Bearing fault diagnosis is of great significance to ensure the safe operation of mechanical equipment. This paper proposes an intelligent fault diagnosis method of rolling bearings based on deep belief network (DBN) with hyperparameter optimization by using parallel computing. Different with traditional diagnosis methods that extract the features manually depending on much prior knowledge about signal processing techniques and diagnostic expertise, DBN extracts fault features automatically by machine learning mechanism. Considering the time consuming problem, parallel computing is adopted to the DBN training process by using a Master/Slave mode to improve the training speed so that the global optimization with Genetic Algorithm and higher diagnosis accuracy can be achieved. Finally, the proposed method is verified with the public datasets provided by Case Western Reserve University (CWRU) with various fault depths in different locations and loads of rolling bearings. The results indicate that the proposed method can identify bearing faults under different conditions correctly which significantly enhances the intelligence of fault classification and reduces the time for parameter selection of deep learning models.

Tool wear condition monitoring and reaming life prediction are critical for near-zero downtime machining. Recent manufacturing outsourcing business environment necessitates more focus on machine performance degradation to optimize the tool management for improved six-sigma productivity and manufacturing performance. The unmet needs for drilling monitoring is how to effectively predict its remaining life and manage the tool change to minimize downtime and costs. This paper presents a hybrid method for on-line assessment and performance prediction of remaining tool life in drilling operations based on the vibration signals. Logistic regression (LR) analysis combined with maximum likelihood technique is employed to evaluate tool wear condition based on features extracted from vibration signals using wavelet packet decomposition (WPD) technique. Auto-regressive moving average (ARMA) model is then applied to predict remaining useful life based on tool wear assessment result. In addition, failure risk distribution is discussed. The developed prognostic method is validated in drilling operations, which can be also implemented to other manufacturing processes.

The assembly process of large-scale and non-standard industrial equipment poses significant challenges due to its inherent scale-related complexity and proneness to errors, making it difficult to ensure process cost, production cycle, and assembly accuracy. In response to the limitations of traditional ineffective production models, this paper aims to explore and propose a digital twin (DT)-based technology paradigm for the intelligent assembly of large-scale and non-standard industrial equipment, focusing on both the equipment structure and assembly process levels. The paradigm incorporates key technologies that facilitate the integration of virtual and physical information, including the establishment and updating of DT models for assembly structures using actual data, the assessment of structural assemblability based on DT models, the planning and simulation of assembly processes, and the implementation of virtual commissioning technology tailored to the actual assembly process. The effectiveness of the proposed paradigm is demonstrated through a case study involving the actual assembly of a large-scale aerodynamic experimental equipment. The results confirm its ability to provide valuable technical support for the design, evaluation, and optimization of industrial equipment assembly processes. By leveraging the DT-based methodological system proposed in this paper, significant improvements in the transparency and intelligence of industrial equipment production processes can be achieved.

Crack fault diagnostics plays a critical role for rotating machinery in the traditional and Industry 4.0 factory. In this paper, an experiment is set up to study the dynamic response of a rotor with a breathing crack as it passes through its 1/2, 1/3, 1/4 and 1/5 subcritical speeds. A cracked shaft is made by applying fatigue loads through a three-point bending apparatus and then placed in a rotor testbed. The vibration signals of the testbed during the coasting-up process are collected. Whirl orbit evolution at these subcritical speed zones is analyzed. The Fourier spectra obtained by FFT are used to investigate the internal frequencies corresponding to the typical orbit characteristics. The results show that the appearance of the inner loops and orientation change of whirl orbits in the experiment are agreed well with the theoretical results obtained previously. The presence of higher frequencies 2X, 3X, 4X and 5X in Fourier spectra reveals the causes of subharmonic resonances at these subcritical speed zones. The experimental investigation is more systematic and thorough than previously reported in the literature. The unique dynamic behavior of the orbits and frequency spectra are feasible features for practical crack diagnosis. This paper provides a critical technology support for the self-aware health management of rotating machinery in the Industry 4.0 factory.

Grinding robots are widely used in the automotive, mechanical processing, aerospace industries, among others, due to their strong adaptability, high safety and intelligence. The grinding process parameters are the main factors that affect the quality and efficiency of grinding robots. However, it is difficult to obtain the optimal combination of the grinding process parameters only by manual experience. This study proposes an artificial intelligence-based method for optimizing the process parameters of a grinding robot using neural networks and a genetic algorithm, with the aim to reduce the workpiece surface roughness and shorten the grinding time. Specifically, this is the first study utilizing a multi-objective optimization approach to optimize the process parameters of a grinding robot. Based on the experimental data of the grinding robot ROKAE XB7, the long short-term memory (LSTM) and multilayer perceptron (MLP) neural networks were trained to fit the quantitative relationships between the process parameters of the grinding robot, such as feed rate, spindle pressure and pneumatic motor pressure, and the result of grinding surface roughness and grinding time. After that, the non-dominated sorting genetic algorithm II (NSGA-II) was used to calculate the Pareto optimal process parameter combinations using the trained LSTM and MPL model as the objective function. Compared with the method based on manual experience, the process parameters optimized with this method achieved a reduction in surface roughness of at least 13.62% and a reduction in the whole grinding process time of 28%. The excellent grinding results obtained for grinding time and surface roughness validated the feasibility and efficiency of the proposed multi-objective method for the optimization of grinding robots’ process parameters in practical manufacturing applications.