Table of Content

25 April 2026, Volume 37 Issue 4

Previous Issue   

Overview and Prospects of Data-driven Low-carbon Design and Manufacturing of Electromechanical Products

WANG Liming, XIAO Xingyuan, LI Fangyi, WANG Xiaoguang, LI Jianfeng, NIE Yanyan, LIU Weitong, LI Liuyuan, WANG Yitong, WANG Boyun, CUI Yuqi

2026, 37(4):  764-779.  DOI: 10.3969/j.issn.1004-132X.2026.04.001

Asbtract ( )   HTML ( )   PDF (3560KB) ( )  

Figures and Tables | References | Related Articles | Metrics

Carbon footprint data served as the core basis for quantifying the full life-cycle carbon emissions of electromechanical products and driving the low-carbon transformation of the manufacturing industries. Focusing on the whole processes of carbon footprint data from acquisition to application， the relevant research approaches were systematically reviewed. The acquisition technologies for multi-source heterogeneous carbon footprint data and the data quality evaluation system were organized， addressing the question of "how data comes". Focusing on “how to use”， applications of data-driven technologies in low-carbon design and manufacturing were elaborated， including data-based carbon footprint correlation modeling， intelligent prediction， generation of low-carbon design solutions， and multi-objective decision-making methods， as well as data-driven manufacturing energy consumption prediction， low-carbon process planning， and intelligent workshop scheduling strategies. Finally， challenges and future directions for data integrity and system integration in low-carbon manufacturing were discussed， offering theoretical references for the green and low-carbon development of electromechanical products.

Eco-design for Additive Manufacturing： Knowledge-driven Framework and Applications

WANG Yanan, PENG Tao, XIONG Yi, WANG Liming, TANG Yunlong, TANG Renzhong

2026, 37(4):  780-791.  DOI: 10.3969/j.issn.1004-132X.2026.04.002

Asbtract ( )   HTML ( )   PDF (14372KB) ( )  

Figures and Tables | References | Related Articles | Metrics

To support designers in developing environmentally sustainable components fabricated by additive manufacturing， the concept of eco-design for additive manufacturing（EcoDfAM） was clarified， and a body of knowledge required in EcoDfAM was constructed. A knowledge-driven EcoDfAM framework was proposed based on knowledge by integrating multiple intelligent technologies. This framework consisted of three layers： knowledge source layer， knowledge model layer， design application layer. Taking the valve body parts of a hydraulic system in the aircraft wing that were printed using the metal powder bed fusion technology as an example， the application analysis and discussion of the proposed framework were carried out. The results show that the framework integrating ontology， machine learning， knowledge graph may effectively integrate the complex multi-domain professional knowledge required by EcoDfAM， and generate eco-design recommendations through flexible reuse of knowledge. This study may be used to develop intelligent design advisor systems， providing appropriate knowledge feedback in different design stages to guide designers， effectively improving the efficiency and quality during the eco-design processes of parts fabricated by additive manufacturing.

Conflict Negotiation Model for Cross-organizational Collaboration in Product Life Cycle Low-carbon Design

TAO Jing, DU Siqi, CHEN Yingyu, ZHANG Yingjie, SHENG Bo

2026, 37(4):  792-801.  DOI: 10.3969/j.issn.1004-132X.2026.04.003

Asbtract ( )   HTML ( )   PDF (1668KB) ( )  

Figures and Tables | References | Related Articles | Metrics

Product life cycle low-carbon design refered many stakeholders， design links， difficulties， and high risks. How to conduct design coordination and conflict negotiation among various stakeholders on different goals of economic， technical and green performance was a bottleneck problem that needed to be solved urgently. The cross-organization collaboration principle of product life cycle low-carbon design was elaborated， and the generation path of low-carbon design conflict negotiation issues in the whole life cycle was clarified. A multi-agent multi-issue negotiation model was built based on multi-agent system for product life cycle low-carbon design and an auto-negotiation decision-making algorithm was built based on NSGA-Ⅱ. The applicability and effectiveness of the proposed negotiated model were verified by numerical simulations， which might provide an effective strategic basis for cross-organization collaborative conflict resolution of product life cycle low-carbon design.

Green Design Knowledge Proactive Recommendation Method Based on Designers’ Lifecycle Dynamic Features

KE Qingdi, NIE Haichuan, HU Jiaqi, HE Haodong, NI Chenxi

2026, 37(4):  802-813.  DOI: 10.3969/j.issn.1004-132X.2026.04.004

Asbtract ( )   HTML ( )   PDF (5822KB) ( )  

Figures and Tables | References | Related Articles | Metrics

In response to the difficulty in acquiring green design knowledge， designer profile matching， and the low efficiency knowledge recommendation throughout the lifecycle design of electromechanical products， various features involved in the designers' green design processes over the full lifecycle， including type， operational， knowledge， and task features， were identified. Then， feature quantization， weight optimization， and dimensionality reduction fusion were conducted. Subsequently， indicator functions for green design knowledge matching， including designer dynamic profile similarity， lifecycle feature distance， and knowledge greenness， were developed， and a corresponding matching model was constructed. Furthermore， a green design knowledge push mechanism was proposed based on the transfer prediction of designer profile， and an active green design knowledge recommendation method driven by full-lifecycle distance prediction and verification was established. Finally， an application validation was conducted using a new refrigerator model with green design requirements in terms of high volume efficiency， reduced material consumption， and low energy use. The results demonstrate that the recommended green design knowledge sets may effectively support designers in achieving green structural optimization and energy-efficient design for electromechanical products.

Large Language Model-driven Knowledge Modeling Method for Low-carbon Product Design

YU Sheng, HE Bin

2026, 37(4):  814-820.  DOI: 10.3969/j.issn.1004-132X.2026.04.005

Asbtract ( )   HTML ( )   PDF (3247KB) ( )  

Figures and Tables | References | Related Articles | Metrics

To address the challenges of multi-source heterogeneity and semantic complexity in low-carbon product design knowledge， a large language model-driven approach for low-carbon design knowledge modeling was proposed. A domain ontology covering product structure， low-carbon factors， and performance constraints was constructed to achieve semantic hierarchical representation from structural design to performance verification. A data annotation paradigm based on large language models was developed， which enables automated semantic labeling of low-carbon design data through a dual-path collaborative mechanism. A contrastive learning-based knowledge extraction model was designed to enhance BERT’s capability in recognizing semantic boundaries and to improve the semantic encoding of the set prediction networks， thereby achieving accurate extraction of multi-entity and multi-relation information. Experimental results show that the proposed method achieves accuracy， recall， and F1 scores of 84.2%， 82.7%， and 83.4%， respectively， providing an intelligent pathway for semantic modeling of low-carbon design knowledge.

Optimization Model Construction Method of CNC Milling Energy Efficiency Based on Specific Energy Values and ELM-AdaBoost under Small Samples

BAO Hong, YANG Shuo, YAO Hang, LI Yapeng

2026, 37(4):  821-830.  DOI: 10.3969/j.issn.1004-132X.2026.04.006

Asbtract ( )   HTML ( )   PDF (2560KB) ( )  

Figures and Tables | References | Related Articles | Metrics

Aiming at the problems of high cost of energy efficiency data acquisition in CNC milling processes and low prediction accuracy of traditional CNC milling energy efficiency model under small sample data， an energy efficiency optimization model was proposed based on specific energy values and extreme learning machine（ELM）-adaptive enhancement algorithm（AdaBoost）. The experimental data were obtained through orthogonal experimental design， a mechanism model was constructed based on specific energy value， combined ELM and AdaBoost to form ELM-AdaBoost data model， and finally integrated the energy efficiency prediction model， which might guarantee the prediction accuracy while effectively reduce the model's demands for data volume. The energy efficiency optimization models were established with the objectives of minimum specific energy value and minimum machining costs， and the optimal processing parameters were solved and optimized by non-dominated sorting genetic algorithm Ⅱ and entropy weight-TOPSIS， and the machining experiments were conducted to verify the feasibility of the proposed method.

Energy Consumption Prediction of Industrial Robots Based on Bayesian Optimized Temporal Convolutional Network

XIAO Wei, ZHANG Cong, CHEN Xubing

2026, 37(4):  831-836.  DOI: 10.3969/j.issn.1004-132X.2026.04.007

Asbtract ( )   HTML ( )   PDF (11630KB) ( )  

Figures and Tables | References | Related Articles | Metrics

To achieve online and efficient prediction of industrial robot energy consumptions， a method was proposed based on Bayesian-optimized TCN. Specifically， TCN was utilized to establish a nonlinear mapping relationship between kinematic parameters and robot energy consumption， which effectively captured the temporal characteristics of energy consumption prediction data. Meanwhile， the Bayesian method was adopted to optimize the hyperparameters in the model， thereby improving the accuracy of the energy consumption prediction model. Ablation experiments and comparative experiments on the IRB 1600-10/145 industrial robots show that， under no-load and 1.5 kg load conditions， the average relative errors of the total energy consumptions of the robot predicted by the proposed method are as 1.04% and 1.78% respectively. These results demonstrate that the proposed method outperforms other commonly used energy consumption prediction models at present.

Pneumatic Atomizer Design and Droplet Characterization for MQL in Metal Cutting

YUAN Yaohui, WANG Chengyong, LI Weiqiu, ZHENG Lijuan, YAN Bingjiang

2026, 37(4):  837-845.  DOI: 10.3969/j.issn.1004-132X.2026.04.008

Asbtract ( )   HTML ( )   PDF (2295KB) ( )  

Figures and Tables | References | Related Articles | Metrics

The atomizer， as the core component of the built-in atomized MQL system， directly affected the atomization efficiency of MQL oil and cutting performance. The atomization mechanism of lubricating oil and the key factors influencing atomizer performances were thoroughly analyzed based on the liquid spray theory herein. A pneumatic atomization-type MQL oil mist atomizer was developed. Through a combination of atomizer performance evaluation methods and experimental research， the coupling mechanism of factors such as throat diameter， suction aperture， and gaps between orifice walls on atomizer performance was clarified. Based on the analysis of the laser spray particle size analyzer， the superior atomization performance of the atomizer with oils of varying viscosities was verified. The particle size number frequency distribution is concentrated between 0.8-10 µm， with the volume frequency distribution peaking at 25 µm and the Sauter mean diameter D₃₂ ranging from 5-12 µm.

Hyper-heuristic Optimization and Decision-making of Hobs and Control Parameters

CAO Weidong, WANG Yuanshuo, LI Minrong, CHEN Fuqi, CHEN Xingzheng, WU Dianjian, HU Kexin

2026, 37(4):  846-854.  DOI: 10.3969/j.issn.1004-132X.2026.04.009

Asbtract ( )   HTML ( )   PDF (3596KB) ( )  

Figures and Tables | References | Related Articles | Metrics

To address the issues of automatic selection of heuristic algorithms in hobbing tools and control parameter optimization， as well as parameter decision-making under the fuzzy expression of users' emphasis on machining performance， a method for optimizing and deciding hobbing tools and control parameters was proposed based on an improved hyper-heuristic algorithm and fuzzy TOPSIS. The spectral clustering algorithm was used to determine the upper and lower limits of parameters based on historical machining datas. With carbon emissions， cutting time， and quality as optimization objectives， an improved multi-objective hyper-heuristic algorithm was used to obtain optimized hobbing parameters （non-dominated solutions）. Based on the user's emphasis on machining performance， fuzzy TOPSIS was employed to rank the optimized hobbing parameters and select the parameters that best met the user's requirements. Experiments verified the feasibility and effectiveness of the proposed method.

Customer-benefit-oriented Energy-efficient Scheduling of Machine Tool Service Resources in Cloud Manufacturing Based on Non-cooperative Game

ZHOU Lirong, CHEN Zihan, WANG Guangcun, KONG Lin, ZENG Guiyuan, REN Yayun

2026, 37(4):  855-865.  DOI: 10.3969/j.issn.1004-132X.2026.04.010

Asbtract ( )   HTML ( )   PDF (3157KB) ( )  

Figures and Tables | References | Related Articles | Metrics

To solve the competition for machine tool resources and conflicts of customer benefits in cloud manufacturing， and to promote the balanced conversion of manufacturing service energy efficiency into customer benefits， a non-cooperative game-based energy-efficient scheduling method for machine tool resources in cloud manufacturing was proposed. a customer-benefit-oriented energy-efficient scheduling model for cloud manufacturing machine tools was constructed by classifying customer preferences into five types： time-sensitive， energy-efficient， cost-effective， quality-focused， and comprehensive. An improved non-cooperative game genetic algorithm was employed to solve the Nash equilibrium. The impact trends of differentiated discount strategies on scheduling outcomes were analyzed based on customer value classification. Simulation experiments conducted on the manufacturing of multiple cylinder piston rod parts for typical construction machinery products from Xuzhou Construction Machinery Group（XCMG） demonstrate that the proposed method increases the average customer manufacturing task benefit by 4.1% and reduces task energy consumption by up to 28%.

Experimental Investigation of Ultrasonically-assisted Energy-saving Surface Drying Techniques for Wire Materials

GU Wenting, CHEN Xiuhong, FENG Shaoke, YAN Lutao, GAO Yaxiang

2026, 37(4):  866-874.  DOI: 10.3969/j.issn.1004-132X.2026.04.011

Asbtract ( )   HTML ( )   PDF (3396KB) ( )  

Figures and Tables | References | Related Articles | Metrics

Conventional wire drying typically relied on high-speed air-jet blowing， which had issues such as extremely thin liquid films on the wire surfaces and a small area affected by the airflow. Applying ultrasonic energy directly to the wire structures so that the solid-liquid interface underwent high-frequency vibration and an ultrasonic atomization effect was produced. An energy-consumption model coupling ultrasonic atomization and high-speed air jets was developed， and comparative experiments were carried out focusing on key parameters such as ultrasonic power and droplet volume. Results show that ultrasound induces rapid droplet spreading and atomization via capillary waves and acoustic cavitation， and the liquid-film removal efficiency is markedly superior to that of air jets. At a droplet volume of 10 \mathrm{\mu }\mathrm{L} and an exposure time of 18 ms， 360 W ultrasound achieves a removal rate of 84.37%， whereas a 450 W high-speed air jet removed only 8.58% of the film. In terms of energy consumption， under conditions producing equivalent removal， ultrasound reduces system energy use by at least approximately 46.7% compared with air jets. This paper demonstrates the feasibility of ultrasonic-assisted rapid， energy-efficient dewatering and drying of wire surfaces.

Digital Twin-driven Performance Modeling and Dynamic Optimization Methodology for Precision Milling Machines

MEI Shulong, XIE Yang, ZHANG Chaoyong, WU Jianzhao, LIU Jinfeng

2026, 37(4):  875-884.  DOI: 10.3969/j.issn.1004-132X.2026.04.012

Asbtract ( )   HTML ( )   PDF (4238KB) ( )  

Figures and Tables | References | Related Articles | Metrics

A digital twin-based dynamic multi-objective optimization method for machining processes was proposed herein. By integrating historical machining data with real-time operational data， a digital twin system was established， comprising geometric， physical， behavioral， and rule-based sub-models. This system combined an Optuna-GBR model and an IMORIME to dynamically adjust machining parameters. The cutting force fluctuations were monitored in real time by the digital twin system. When the fluctuations exceeded the adaptive threshold， a dynamic optimization process was triggered， during which a new Pareto solution set was regenerated and the optimal machining parameter combination was determined using the entropy-weighted technique for order preference by similarity to an ideal solution（TOPSIS） method. Experimental validation under actual machining conditions demonstrates that the dynamic optimization method of the digital twin system achieves a 19.99% reduction in spindle energy consumption， a 29.02% reduction in specific cutting energy， and an 11.22% reduction in machining noise. These results indicate a significant improvement in machining efficiency and a remarkable reduction in spindle energy consumption and machining noises.

Research Progresses on Synergistic Technology of Micro-textured and Minimum Quantity Lubrication on Tool Surfaces

NIU Qiulin, ZHU Chenyi, WU Binghui, ZHANG Shengfeng, GUO Tao, GAO Jingyi

2026, 37(4):  885-899.  DOI: 10.3969/j.issn.1004-132X.2026.04.013

Asbtract ( )   HTML ( )   PDF (5353KB) ( )  

Figures and Tables | References | Related Articles | Metrics

This paper systematically reviewed the research progresses of MQL and micro-textured tool technology in DMMs cutting in recent years， and focused on analyzing the synergistic mechanism and the influences on DMMs cutting performance. Compared with dry cutting， the synergistic effects of micro-textured+MQL may reduce the average cutting force by 25% to 48.3%， reduce the cutting temperature by 20% to 50%， improve the surface roughness by 15% to 40%， and extend the tool life by 1.5 to 2.4 times. The existing research is yet to be strengthened in the following aspects： first， quantitative theoretical model of synergistic mechanism； second， adaptability analysis for different materials； and third， experimental verification of long-term stability under industrialized conditions. Future research should focus on the construction of quantitative synergistic models based on thermodynamic and tribological behaviors， especially the optimization of the applications in the machining of high-strength and high-temperature materials such as titanium-based alloys and aluminum-based alloys， and the verification of the industrial application value through the evaluation of the long-term performance under actual working conditions.

Interpretable Modeling and Optimization of Laser Hardening Process Parameters for QT550-5

LIANG Qiang, CHEN Hong, ZHENG Yinpeng, WANG Bing, DU Yanbin, LONG Shuai

2026, 37(4):  900-912.  DOI: 10.3969/j.issn.1004-132X.2026.04.014

Asbtract ( )   HTML ( )   PDF (7418KB) ( )  

Figures and Tables | References | Related Articles | Metrics

Aimed to achieve laser surface hardening and optimize processing parameters for nodular cast iron QT550-5， a finite element model coupling the temperature and phase transformation fields was developed herein. Using laser power， scanning speed， and overlap rate as experimental variables， and targeting the hardened layer depth and molten layer depth as optimization objectives， Latin hypercube sampling was first employed for the experimental design. A Bayesian-optimized multi-task neural network prediction model was constructed based on the experimental data. SHAP were introduced for interpretability analysis to clarify the contribution mechanism of various parameters to the hardening results. Subsequently， the multi-objective hippopotamus optimization algorithm was used for parameter optimization. A comprehensive evaluation system integrating the entropy weight method and the technique for order preference by similarity to ideal solution was established to rank the non-dominated solution set and determine the optimal parameter combination. Experimental validation under the optimal parameters confirmes the significant surface hardening effectiveness in QT550-5.

Research on Equivalent Models for Dynamic Response Analysis of Self-piercing Rivet Joints

QIU Kangbo, SONG Haisheng, ZHANG Shenglan, GUO Haotian, YANG Na, WANG Wenxin

2026, 37(4):  913-919.  DOI: 10.3969/j.issn.1004-132X.2026.04.015

Asbtract ( )   HTML ( )   PDF (2573KB) ( )  

Figures and Tables | References | Related Articles | Metrics

The vehicle bodies were subjected to dynamic loads arising from complex road conditions during actual operations， current research on SPR primarily focused on static load response analysis， with insufficient attention given to dynamic load response analysis. A dynamic modelling and analysis method for SPR joints was proposed based on a mass-spring system herein. ABAQUS finite element analysis was employed to identify the tensile and shear stiffness parameters of the SPR joints. The connection between the SPR joints and the sheet metal was abstracted as an interaction between mass-spring systems， establishing a three-degree-of-freedom nonlinear dynamic response model for SPR. Finite element simulation was used to validate the SPR dynamic response equivalent model. Research indicates that the SPR three-degree-of-freedom nonlinear dynamic response model enables efficient and accurate prediction of SPR joints dynamic responses.

Fine-grained Carbon Emission Accounting in Aluminum Casting Production Processes Based on Data Allocation

WANG Zhihui, PENG Tao, LIU Weipeng, TANG Renzhong

2026, 37(4):  920-928.  DOI: 10.3969/j.issn.1004-132X.2026.04.016

Asbtract ( )   HTML ( )   PDF (1999KB) ( )  

Figures and Tables | References | Related Articles | Metrics

A data allocation method that accommodated the heterogeneous on-site data collection capabilities of aluminum casting enterprises was proposed， enabling fine-grained carbon emission accounting at the single aluminum casting level， differentiated by product， model， and batch， without requiring additional investment in data collection infrastructure. According to the granularity and interval of carbon emission-related production data， enterprise data acquisition capabilities were categorized into three levels， and corresponding allocation strategies were developed for each level to derive fine-grained energy and material activity levels for individual aluminum castings. Product carbon emissions were calculated using the emission factor method， and a correction mechanism was established to incorporate the effects of internal scrap recycling and external scrap sales. A case study on aluminum wheel production was conducted to validate the proposed method. The results demonstrate that the method effectively overcomes the bias induced by the constant operating condition assumption inherent in traditional literature-based approaches， thereby improving the accuracy of carbon accounting results.

A Low-carbon Process Optimization Method for Parts Driven by Intelligent Parsing of Manufacturing Features

ZHANG Lei, ZHANG Zhen, LIU Runze

2026, 37(4):  929-938.  DOI: 10.3969/j.issn.1004-132X.2026.04.017

Asbtract ( )   HTML ( )   PDF (3450KB) ( )  

Figures and Tables | References | Related Articles | Metrics

To address the challenges of optimizing low-carbon manufacturing processes， a structured feature analysis and an intelligent mapping-driven method were proposed to link part design models to carbon emissions assessment during the manufacturing phases. First， the geometric and topological information were extracted by parsing STEP files， and an extended attributed adjacency graph and the corresponding matrix representation were constructed， then feature submatrix matching against a feature library was utilized to identify typical manufacturing characteristics. Second， based on the identification results， corresponding parameter-extraction rules were triggered to obtain geometric feature dimensions， thereby achieving a structured representation of manufacturing information. Finally， a collaborative quantification model linking manufacturing carbon emissions and processing time was constructed. A multi-objective process optimization framework targeting low-carbon efficiency was established， and the NSGA-Ⅱ algorithm was employed to determine the Pareto-optimal solution set， providing decision support for low-carbon manufacturing process planning of the parts.

Energy Consumption Prediction of Industrial Robots under High-load Dynamic Conditions

SUN Yue, HUANG Hui, YIN Fangchen

2026, 37(4):  939-947.  DOI: 10.3969/j.issn.1004-132X.2026.04.018

Asbtract ( )   HTML ( )   PDF (10846KB) ( )  

Figures and Tables | References | Related Articles | Metrics

The power of industrial robots under high-load and highly fluctuating processing conditions showed non-stationary and multi-source coupling characteristics， which led to the problems of reduced accuracy and stability of energy consumption prediction models under cross-operation conditions. Multi-source time series data were collected from the self-built processing experimental platform. The heterogeneous data were synchronized and resampled through timestamps， and power tags were constructed using sliding windows. The prediction results of random forest， gradient boosting tree， support vector regression， multi-layer perceptive machine and two fusion structure models under multiple working conditions were compared. The results show that the energy consumption prediction result of the gradient boosting tree + support vector regression fusion model is the best in the working conditions without participation in training， with an average absolute error of 3.73%. The research reveales the predictive characteristics of different models under high-dynamic processing conditions， which may provide technical support for energy efficiency modeling， process optimization and green operations of high-load processing of the industrial robots.

Carbon Emission Prediction and Uncertainty Analysis Method for Machining Processes Driven by Manufacturing Scenarios

KONG Lin, ZENG Qingliang, WANG Liming, LI Fangyi, ZHANG Xin, LU Zhenguo, WANG Guijie

2026, 37(4):  948-958.  DOI: 10.3969/j.issn.1004-132X.2026.04.019

Asbtract ( )   HTML ( )   PDF (4024KB) ( )  

Figures and Tables | References | Related Articles | Metrics

Conventional prediction methods faced challenges such as multi-source heterogeneity and strong parameter uncertainty， so equipment， process， and resource-related factors were integrated to identify and define manufacturing scenarios， enabling the unified representation and description of carbon emission influences. The ensemble mechanism of random forest decision trees was combined with Bayesian adaptive hyperparameter optimization to establish a three-stage prediction framework “feature selection， model training， parameter tuning” for the high-efficiency prediction of carbon emissions. A Monte Carlo-Bayesian optimized random forest approach for uncertainty analysis was developed， where sensitive carbon emission parameters were identified and their impacts were quantified to enhance reliability through targeted parameter optimization. A case study on wind turbine blade machining demonstrated the effectiveness of the proposed method. The results show excellent agreement between predicted and actual carbon emissions. After uncertainty analysis， the coefficient of variation is reduced by 0.0347， significantly improving the reliability of the prediction results and supporting more robust decision-making.

A Novel SiC MOSFET Lifetime Prediction

HU Yawei, FANG Xiang, YIN Chuanan, LIN Zijun, LIN Xiaowei

2026, 37(4):  959-966.  DOI: 10.3969/j.issn.1004-132X.2026.04.020

Asbtract ( )   HTML ( )   PDF (7094KB) ( )  

Figures and Tables | References | Related Articles | Metrics

To address the reliability challenges faced by SiC MOSFETs in high-frequency， high-temperature， and high-power density applications， a novel lifetime prediction method was proposed that integrated a CNN， an ECA mechanism， and a BiLSTM. This method used the drain-source on-state voltage drop as the core degradation feature， incorporated preprocessing strategies such as outlier removal， normalization， and exponential smoothing， and reconstructed the degradation time series through a sliding window to achieve effective modeling under small sample conditions. Comparative experimental results demonstrate that the proposed method offers significant advantages in prediction accuracy， stability， and robustness.

GraphRAG-based Disassembly Sequences Planning for End-of-life Power Batteries

WANG Hang, YAN Wei, ZHANG Xumei, ZHU Shuo, JIANG Zhigang, ZHU Zerui

2026, 37(4):  967-976.  DOI: 10.3969/j.issn.1004-132X.2026.04.021

Asbtract ( )   HTML ( )   PDF (3829KB) ( )  

Figures and Tables | References | Related Articles | Metrics

To address the challenges of low-efficiency disassembly sequence generation caused by the lack of knowledge reusability of end-of-life power batteries， a disassembly sequence planning method was proposed based on GraphRAG， by integrating the complementary strengths of KGs in structured knowledge representation and LLMs in semantic reasoning. First， a KG label-matching-based subgraph generation method was proposed， utilizing the Cypher query language to form custom disassembly sequence subgraphs for specific battery models. Second， a coarse-grained disassembly knowledge retrieval mechanism employing hybrid retrieval and re-ranking was established to locate target battery components precisely. Finally， a fine-grained retrieval model of disassembly knowledge was constructed based on multi-hop reasoning of hierarchical constraint relations. By extracting the disassembly sequence information associated with the components， the intelligent generation of disassembly sequences was achieved using the large language model. The experimental results indicate that the proposed method achieves an accuracy of 93.9% in disassembly sequence generation across five mainstream power batteries， demonstrating the excellent feasibility and effectiveness.

Selective Disassembly Sequence Planning for Retired Electromechanical Products Based on Heterogeneous Graph with Improved Proximal Policy Optimization

GUO Hongfei, FU Wenjie, REN Yaping

2026, 37(4):  977-986.  DOI: 10.3969/j.issn.1004-132X.2026.04.022

Asbtract ( )   HTML ( )   PDF (2123KB) ( )  

Figures and Tables | References | Related Articles | Metrics

To address the issues of complex physical modeling， poor adaptability， and insufficient algorithm generalization in the current selective disassembly sequence planning problem， a structured heterogeneous graph modeling method was proposed， which combined with an adaptive proximal policy optimization algorithm to achieve efficient disassembly sequence optimization. Through the structured heterogeneous graph modeling， the multi-constraint relationships of the product components were unified， providing a more expressive state representation for subsequent optimization. Additionally， in the optimization algorithm， advantage function normalization and entropy regularization mechanism were introduced to standardize the data distribution inconsistency caused by dimensional differences across different training stages， while adaptively adjusting the exploration intensity during the training processes to enhance the model's training stability and generalization ability. Experimental results show that the introduction to advantage function normalization significantly improves the algorithm's convergence speed and training stability， while the entropy regularization mechanism enhances the algorithm's exploration ability. Compared with traditional deep reinforcement learning algorithms， the proposed method performes better in terms of convergence and the quality of the optimal policy.

Research Progresses on Reliability Design of Remanufactured Electromechanical Products with Multi-life Characteristics across Whole Working Ranges

YANG Jie, JIANG Zhigang, ZHU Shuo, CHEN Xin, ZHANG Hua

2026, 37(4):  987-998.  DOI: 10.3969/j.issn.1004-132X.2026.04.023

Asbtract ( )   HTML ( )   PDF (3842KB) ( )  

Figures and Tables | References | Related Articles | Metrics

The low reliability of remanufactured electromechanical products stemmed from multiple factors： complex and variable operational scenarios leading to performance disparities across working ranges， the imbalance in multi-attribute life（physical， technical， and economic） of various substance-field units within the products， and the non-uniform evolution of reliability across different attribute life under varying working ranges. These factors collectively posed significant challenges to enhancing product reliability. In response， research on whole working range reliability design was emerged. By exploring the design connotations， a framework for whole working range reliability design under multiple life characteristics was advanced. Building upon this framework， representative studies， ranging from the generation to the optimization of reliability design schemes， were reviewed. These included the evolution of whole working range reliability based on life characteristics， reliability design and multi-objective reliability optimization for remanufactured products. Furthermore， current research challenges were discussed. Finally， key research focuses were summarized， and future directions for reliability design were proposed.

Decision-making of Disassembly Sequence Alternatives for End-of-life Mechanical Products Based on Intuitionistic Normal Cloud

ZHANG Honghao, HAN Tao, ZHANG Shaolin, WANG Danqi, TIAN Guangdong, HUANG Zhongwei

2026, 37(4):  999-1006.  DOI: 10.3969/j.issn.1004-132X.2026.04.024

Asbtract ( )   HTML ( )   PDF (5447KB) ( )  

Figures and Tables | References | Related Articles | Metrics

To address the challenges of disassembly end-of-life mechanical products in complex and fuzzy environments， which complicated the sequencing of disassembly alternatives， a hybrid multi-criteria decision-making method was proposed based on INC fuzzy theory and incorporating integrated subjective and objective weights. According to dismantling features， a comprehensive optimization criteria system for evaluating disassembly sequence alternatives of end-of-life mechanical products was constructed across the environmental， economic， and human-related dimensions. Combining Monte Carlo simulation with cloud model distance， improvements were made to the cross-entropy and full consistency method to determine criteria fusion weights. The INC-measurement of alternatives and ranking according to compromise solution method was subsequently employed to rank the disassembly sequence alternatives and determine the optimal solution. Finally， an end-of-life coal-fire power products was employed as a case study to verify the scientific rigor and effectiveness of the proposed method by comparing the impacts of different weighting strategies and decision-making methods on the final results.

Reliability Dynamic Prediction Method for Remanufactured Products Based on Data-model Integration and Transfer

FENG Yukang, ZHU Shuo, JIANG Zhigang, YAN Wei, ZHANG Hua

2026, 37(4):  1007-1015.  DOI: 10.3969/j.issn.1004-132X.2026.04.025

Asbtract ( )   HTML ( )   PDF (2592KB) ( )  

Figures and Tables | References | Related Articles | Metrics

To address the problems that the reliability data samples of remanufactured products were scarce， leading to difficulties in accurately predicting their reliability status during service， a dynamic reliability prediction method for remanufactured products was proposed， integrating substance-field degradation data from similar products with model transfer fine-tuning. Firstly， the “physical form” and “field properties” degradation indicators affecting product reliability were analyzed using the substance-field model. Then， a comprehensive degradation index reflecting the multi-dimensional substance-field degradation characteristics of products was constructed using a linear regression model， and a three-stage similarity calculation method was designed to screen and transfer historical degradation data from similar products for sample expansion. Secondly， to address the spatial coupling and temporal dependency characteristics of the historical substance-field degradation data of similar products， a reliability prediction model was established based on a convolutional long short-term memory neural network. Furthermore， the parameters of the prediction model were dynamically adjusted through deep transfer learning techniques to improve the prediction accuracy for the reliability of remanufactured products under personalized service scenarios. Finally， the proposed prediction method was validated using a remanufactured spindle system as a case study， and the coefficient of determination （R²） of the prediction results reache 0.92， which indicates the effectiveness of the method.