Research On Disassembly Sequence Planning And Object Detection Of Electromechanical Products For Remanufacturing

The Remanufacturing is an important means of recycling waste resources.Disassembly is an important part of remanufacturing.Scientific dismantling methods can effectively improve the efficiency of remanufacturing.Disassembly sequence planning(DSP)is a method to find the optimal or approximate optimal disassembly sequence of a product.A good method of DSP can effectively improve disassembly efficiency and reduce disassembly costs.In this paper,aiming at the characteristics of DSP in remanufacturing and the current low level of disassembly automation,this paper conducts research on dismantling sequence planning and target detection for remanufacturing electromechanical products.The main research work is as follows:(1)A product disassembly information model is constructed based on the fastener constraint matrix and the part constraint matrix.Based on the model,a feasible disassembly condition is derived and a feasible solution is constructed.The optimization objective is to minimize the disassembly time and maximize the disassembly benefits.At the same time,a mathematical model for selective disassembly sequence planning of multi-objectives is constructed.The Pareto idea is incorporated to construct an improved multi-objective frog-leaping algorithm,and a multi-objective selective disassembly sequence planning method based on the multi-objective frog-leap algorithm is proposed(2)Aiming at the problem of asynchronous parallel disassembly sequence planning,taking disassembly personnel as the research object and minimizing the completion time of disassembly as the optimization goal,a mathematical model of selective asynchronous parallel disassembly sequence planning for multiple people and multiple target pieces is constructed.The coding structure,decoding method,local evolution formula and global information interaction method of frog individuals are improved,and a multi-person and multi-objective selective selective parallel disassembly sequence planning method based on an improved hybrid frog jumping algorithm is proposed.(3)Aiming at the characteristics of remanufactured parts,the feature extraction network,anchor box parameters,and loss function parameters in YOLOv3 were improved,and a target detection method based on deep learning neural network YOLOv3 was proposed(4)Based on VS2015,a prototype system of disassembly sequence planning and target detection was developed.Taking the disassembly example in this article as an experimental object,the automatic disassembly sequence planning and target detection functions under various algorithms and modes were implemented.