Vibration-based fault diagnosis methods are widely used in industrial applications due to their high accuracy and reliability. However, their implementation is often hindered by challenges such as confined spaces, harsh environmental conditions, and cost limitations. Moreover, traditional deep learning-based approaches frequently overlook the critical relationship between virtual and physical signals, which can limit their diagnostic performance. To overcome these challenges, this study introduces a novel digital twin-driven feature imputation framework (NIFD-Net) for non-contact intelligent fault diagnosis. The NIFD-Net framework consists of three key stages: (1) codebook generation, where features from simulated vibration signals (SVS) and non-contact signals (NCS) are compressed and regularized using diagnosis-oriented codebooks ; (2) mapping construction, where a Transformer is employed to establish the relationship between SVS and NCS in a compressed feature space ; and (3) diagnostic enhancement, where imputed SVS features are integrated with original NCS features to improve diagnostic performance. First, digital twin (DT)-enabled models, developed using structural parameters and fault characteristics, dynamically simulate the operating conditions of rolling bearings to generate high-quality SVS samples. The relationship between the two feature spaces of SVS and NCS is then learned by a transformer to impute the SVS features. Finally, the imputed SVS features are integrated with the original NCS features to enhance fault diagnostic performance. Extensive experiments on three case studies demonstrate that NIFD-Net outperforms state-of-the-art methods, achieving an average accuracy of 98.34% across diverse tasks. Notably, the framework exhibits strong noise robustness, maintaining high diagnostic performance even under adverse conditions (e.g., SNR = -10 dB). The results highlight the effectiveness of integrating simulated and real signals for non-contact fault diagnosis, offering a promising solution for industrial applications where traditional contact-based methods are impractical.

This paper introduces an observer-based neural quadratic sliding mode control strategy for interconnected Markov jump systems faced with unknown interconnections, regardless of the high dimensionality of the systems. Firstly, a dynamic event-triggered scheme is constructed in the communication channel to the Lebesgue state observer, with which an integral quadratic sliding mode hyperplane is put forward; Secondly, a neural-based control method is put forward to make sure that predefined sliding hyperplane is attractive; In addition, the occurrence of Zeno phenomenon is also verified to be avoided with the implementation of the controller; Thirdly, linear matrix inequality technique and Lyapunov stochastic stability theory are proposed to check the stochastic stability of closed-loop systems, including sliding mode dynamics and error dynamics; Finally, simulation results on single-link robot arms are given to reveal the validity of the obtained results.

Wind speed forecasting plays a crucial role in reducing the risk of wind power uncertainty, which is vital for power system planning, scheduling, control, and operation. However, it is challenging to obtain accurate wind speed forecasting results since wind speed series contain complex fluctuations. In this paper, a novel wind speed forecasting model is proposed by using a genetic algorithm (GA) and long short-term memory neural network (LSTM), where GA is used for evolving the architectures and hyper-parameters of the LSTM, called EvLSTM, because there is no clear knowledge to determine these crucial parameters. In the proposed EvLSTM forecasting model, a flexible gene encoding strategy, crossover operation, and mutation operation are proposed to describe the different architectures and hyper-parameters of the LSTM during the evolutionary process. In addition, to overcome the weakness of the single method for wind speed forecasting, the ensemble EvLSTM (EnEvLSTM) is proposed by using the no negative constraint theory ensemble learning strategy, whose weight coefficients are determined by the differential evolution algorithm. The proposed EvLSTM and EnEvLSTM forecasting models are evaluated on two real-world wind farms located in Inner Mongolia, China and Sotavento Galicia, Spain. Experimental results on two forecasting horizons demonstrate the superiority of EvLSTM and EnEvLSTM in terms of three performance indices and three statistical tests.

The paper proposes an adaptive controller design for Markov jump systems with mixed mode transition information through a reduced-order sliding mode approach. The stability criteria and mode-dependent adaptive control law are achieved using linear matrix inequality technique. Firstly, a linear reduced-order sliding surface function is proposed to achieve the reduced-order sliding mode dynamics. Secondly, a feasible approach is presented to check the stochastic stability of resulting sliding motion corresponding to different mode transition information, and to solve the controller gains from stability criteria. Thirdly, an adaptive sliding mode controller is also designed to ensure the finite-time reachability of the predefined hyperplane even when no mode information is available. Finally, the application of the proposed control strategy to the RLC circuit is provided.