Table of Content

25 July 2023, Volume 34 Issue 14

Previous Issue    Next Issue

A Wafer Cycle Processing Time Prediction Method Incorporating Double Attention Mechanism and Parallel GRU

DAI Jiabin, ZHANG Jie, WU Lihui

2023, 34(14):  1640-1646.  DOI: 10.3969/j.issn.1004-132X.2023.14.001

Asbtract ( )   PDF (4600KB) ( )  

References | Related Articles | Metrics

Low efficiency and low prediction accuracy were caused by the large scale of production feature data, complex correlation among features, and strong correlation of feature samples in wafer fabrication processes, so a wafer processing cycle prediction method integrating double attention mechanism and parallel GRU was proposed. Firstly, Relief-F algorithm was used to reduce the dimensionality of production feature data. Secondly, a fuzzy C-mean algorithm was used to cluster the process similarity of data samples and design a parallel GRU network to explore the strong correlation among wafer feature samples. Finally, a double attention mechanism was designed to learn the complex correlation information within key features and among features and processing cycle. The experimental results show that the proposed method may effectively reduce the prediction training time and improve the prediction accuracy.

Hybrid Flow Shop Scheduling Problems with Unrelated Parallel Machine Solved by Improved Adaptive Genetic Algorithm(IAGA) with ITPX

ZHENG Kun, LIAN Zhiwei, GU Xinyan, ZHU Changjian, XU Hui, FENG Xueqing

2023, 34(14):  1647-1658，1671.  DOI: 10.3969/j.issn.1004-132X.2023.14.002

Asbtract ( )   PDF (4458KB) ( )  

References | Related Articles | Metrics

Aiming at the hybrid flow-shop scheduling problems, an adaptive genetic algorithm with ITPX was proposed. Firstly, the solution performance of two-points crossover（TPX） was improved by exacting point taking method. Secondly, adaptive selection probability was demonstrated based on hormonal regulation guiding convergence trend of populations. Then, a pool of high-quality chromosomes and a memory factor were established to record the high-quality chromosomes during population evolution, and two different regional crossovers were implemented. Experimentsal results show that ITPX may save optimization time and improve solution performance; the adaptive probability may enhance convergence; ITPX-IAGA may reduce solution time by more than 40% and improve solution performance.

Time-series Correlation Prediction of Quality in Process Production Processes Based on Deep TCN and Transfer Learning

YIN Yanchao, SHI Chengjuan, ZOU Chaopu, LIU Xiaobao

2023, 34(14):  1659-1671.  DOI: 10.3969/j.issn.1004-132X.2023.14.003

Asbtract ( )   PDF (13154KB) ( )  

References | Related Articles | Metrics

 To address the problems which were difficult to accurately predict production quality due to the temporal coupling of multiple processing parameters in process production, a fast and efficient production quality prediction method was proposed based on deep TCN networks and migration learning. With a sequence-to-sequence learning structure, a deep TCN and a temporal attention mechanism formed the encoding component for extracting key temporal features from multiple sources, while a residual long short term memory network formed the decoding component for simultaneous extraction of quality temporal information, and migration learning was introduced to address the adaptability of the prediction model to online production quality prediction. The experiments show that the proposed method has significant advantages in prediction accuracy and stability, and has high prediction accuracy and computational efficiency in predicting small sample data.

Supply Chain Inventory System Optimization Model under Demand Disturbances

WU Yingnian, ZHANG Jing, LI Qingkui, JIAO Shuai,

2023, 34(14):  1672-1682，1700.  DOI: 10.3969/j.issn.1004-132X.2023.14.004

Asbtract ( )   PDF (7177KB) ( )  

References | Related Articles | Metrics

 Aiming at the supply chain inventory systems under demand disturbances, a supply chain inventory system optimization model was designed combining improved sliding mode controller and disturbance observer. A dynamic inventory model of a three-echelon networked supply chain system was established based on the product operation logic of the supply chain system. An optimization model combining improved sliding mode controller was designed based on adaptive exponential reaching law and disturbance observer. The model could suppress the influences of demand disturbances on the supply chain inventory system under the premise of ensuring system stability. Simulation comparison experiments verified the effectiveness of the optimization model.

Strategy of Group Maintenance of Multiple Equipment in Stages from the Perspective of Outsourcing Contracts

ZHANG Xinyu, LIU Qinming, YE Chunming, XIE Shirui

2023, 34(14):  1683-1692.  DOI: 10.3969/j.issn.1004-132X.2023.14.005

Asbtract ( )   PDF (2101KB) ( )  

References | Related Articles | Metrics

In order to solve the maintenance outsourcing problems of multi-equipment series production system, a strategy of stage group maintenance was put forward. The average failure response time introduced by third-party maintenance features and system availability were combined to measure the satisfaction of the factory to the maintenance service provider. Based on the satisfaction segmentation, a segmented profit model under the reliability constraint was constructed. The sparrow search algorithm of multi-strategy integration was used to find out the optimal grouping and maintenance planning corresponding to different stage strategies, and the maintenance planning was developed accordingly for reference by both parties of the maintenance outsourcing contracts. The results of the example show that the four-stage group maintenance strategy has obvious advantages over the traditional maintenance modes. 

Open Shop Scheduling Problems Considering Equipment Preventive Maintenance 

ZHU Chuanjun, FENG Shijian, ZHANG Chaoyong, JIN Liangliang, WANG Linlin

2023, 34(14):  1693-1700.  DOI: 10.3969/j.issn.1004-132X.2023.14.006

Asbtract ( )   PDF (2203KB) ( )  

References | Related Articles | Metrics

Based on the actual operations of a production workshop, an open shop scheduling model with equipment preventive maintenance was established, and a hybrid genetic taboo search algorithm was designed to solve the problems. According to the characteristics of the problems, the genetic coding, decoding, improved crossover, and mutation operations of the hybrid algorithm, and the neighborhood structure of the taboo search algorithm were designed to achieve an balance between centralized searches and decentralized searches. The proposed algorithm was applied to the Taillard benchmark instance of open shop scheduling and the open shop scheduling case with preventive maintenance, which verifies the efficiency and effectiveness of the proposed hybrid algorithm.

Job-shop Scheduling Problems Considering Similar Learning Effect in One-worker and Multiple-machine Partterns

ZHANG Weicun, GU Hongyu

2023, 34(14):  1701-1709.  DOI: 10.3969/j.issn.1004-132X.2023.14.007

Asbtract ( )   PDF (4248KB) ( )  

References | Related Articles | Metrics

 A multi-objective job shop scheduling model was established considering the effects of job similarity and personnel learning under the one-worker multiple-machine production mode, and a grid filtering external archive genetic algorithm(GFEAGA) was designed to solve the scheduling problems. In order to improve solution efficiency, a two-stage coding and decoding approach was adopted, and an improved N6 neighbor structure search method was applied. The personnel selection method was designed to balance personnel workload. Non-dominant individuals were filtered based on grid sorting to enhance the diversity of the solution sets. The experiments verified the high efficiency and superiority of GFEAGA solution, and the sensitivity of the similarity and learning rate in the model were analyzed.

Human Factor Engineering for Human-Cyber-Physical System Collaboration in Intelligent Manufacturing

YANG Xiaonan, FANG Haonan, LI Jianguo, XUE Qing

2023, 34(14):  1710-1722,1740.  DOI: 10.3969/j.issn.1004-132X.2023.14.008

Asbtract ( )   PDF (5740KB) ( )  

References | Related Articles | Metrics

The theoretical system of intelligent manufacturing for HCPS confirmed the central position of human in the intelligent manufacturing system. Starting from the demand of human-machine collaboration in the intelligent manufacturing system, the emphases of human factors in HCIM were discussed from three levels such as behavior, intention, and cognition, based on the theory of gulf. Focusing on virtual-real fusion scenarios, multimodal human-machine interaction, cognitive quantification and other methods, the importance of human factor engineering in promoting the integration of human-computer intelligence was expounded. Finally, research direction and development suggestions of human-centered intelligent manufacturing from the implementation of HCPS intelligent manufacturing systems were put forward.

Robot Welding Trajectory Planning and High Frequency Control for Curved Seams

WU Chaoqun, ZHAO Song, LEI Ting

2023, 34(14):  1723-1728.  DOI: 10.3969/j.issn.1004-132X.2023.14.009

Asbtract ( )   PDF (6070KB) ( )  

References | Related Articles | Metrics

In a robotic real-time seam tracking system, the trajectory planning and control delay affected the tracking accuracy and welding quality. To solve this problem, a piecewise real-time trajectory planning and control method for curved seams was proposed by combining B-spline curve interpolation algorithm and EGM module. Firstly, the trajectory was segmented according to the principle of optimal interpolation time. Secondly, three times non-uniform B-spline was used to interpolate each trajectory to obtain the interpolation points. Finally, the high-frequency controller of the robot was designed. The interpolation points were sent to the robot by EGM module in a cycle of 4 ms to guide the robot movements. The experimental results show that this method may complete the planning of sine curve weld and guide the robot welding in 100 ms, and the tracking errors were controlled within ±0.2 mm, which realizes the rapid trajectory planning and high-frequency control.

A Robotic Multi-directional Polishing Trajectory Generation Method Based on Preston-PSO Algorithm

LI Jiaxuan, LI Lun, ZHOU Bo, ZHAO Jibin, ZHU Guang, WANG Zhengjia,

2023, 34(14):  1729-1740.  DOI: 10.3969/j.issn.1004-132X.2023.14.010

Asbtract ( )   PDF (13506KB) ( )  

References | Related Articles | Metrics

Aiming at the texture phenomenon in robotic polishing processes, a multi-directional trajectory generation method used for robotic polishing was proposed based on Preston equation and particle swarm optimization algorithm，and the generated multidirectional polishing path was smooth, evenly distributed and without corners. Simulations and experiments show that compared with the polishing effectiveness of parallel trajectory, the trajectory generated by this algorithm may effectively inhibit the generation of surface polishing texture, reduce the roughness of polished surfaces, improve the surface quality, and obtain good mirror polishing effectiveness.

Calibration Parameter Optimization and Accuracy Evaluation of Complex Visual Measurement Systems

SUN Jiale, LUO Chen, ZHOU Yijun, WANG Wei, ZHANG Gang

2023, 34(14):  1741-1748,1755.  DOI: 10.3969/j.issn.1004-132X.2023.14.011

Asbtract ( )   PDF (3739KB) ( )  

References | Related Articles | Metrics

There were many calibration errors in complex visual measurement systems， and the coupling between errors directly affected the accuracy of system measurement， so the accuracy improvelment of system calibration parameters was the key to ensure the system measurement accuracy. Thus， a calibration parameter optimization method was proposed based on multi-dimensional angular point error compensation. Firstly， a multi-dimensional angular point error function was defined， and the corresponding optimization model whose parameters were solved by LM algorithm was established. Then， the effects of the parameter optimization method on the system calibration optimization were evaluated by the optimization rate of the system calibration errors. Experimental results show that the optimization rate of the system calibration errors may reach 48%， the system measurement accuracy is high and meets the measurement requirements. 

Online Diagnostic Inspection and Prediction of Product Quality in Injection Molding Intelligent Factories Based on Data Mining

CHEN Yu, XIANG Wei, GONG Chuan

2023, 34(14):  1749-1755.  DOI: 10.3969/j.issn.1004-132X.2023.14.012

Asbtract ( )   PDF (1793KB) ( )  

References | Related Articles | Metrics

The dimensional accuracy of injection products was related to the injection processing parameters, and the real-time conditions of each stage in the injection processes and the changes of real-time working conditions. A workpiece quality diagnosis model was developed herein based on data mining. The real-time data collected by high-frequency sensors such as temperature, pressure, and displacement etc. in mold were used to construct the high-dimensional time series feature set. A three-stage feature selection method was used to determine the key feature subset, which was used to train the online quality detection model based on LightGBM classifier. The future values of each features were predicted based on the CNN-LSTM temporal prediction model, and the product quality was forecasted in advance with the classifier. The results show that the average recall rate of the macros is as 89.1%, and the average recall rate of the macros is as 81.6%.

Digital Transformation Mode and Strategy of SMEs in China

WANG Baicun, ZHU Kailing, XUE Yuan, BAI Jie, ZANG Jiyuan, XIE Haibo, YANG Huayong,

2023, 34(14):  1756-1763.  DOI: 10.3969/j.issn.1004-132X.2023.14.013

Asbtract ( )   PDF (5220KB) ( )  

References | Related Articles | Metrics

Promoting the digital transformation of SMEs was of great significance for Chinas manufacturing industries to improve quality and increase efficiency. SMEs were facing problems in digital transformation, such as high cost, fuzzy path, talent shortage, and lacking analytical framework and reference paradigm for digital transformation. The key factors to achieve digital transformation were clarified by building an analytical framework for SMEs digital transformation herein. Through case studies, 4 basic path models of digital transformation of SMEs were summarized and proposed. Based on the above researches, targeted suggestions were proposed for SMEs digital transformation in China, so as to promote the digital and intelligent development of SMEs.