In the global pursuit of carbon neutrality, the manufacturing industry is under increasing pressure to reduce energy waste. Excess consumption not only depletes resources but also hinders sustainable development. Accurate energy consumption prediction is therefore essential for scientific production scheduling and resource allocation, enabling loss reduction, efficiency improvement, and environmental performance enhancement. However, the complexity of modern manufacturing environments results in energy consumption data that is high-dimensional, noisy, and strongly spatiotemporal, which poses challenges to traditional prediction methods. To address these issues, this paper constructs an energy consumption behavior model considering key factors such as equipment status, processing techniques, and environmental conditions. A comprehensive feature analysis and data preprocessing are carried out to identify the key factors influencing consumption. Based on this, an optimization model is proposed that integrates an improved Variational Autoencoder (VAE) with an enhanced Graph Attention Network (GAT). VAE extracts compact latent representations from high-dimensional noisy inputs, suppressing redundancy while preserving essential patterns. GAT then captures complex spatiotemporal dependencies among energy-related features, thereby revealing intrinsic consumption dynamics. Experimental evaluations on both public and real-world datasets demonstrate that the proposed VAE-GAT model achieves superior prediction accuracy and generalization compared with other deep learning baselines. This approach provides a reliable foundation for energy management and contributes to advancing green intelligent manufacturing.

Today's automotive assembly is highly automated, and automatic identification of items could enable an even higher level of automotive assembly automation and more flexibility. For this reason the radio-frequency identification (RFID) technology has received serious backing from the automotive manufacturer. In this paper, a new kind of conveyance called friction drive conveyor used in automotive assembly lines is introduced at first, and practical applications of RFID in Painted Body Storage (PBS) and Assembly Shop equipped with friction drive conveyors are described in detail. As a mobile data carrier through a mixed-model assembly line, an RFID tag specifying the car model information is attached to a skeleton car or a carriage, which can enable automatic PBS input/output control and help the assembly operators to achieve better operational efficiency.

Reverse engineering is to determine the relevant form and functions and their relationships for a component or product in order to generate a complete engineering representation. Normally reverse engineers are focusing on the geometric model from point cloud data that represent the form of the physical object being considered. This in essence is not reverse engineering, but geometric reverse modeling. Actually the original intrinsic knowledge is located in the function model of the existing product, and the typical reverse geometric modeling and redesign may cause the problem of knowledge property. In this paper, it is considered that the reverse engineering shall be extended to the functional modeling phase, and a functional reverse design method is proposed to overcome the shortcoming of geometric reverse modeling and redesign. Functional reverse design is conducted through two sequential steps: functional reverse engineering and functional reengineering. Functional reverse engineering is to reproduce the original function model through a form-to-function mapping procedure, and functional reengineering is to improve or update the mapped original function model for derivative function modeling. A case study is offered to illustrate the proposed functional reverse design method.

To ensure efficient knowledge capture, tracing, sharing and reuse to support product design and development, it is important to find a structured way to organize the past experience and information. Design structure matrix (DSM), a structured method which has advantages on representing and analysing interaction relations between system elements (such as development tasks, design parameters, architecture concepts, and organizational teams), is proposed to be suitable means to define, capture and organize and distribute the system level knowledge during the product development process. The captured knowledge through DSM can improve understanding of the design routes and design history by linking designed items to rationales, decisions and assumptions behind them. Meanwhile, the knowledge captured through DSM could assist on predicting changes on existing solutions, reusing of the existing solutions in new projects, and educational process for inexperienced designers. How to capture, trace and manage the design knowledge through DSM is presented in detail, and a DSM-based knowledge management system has been developed.

It is widely acknowledged that the automotive industry is more than ever obliged to improve its development strategy according to the increasing pressure of product innovation and complexity, the changing market demands and increasing level of customer awareness. Due to the complex development cycle, the automotive OEM has begun to adopt the supplier integration into its product development process. To respond to this trend, the collaboration and partnership management between the automotive OEM and associated suppliers need to be investigated. Regarding the depth of collaboration, the integration of supplier into automotive OEM process chain has been defined in two ways, quasi supplier integration and full supplier integration. To enable the success of supplier integration, this paper has investigated how to control the collaboration between the automotive OEM and its suppliers, through deciding on an appropriate supplier integration way. The collaboration tools enabling supplier integration for automotive development have been proposed. Taking the tool supplier as a case study, a web-based system called "Cyberstamping" has been developed to realize the collaborative supplier integration for automotive product design and development.

This paper reports on a collaborative design environment to facilitate active die-maker involvement into metal stamping part design. Using the agent-based approach, a multi-agent based system is constructed to integrate die-maker's activities into customer part design process within a collaborative and concurrent environment. The overall architecture incorporates three agent communities: part design agent, die-maker involvement agent, and coordination agent. Each agent community has a facilitator which provides an intermediary between a local collection of sub-agents and remote agents through two main services: routing outgoing messages to the appropriate destinations and translating incoming messages for consumption. A KQML/XML communication and data exchange method between agents is proposed. A case study is presented to illustrate how the die-maker is involved and cooperates with the part designer to make an optimal stamping part design