IET Collaborative Intelligent Manufacturing最新文献

Current manufacturing systems are forced to meet the most dynamic market demands under sustainable factors. However, not only technical transformations will address the challenge but, to fully cover social needs, the analysis of the human role in highly interactive systems is still decisive, following a socially sustainable approach. To fully extract the most from both agents under a performance point of view, the main added value of agents in the work environment needs to be carefully analysed, captured, and boosted. The context shapes a specific operation or task, which consequently drives the final outcome according to individual necessities. Furthermore, a methodology that potentially helps a proper assessment of these performance-based interactions is still missing. The contribution focusses on the definition of a novel human-centric methodology under a holistic point of view to analyse performance-based interactions and to define appropriate indices and metrics that helps assessing the human-system interactions in the manufacturing domain. The methodology is applied in a case study to guide practitioners with its use.

Job shop scheduling problem is a classical scheduling problem, and it is very difficult to work out. To solve it well, the meta-heuristic algorithm is a good choice, and the evaluation method of neighbourhood solutions, which affects the efficiency of the algorithm and the quality of the solution, is one of the keys in the algorithm. We propose an approximate evaluation method by exploring domain knowledge in neighbourhood solutions. Firstly, we reduce the computational time of the evaluation by analysing the unnecessary computational operations. Secondly, according to the domain knowledge, we prove that the evaluated value of the neighbourhood solution is the exact value under certain conditions. At the same time, a set of critical parameters are calculated to correct the estimated value of the neighbourhood solutions that do not meet the conditions to improve the evaluation accuracy. With all of these, an approximate evaluation method for neighbourhood solutions in job shop scheduling problems is proposed. The experiments on different numerical instances show the superiority of the method proposed.

Data-driven fault diagnosis approaches have been widely adopted due to their persuasive performance. However, data are always insufficient to develop effective fault diagnosis models in real manufacturing scenarios. Despite numerous approaches that have been offered to mitigate the negative effects of insufficient data, the most challenging issue lies in how to break down the data silos to enlarge data volume while preserving data privacy. To address this issue, a vertical federated learning (FL) model, privacy-preserving boosting tree, has been developed for collaborative fault diagnosis of industrial practitioners while maintaining anonymity. Only the model information will be shared under the homomorphic encryption protocol, safeguarding data privacy while retaining high accuracy. Besides, an Autoencoder model is provided to encourage practitioners to contribute and then improve model performance. Two bearing fault case studies are conducted to demonstrate the superiority of the proposed approach by comparing it with typical scenarios. This present study's findings offer industrial practitioners insights into investigating the vertical FL in fault diagnosis.

The distributed blocking flowshop scheduling problem (DBFSP) with new job insertions is studied. Rescheduling all remaining jobs after a dynamic event like a new job insertion is unreasonable to an actual distributed blocking flowshop production process. A deep reinforcement learning (DRL) algorithm is proposed to optimise the job selection model, and local modifications are made on the basis of the original scheduling plan when new jobs arrive. The objective is to minimise the total completion time deviation of all products so that all jobs can be finished on time to reduce the cost of storage. First, according to the definitions of the dynamic DBFSP problem, a DRL framework based on multi-agent deep deterministic policy gradient (MADDPG) is proposed. In this framework, a full schedule is generated by the variable neighbourhood descent algorithm before a dynamic event occurs. Meanwhile, all newly added jobs are reordered before the agents make decisions to select the one that needs to be scheduled most urgently. This study defines the observations, actions and reward calculation methods and applies centralised training and distributed execution in MADDPG. Finally, a comprehensive computational experiment is carried out to compare the proposed method with the closely related and well-performing methods. The results indicate that the proposed method can solve the dynamic DBFSP effectively and efficiently.

Deep learning-based methods have been widely used in the field of rotating machinery fault diagnosis. It is of practical significance to improve the calculation speed of the model on the premise of ensuring accuracy, so as to realise real-time fault diagnosis. However, designing an efficient and lightweight fault diagnosis network requires expert knowledge to determine the network structure and adjust the hyperparameters of the network, which is time-consuming and laborious. In order to design fault diagnosis networks considering both time and accuracy effortlessly, a novel lightweight network with modified tree-structured parzen estimators (LN-MT) is proposed for intelligent fault diagnosis of rotating machinery. Firstly, a lightweight framework based on global average pooling and group convolution is proposed, and a hyperparameter optimisation (HPO) method based on Bayesian optimisation called tree-structured parzen estimator is utilised to automatically search the optimal hyperparameters for the fault diagnosis task. The objective of the HPO algorithm is the weighting of accuracy and calculating time, so as to find models that balance both time and accuracy. The results of comparison experiments indicate that LN-MT can achieve superior fault diagnosis accuracies with few trainable parameters and less calculating time.

A multi-agent iterative optimisation method based on deep reinforcement learning is proposed for the balancing and sequencing problem in mixed model assembly lines. Based on the Markov decision process model for balancing and sequencing, a balancing agent using a deep deterministic policy gradient algorithm, a sequencing agent using an Actor–Critic algorithm, as well as an iterative interaction mechanism between these agents' output solutions are designed for realising the global optimisation of mixed model assembly lines. The exchange of solution information including assembly time and station workload in the iterative interaction realises the coordination of the worker assignment policy at the balancing stage and the production arrangement policy at the sequencing stage for the minimisation of work overload and idle time at stations. Through the comparative experiments with heuristic rules, genetic algorithms, and the original deep reinforcement learning algorithm, the effectiveness of the proposed method is demonstrated and discussed for small-scale instances as well as large-scale ones.

Tube-to-tubesheet welding is widely applied in industrial fields. However, the current tubesheet welding robot still mainly relies on manual tubesheet models. Aiming to solve this problem, this paper proposed an improved direct method to automatically establish a tubesheet semi-dense point cloud model based on a selected monocular camera and a one-dimension (1D) laser rangefinder. Firstly, a laser filtering method was designed to acquire the distance between the camera and tubesheet through the 1D laser rangefinder. Then, from combing the 1D laser rangefinder data with keyframe data, the scale factor was obtained and proceeded to be processed by the Kalman filter to reduce the error. Then, the computed scale factor and all the keyframes were calculated to construct the tubesheet point cloud model through the graph optimisation algorithm. The experimental results showed that the semi-dense point cloud model of the tubesheet could be efficiently established by the proposed algorithm with row error and column error both less than 1 mm, satisfying the welding requirements.

Research on Reconfigurable Battery Systems (RBS) is gaining emphasis over the traditional fixed topology of the battery pack due to its advantages of adapting flexible topology (series-parallel) during its operation in the pack for meeting the non-linear time-dependent load requirements. There could emerge serious issues such as those related to safety due to malfunction of the switching circuit, heat generation from switches during frequent switching of circuits, charging temperature rise, increased charging time, sensing issues arising from the use of low-accuracy voltage/current sensors, state of charge/state of health estimation, and cost issues due to the use of increasing number of switches, fuses, contactors, relays, circuit breakers etc. To address these mentioned issues, the problem of optimal switching circuit topology for RBS is formulated as a mathematical multi-objective optimisation problem. An intelligent system framework based on digital twins is proposed. The proposed framework is further extended to a life cycle management approach that includes the interactions among pack design, pack assembly and operational and recycling levels. This could provide greater access of real-time big data cloud storage to the battery designers, manufacturers and recycling industries, who can make use of it to optimise their designs, systems and operations.

Industry 4.0 has ushered in a new era of digital manufacturing and in this context, digital twins are considered as the next wave of simulation technologies. The development and commissioning of Cyber Physical Systems (CPS) is taking advantage of these technologies to improve product quality while reducing costs and time to market. However, existing practices of virtual design prototyping and commissioning require the cooperation of domain specific engineering fields. This involves considerable effort as development is mostly carried out in different departments using vendor specific simulation tools. There is still no integrated simulation environment commercially available, in which all engineering disciplines can work collaboratively. This presents a major challenge when interlinking virtual models with their physical counterparts. This paper therefore addresses these challenges by implementing a holistic and vendor agnostic digital twin solution for design prototyping and commissioning practices. The solution was tested in an industrial use case, in which the digital twin effectively prototyped cost-efficient solar assembly lines.