Research On Product Disassembly Sequence Planning Under Parts Missing Condition Driven By Visual Data

The disassembly of end-of-life products is a key part in remanufacturing.It is of great significance to alleviate the dual pressures of resources and economy,promote the healthy development of society,and realize sustainable manufacturing.Usually,the disassemble sequence generation depends on the original CAD of the product.It can only complete the disassembly of the product according to the disassemble sequence generated by the original CAD data.When the product is used for a long time,some parts are missing.The original disassembly sequence should be adjusted.If the operator cannot obtain the new disassembly constrains data,the original disassemble sequence cannot be applied to the current situation.The disassembly task will not be completed.It will greatly reduce the working efficiency.Therefore,it is of great significance to study the products disassembly sequence planning under parts missing condition.To resolve these problems,this article focuses on the problem of disassembly constraint modeling driven by visual data and disassembly sequence planning under parts missing condition.It can provide corresponding disassembly sequence information for the automatic disassembly.The main research work is as follows:(1)Research on the modeling of products disassembly constraints relationship based on visual data.Industrial data sets of mechanical and electrical product are constructed for the detection and identification of products parts.A variety of preprocessing methods,such as random rotation,shearing,random brightness and random contrast are used to enrich the data sets.This article selects appropriate algorithm to identify product parts from the two-dimensional images after analyzing the existing target detection algorithms based on two-dimensional images.It obtains the parts information of the disassembly object,and establishes the corresponding disassembly constraint model.The article analyzes disassembly feasibility based on the constructed disassembly constraint model.It generates a feasible disassembly sequence,and provides corresponding sequences information for the next step of disassembly sequence planning.(2)Research on the disassembly sequence planning under parts missing condition.To solve the problem of missing parts in the disassembly process,adjustment of the disassembly constraint relationship and disassembly constraint model is carried out based on visual data.The disassembly feasibility of the disassembly model analysis is performed to generate a feasible disassembly sequence.This paper constructs the objective function and the evaluation criteria of disassembly sequence planning problem.Besides,this article proposes an improved gray wolf algorithm to optimize the feasible disassembly sequence.Finally,taking the diaphragm coupling as an example,tests are carried out to verify the effectiveness of the method under complete product and the product of missing parts.(3)Design and develop a product disassembly sequence planning system driven by visual data.This system is based on hardware devices such as Kinect vision sensors.This article designs and develops the system for disassembly of products.It verifies the feasibility of the proposed framework and method,and provides data and management and maintenance support for disassembly tasks.The system includes a basic data management module,an image acquisition module,a component recognition processing module and a disassembly sequence planning module.The user can call the corresponding module of this system,extract the information of the product to be disassembled,and generate the planning disassembly sequence of the product to complete the disassembly process.