公开(公告)号：US20240280982A1

公开(公告)日：2024-08-22

申请号：US18187698

申请日：2023-03-22

Applicant: GUANGDONG UNIVERSITY OF TECHNOLOGY

Inventor： Jiewu LENG ,  Xiaofeng ZHU ,  Caiyu XU ,  Hongye SU ,  Qiang LIU

IPC: G05B23/02

CPC classification number: G05B23/0283 ,  G05B23/0235 ,  G05B2223/06

Abstract: A system and method for monitoring the failure of assembly tooling for a mass-individualized production line are provided. The system realizes real-time monitoring and prediction of the remaining service life of the assembly tooling in the mass-individualized production line through the cooperation of a manufacturing execution system (MES), a supervisory control and data acquisition (SCADA) system, an assembly tooling failure prediction system, a controller network, and an assembly line. When the remaining service life reaches a certain threshold, the system sends an early warning to an operator, and provides decision support for the operator to replace the assembly tooling. The assembly tooling failure prediction system is built with an assembly tooling failure prediction model for predicting the remaining service life of the assembly tooling in real time, which avoids the influence of human factors and greatly improves the prediction accuracy.

公开(公告)号：US20250036111A1

公开(公告)日：2025-01-30

申请号：US18916505

申请日：2024-10-15

Applicant: GUANGDONG UNIVERSITY OF TECHNOLOGY

Inventor： Jiewu LENG ,  Rongjie LI ,  Caiyu XU ,  Yuanwei ZHONG ,  Junxing XIE ,  Keyou ZHENG ,  Qiang LIU

IPC: G05B19/418 ,  H04L9/00

Abstract: In a method for predicting product assembly quality based on longitudinal unified learning, sample alignment is performed on a data sample of each participant to resolve problems of decentralization, nonuniformity, and scarcity of data; data partitioning is performed by a multi-player parallel structure on product assembly data by using a customized data partitioning policy, layer normalization is firstly performed by an encoder on partitioned data, and feature extraction is performed by a multi-thread attention layer to mine a correlation between each assembly production line inside the participant and assembly data of each device, so that a model feature extraction capability is enhanced; gradient security aggregation is performed on the local model of each participant by using a homomorphic encipherment method of secure multi-player computation, to obtain a global model, so that data of a plurality of sub-factories is merged to co-train a high-precision assembly quality prediction model.

公开(公告)号：US20240295873A1

公开(公告)日：2024-09-05

申请号：US18346845

申请日：2023-07-04

Applicant: GUANGDONG UNIVERSITY OF TECHNOLOGY

Inventor： Jiewu LENG ,  Jiwei GUO ,  Xiaofeng ZHU ,  Caiyu XU ,  Hongye SU ,  Qiang LIU

IPC: G05B19/418

CPC classification number: G05B19/41865

Abstract: A method for constructing a topology reference structure for a production line is provided. The present disclosure is based on historical production line topology data of an enterprise to extract a commonly used topology reference structure for a production line of the enterprise by a computer through a machine learning (ML) algorithm, so as to form a typical production line topology group of the enterprise. The present disclosure can record typical production line characteristics and production habits of the enterprise, realize reuse of a production line structure and production line construction knowledge, reduce the workload of production line designers, and improve the production line construction efficiency of the enterprise. In addition, the present disclosure avoids the interference of designers' subjective decisions to a certain extent, and the reference structure extracted by the computer has high reference value, and is objective, mature, and stable.