Celebrating the 70th Anniversary of School of Mechanical Science and Engineering of Huazhong University of Science & Technology

Abstract

The School of Mechanical Science and Engineering (MSE) of Huazhong University of Science & Technology (HUST-MSE) is one of the best mechanical engineering schools in China. HUST-MSE not only leads the development of equipment automation, digitization and intelligence in China but also wins a high reputation in the field of mechanical engineering in the world. To celebrate the 70th anniversary of HUST-MSE, this special issue aims at presenting the new methodologies and techniques for the application of intelligent manufacturing.

This special issue contains seven contributions on the topic areas of manufacturing scheduling, fault diagnosis, automatic welding, and reconfigurable battery systems, which are the important topics in intelligent manufacturing. All the papers are invited from the scholars who were graduated from HUST-MSE.

The first paper, ‘an approximate evaluation method for neighbourhood solutions in job shop scheduling problem’ by Gui et al., investigates the approximate evaluation method for the meta-heuristic algorithm solving the Job Shop Scheduling problem. The authors prove that the evaluated value of the neighbourhood solution is under certain conditions by exploring domain knowledge. It can reduce the computational time of the evaluation of meta-heuristics and improve its efficiency.

The second paper, ‘a deep reinforcement learning based approach for dynamic distributed blocking flowshop scheduling with job insertions’ by Sun et al., studies the distributed blocking flowshop scheduling problem (DBFSP) with new job insertions. The authors propose a multi-agent deep deterministic policy gradient method to optimize the job selection model and only make little local modification based on the original plan while minimizing the objective of the total completion time deviation of all products so that all jobs can be finished on time.

The third paper, ‘deep reinforcement learning-based balancing and sequencing approach for mixed model assembly lines’ by Lv et al., proposes a multi-agent iterative optimization method for the balancing and sequencing problem in mixed-model assembly lines. The balancing agent adopts a deep deterministic policy gradient algorithm, while the sequencing agent uses an Actor Critic algorithm. Then an iterative interaction mechanism is developed for these agents to minimize the work overload and the idle time at stations.

The fourth paper, ‘intelligent fault diagnosis of rotating machinery using lightweight network with modified tree-structured Parzen estimators’ by Liang et al., investigates a novel lightweight network with modified tree-structured Parzen estimators to automatically search the optimal hyper-parameters for the fault diagnosis task.

The fifth paper, ‘privacy-preserving gradient boosting tree: vertical federated learning for collaborative bearing fault diagnosis’ by Xia et al., focusses on the insufficient data in real manufacturing scenarios. The authors investigated a vertical federated learning method to break down the data silos while preserving data privacy. Only the model information will be shared for the collaboration to promote its performance.

The sixth paper, ‘construction of semi-dense point cloud model for tube-to-tubesheet welding robot’ by Wang et al., aims to promote the tube-to-tubesheet welding and develops a semi-dense point cloud model based on a selected monocular camera and one-dimension laser rangefinder. A laser filtering method is developed firstly to acquire the distance between the camera and the tubesheet, and the tubesheet point cloud model is constructed through the graph optimization algorithm.

The seventh paper, ‘reconfigurable battery systems: challenges and safety solutions using intelligent system framework based on digital twins’ by Garg et al., presents an intelligent system framework based on digital twins. The proposed framework is further extended to the life cycle management approach, and it can be helpful to optimize the design, manufacturing, operation, and maintenance of batteries.

We appreciate all the authors who have contributed to this special issue. We are also grateful to all the reviewers for their services and commitments to this special issue through their rigorous reviews, timely responses within a tight schedule, and insightful and constructive comments that helped shape the outcome of this issue. All the papers show the good improvements on intelligent manufacturing on the theoretic aspect or the application. Meanwhile, there are still many challenges in the field. The further researches can be conducted on all branches of the collaborative intelligent manufacturing and promote the effectiveness and efficiency of manufacturing systems. We also hope that HUST-MSE is developing better and better.