Computers in Industry最新文献

The current surge in the need for Li-ion batteries to power electric vehicles has also translated in a need for more advanced models that can predict their behavior, but also quantify the uncertainty in their predictions, given the amount of variables involved and the varying operating conditions. This manuscript proposes a new Bayesian physics-informed recurrent neural network, where the battery discharge curve is described using the Nernst and Butler–Volmer equations, the activity correction term within such equations is modeled with two multilayer perceptrons, and approximate Bayesian computation by subset-simulation is used to train the weights, bias and the physical parameters representing the maximum charge available and the internal resistance. The challenges found during the adaptation and implementation of the Bayesian training algorithm to the recurrent physics-informed cell are described, along with the approaches proposed to overcome them. The performance of the Bayesian hybrid model presented in this paper has also been evaluated using data from NASA Ames Prognostics Data Repository, and the results show comparable accuracy to the standard approach with backpropagation, and a flexible and realistic quantification of the uncertainty. Furthermore, the uncertainty related to the physical parameters of the hybrid model can be evaluated in semi-isolation of the weights and bias of the MLPs, providing a sensitivity tool to assess the relative importance between different parameters.

In mechatronic machine design and development, it is no longer enough to think about machine functionality and integration as machines are increasingly digitalized. Virtual upgrades are being made to manufacturing systems to keep up with the need for faster product cycles, higher quality, and the introduction of Industry 4.0 technologies. The design and development of new mechatronic discrete manufacturing machines (MDMM) should thus include these characteristics in their design. However, most machine builders do not have the capabilities and resources to do virtual engineering (VE) at the required level, which means these machines are made with limitations or sometimes without their virtual counterparts. Reusable VE MDMM modularization allows machine builders to obtain these competencies quickly and with fewer resources. This research proposes developing adaptable digital twins (DT) by modularizing all virtual and physical mechatronic machine aspects. DTs are well-explored in literature, but re-engineering them requires massive resources and is often unviable. We introduce a new DT-based approach that allows machine builders to quickly re-engineer, adapt, and test machines, given its modular confined approach. Although VE on different abstraction levels still must be developed, confined modularization allows hiding the complexity into modules rather than addressing the entire machine simultaneously. Building machines through modularization is thus an investment, as machine builders and other stakeholders will be able to use and reuse them later for other machines, reducing the overall resources that go into the development. The paper shows how to develop adaptable DT machines using Siemens tools related to virtual engineering.

In the manufacturing industry, Augmented Reality (AR) has shown significant potential in enhancing operators’ capabilities while performing inspection and assembly activities. However, the augmented visualization of virtual models on physical components can present challenges and potential misunderstandings, as the visualization mode greatly influences the perception of components and the amount of information received. This study investigates the impact of rendering techniques on user performance during industrial assembly inspection tasks. In particular, a set of rendering techniques, suitable for industrial AR applications, have been selected and implemented through a dedicated AR tool. The selected AR rendering techniques have been compared, in terms of qualitative and quantitative metrics, in order to test their efficiency and effectiveness for industrial assembly inspection activities. 17 domain experts and 33 representative users have been involved in the experimental study which outcomes reveal that the rendering techniques play a significant role in assisting operators for identifying design discrepancies in industrial products. Furthermore, a correlation has been observed between the two AR rendering techniques best suited to support industrial inspection activities and the types of assembly errors detected.

Structured enterprise information systems such as Enterprise Resources Planning (ERP) and Product Lifecycle Management (PLM) have reached a maturity plateau and are storing up to hundreds of millions of objects and links. Such data is crucial for enterprise processes and operations. They are frequently the target of data transformation such as migration to a new data system, re-organisation according to new business paradigms, cleansing, purge and archive, etc. To make data transformation manageable, iterative, and achievable, it requires a divide and conquer strategy therefore producing loosely coupled data packages. Most data migration methods recommend divide and conquer strategy but do not explain how to produce these loosely coupled data packages. This paper outlines there exist two different approaches relying on a wide range of algorithms: clustering and community detection. Also, data package must be PLM business meaningful and fit into a mesoscopic scale to provide operational and achievable options for data transformation. Finally, a PLM specific algorithm is proposed for pre-processing data before clustering. A multi-pass tooled-up method able to combine and sequence data clustering approaches/algorithms has been developed for this purpose: Data Systemizer (D6). Using graph-based clustering metrics will help to assess the benefit of multi-pass data clustering approach and provide some principles to select right clustering approaches/algorithms chain.

In the context of Industry 4.0, the design of efficient Material Handling Systems (MHS) plays a critical role in optimizing industrial operations and enhancing productivity. The integration of advanced technologies, automation, and complex systems has revolutionized industrial processes, emphasizing the need for defining coherent MHS alternatives prior to the deployment of a specific solution. This paper presents an integrated MHS alternative generation approach that combines MHE allocation, fleet sizing, and key performance indicator (KPI)-based evaluation to enhance MHS. The proposed approach deploys a Constraint Satisfaction Problem (CSP) framework, which defines the conditions that must be satisfied by the values assigned to the MHE allocation and fleet sizing variables. The constraints capture the physical and operational characteristics of MHE and MHS, ensuring that the generated alternatives adhere to the specific requirements and limitations of the system. Once a set of feasible MHS alternatives complying with various constraints is established, further assessments can be carried out to gauge the dynamic behavior of these systems using methods like Discrete Event Simulation. The paper emphasizes the importance of incorporating a broader range of data categories, such as layout constraints, product characteristics, and Industry 4.0 challenges, to enhance the effectiveness and accuracy of the MHS design process. To evaluate the generated alternatives, a set of KPIs is proposed, encompassing operational, economic, environmental, safety, and security aspects. These KPIs enable decision-makers to assess alternative performance across multiple metrics, aiding in the identification of relevant solutions aligned with company objectives. The incorporation of KPIs provides a pre-evaluation of MHS alternatives, streamlining subsequent simulation and multicriteria decision-making processes. This research acknowledges the importance of alternative generation as an integral part of the broader MHS design problem.

The recent convergence between two industrial transitions towards digitalization on the one side and servitization on the other side led to the new business strategies of digital servitization and smart PSS delivery. While inheriting from the previous scientific literature on PSS, because of the multiple impacts of digitalization in the overall system, the processes of ensuring the design and engineering of smart PSS solutions poses new challenges. This research addresses the specific needs to develop conceptual prototypes of smart PSS value offers, at early stages of the design process. The paper presents the development and experimentation of a modelling language and its associated modelling toolkit (sPS²Modeller). The application case study addresses the design of a smart PSS in the field of heating appliances, developed in collaboration with the company elm.leblanc, Bosch Group – France.

In the context of an increasingly automated and personalized manufacturing mode, efficient assembly sequence planning (ASP) has emerged as a critical factor for enhancing production efficiency, ensuring product quality, and satisfying diverse market demands. To address this need, our study first transforms the assembly topology and process into a weighted precedence graph, wherein parts represent nodes, and the assembly interconnections between parts constitute weighted edges. Then, we formulate the quantitative models of semantic knowledge, encompassing three facets: assembly direction changes, assembly stability, and part assembly interference, and thus constructs a heuristic function. We propose a novel dynamic graph learning algorithm, i.e., assembly-oriented graph attention sequence (A-GASeq), utilizing the heuristic information as edge weights of the assembly graph structure to incrementally direct the search towards optimal sequences. The performance of A-GASeq is first evaluated utilizing three key metrics: area under the receiver operation characteristic curve (AUC), precision score, and time consumption. The results reveal the superiority of our model over competing state-of-the-art graph learning models using a real-world dataset. Concurrently, we apply the algorithm to actual industrial products of diverse complexity, thereby demonstrating its broad utility across different complex products and its potential for addressing complex assembly sequence planning problems in the field of smart manufacturing.

Continuous high-frequency wood drying, when integrated with a traditional wood finishing line, allows correcting moisture content one piece of lumber at a time in order to improve its value. However, the integration of this precision drying process complicates sawmills logistics. The high stochasticity of lumber properties and less than ideal lumber routing decisions may cause bottlenecks and reduces productivity. To counteract this problem and fully exploit the technology, we propose to use reinforcement learning (RL) for learning continuous drying operation policies. An RL agent interacts with a simulated model of the finishing line to optimize its policies. Our results, based on multiple simulations, show that the learned policies outperform the heuristic currently used in industry and are robust to sudden disturbances which frequently occur in real contexts.

Asset management and digitalization are two timely research topics, especially for small and medium-sized enterprises (SMEs). SMEs continue to struggle with implementing digital technologies while retaining effective asset management processes within maintenance and after-sales service. Therefore, this paper develops a conceptual framework that integrates technological and organizational factors for digitalizing maintenance processes in SMEs. Furthermore, this paper steps away from the prevalent high-tech narrative, to explore how SMEs practically can implement digital technologies amidst common constraints. A focal point of this study is leveraging existing technologies and data for low-cost optimizations as a practical steppingstone for SMEs, bridging the gap between high-tech digital discussions and SMEs' real-world scenarios. The framework addresses three theoretical challenges in the existing literature, drawing on multiple case studies, and employing pragmatic research methods. It offers a conceptual tool for assessing the digital maturity of maintenance processes in SMEs and identifying areas for improvement. Based on the framework, low-performing domains across the case studies are pinpointed for future research. These domains encompass technical competencies within information systems, strategic efforts, and human capital within prototyping, and organizational culture within digital data collection and analysis.

Metal-based Additive Manufacturing (AM) can realize fully dense metallic components and thus offers an opportunity to compete with conventional manufacturing based on the unique merits possible through layer-by-layer processing. Unsurprisingly, Machine Learning (ML) applications in AM technologies have been increasingly growing in the past several years. The trend is driven by the ability of data-driven techniques to support a range of AM concerns, including in-process monitoring and predictions. However, despite numerous ML applications being reported for different AM concerns, no framework exists to systematically manage these ML models for AM operations in the industry. Moreover, no guidance exists on fundamental requirements to realize such a cross-disciplinary platform. Working with experts in ML and AM, this work identifies the fundamental requirements to realize a Machine Learning Operations (MLOps) platform to support process-based ML models for industrial metal AM (MAM). Project-level activities are identified in terms of functional roles, processes, systems, operations, and interfaces. These components are discussed in detail and are linked with their respective requirements. In this regard, peer-reviewed references to identified requirements are made available. The requirements identified can help guide small and medium enterprises looking to implement ML solutions for AM in the industry. Challenges and opportunities for such a system are highlighted. The system can be expanded to include other lifecycle phases of metallic and non-metallic AM.