Prognostics is an emerging science of predicting the health condition of a system (or its components) based upon current and previous system states. A reliable predictor is very useful to a wide array of industries to predict the future states of the system such that the maintenance service could be scheduled in advance when needed. In this paper, an adaptive recurrent neural network (ARNN) is proposed for system dynamic state forecasting. The developed ARNN is constructed based on the adaptive/recurrent neural network architecture and the network weights are adaptively optimized using the recursive Levenberg-Marquardt (RLM) method. The effectiveness of the proposed ARNN is demonstrated via an application in remaining useful life prediction of lithium-ion batteries.

Most rotary machinery imperfections are related to defects in rolling element bearings. Unfortunately, reliable bearing fault detection still remains a challenging task, especially when bearing defect-related features are nonstationary. A new morphological Hilbert-Huang (MH) technique is proposed in this paper for incipient bearing fault detection. In the proposed MH technique, a new linearity measure method is suggested to demodulate characteristic feature functions, and a mathematical morphological filter is proposed to reduce impedance effect of the measured vibration signal to improve fault detection accuracy. The effectiveness of the proposed MH technique is verified by a series of experimental tests corresponding to different bearing conditions.

The particle filter (PF) has been used for the analysis of nonlinear, non-Gaussian dynamical systems with hidden state variables. However, PF has some limitations in real-world applications, for instant, the sample degeneracy and the impoverishment, which are considered as long-standing challenges in this research and development field. Although several techniques have been proposed in the literature for this purpose, they have some limitations: for example, they cannot represent the entire probability density function (pdf) effectively and are usually problem dependent. In this paper, an enhanced mutated PF (EMPF) technique is proposed to improve the performance of PFs. In the EMPF technique, first, a novel enhanced mutation approach is proposed to actively explore the posterior pdf to locate the high-likelihood area. Second, a new selection scheme is suggested to process low-weight particles for optimizing the posterior distribution and tackling sample degeneracy. Third, an outlier assessment method is adopted to monitor the overall pattern of the posterior distribution based on the interquartile range statistical analysis. The effectiveness of the proposed EMPF technique is verified by simulation tests. It is also implemented for the remaining useful life prediction of lithium-ion batteries. Test results show that the developed EMPF technique can capture a system's dynamic behavior and track system characteristics effectively.

Lithium-ion (Li-ion) batteries have become essential in modern industries and domestic applications due to their high energy density and efficiency. However, they experience gradual degradation over time, which presents significant challenges in maintaining optimal battery performance and increases the risk of unexpected system failures. To ensure the reliability and longevity of Li-ion batteries in applications, various methods have been proposed for battery health monitoring and remaining useful life (RUL) prediction. This paper provides a comprehensive review and analysis of the primary approaches employed for battery health monitoring and RUL estimation under the categories of model-based, data-driven, and hybrid methods. Generally speaking, model-based methods use physical or electrochemical models to simulate battery behaviour, which offers valuable insights into the principles that govern battery degradation. Data-driven techniques leverage historical data, AI, and machine learning algorithms to identify degradation trends and predict RUL, which can provide flexible and adaptive solutions. Hybrid approaches integrate multiple methods to enhance predictive accuracy by combining the physical insights of model-based methods with the statistical and analytical strengths of data-driven techniques. This paper thoroughly evaluates these methodologies, focusing on recent advancements along with their respective strengths and limitations. By consolidating current findings and highlighting potential pathways for advancement, this review paper serves as a foundational resource for researchers and practitioners working to advance battery health monitoring and RUL prediction methods across both academic and industrial fields.

Reliable induction motor (IM) fault detection techniques are very useful in industries to diagnose IM defects and improve operational performance. A smart sensor-based technology is proposed in this paper to synergistically use vibration and current harmonics for rotor bar fault detection in IMs. The vibration signal is used for analysis of shaft speed variations and the current harmonics information is applied for rotor bar fault detection. A wireless smart sensor network is developed and used for data collection, allowing for low-cost, low space footprint, and non-invasive installation. The effectiveness of the proposed synergistic technique is examined experimentally, with results demonstrating advantages over conventional methods in terms of accurately differentiating between a healthy and faulty motor, as well as estimating the fault severity, even under zero-load IM conditions. A means to quantify the fault states as diagnostic indices is also proposed for online IM health condition monitoring.

The detection of the onset of damage in gear systems (e.g., gearboxes) is of great importance to a wide array of industries. In this paper, an enhanced diagnostic (ED) system is developed for real-time gear system condition monitoring. A neurofuzzy (NF) paradigm is adopted for pattern classification of the features from the energy, amplitude, and phase domains. The diagnostic reliability is enhanced by properly integrating predicted future machinery states that are forecast by recurrent NF predictors. An online training technique is proposed to improve the classifier's adaptive capability to accommodate different machinery conditions. The viability of this new monitoring system has been verified by experimental tests under different gear conditions. This proposed ED system has also been applied for real-time condition monitoring in multistage printing machines. The primary application has demonstrated its reliability as an effective monitoring tool for both production quality control and maintenance planning.