to directly minimize the performance criterion instead of its linear approximation in each iteration. Furthermore, an unbiased estimate method of the auxiliary variables is developed based on the impulse response experiment. These guarantee the proposed approach high convergence rate with less experiments required per iteration. Comparative simulation and experimental results demonstrate that the proposed approach converges much faster than the traditional IFT, and outperforms the model-based approach in terms of the tracking performance. The ease of implementation and effectiveness makes the proposed approach highly suitable for industrial applications.