| The recovery characteristic of electromechanical products relates to the strategic issues such as the resources recycling and the sustainable development. The key to promote the recovery decision-making and improve the efficiency of recovery is to study how to establish an understandable and interoperable product recovery information model for the whole product lifecycle, build the semantic relationships between the product lifecycle information and the recovery activities, and accurately deliver them to the subjects of product recovery activities through an appropriate carrier. As a non-contact information identification technology, RFID (Radio Frequency Identification) has a strong information processing capability with small volume size and low costs. It can quickly and accurately obtain the information related to the recovery of end products. In this thesis, RFID was used as the carrier of electromechanical product lifecycle information; the recovery information model and semantic relationships of electromechanical product were constructed by ontology; a series of automated disassembly decision-making methods in the recovery activities were studied. The main work and research results of this paper includes:(1) The disassembly of electromechanical products is closely related to their assembly structures. In order to save the assembly structure in RFID tags, three assembly features (i.e. fir, cover and bearing) for express the assembly relations between parts/components were proposed, and an assembly/disassembly relation model was constructed based on these assembly features. The model used three matrices to express all the assembly relations in an assembly completely. A feasible disassembly sequence iteration method for electromechanical product was designed based on the assembly/disassembly relation model. On the basis of the Extensible Markup Language, a structured semantic encoding method for recovery information was proposed to realize the flexibility of RFID tag information storage and support the ontology mapping of heterogeneous information.(2) In order to achieve the automation of disassembly decision-making process and the co-building/sharing of disassembly knowledge, an ontology of electromechanical product recovery information was constructed, which expressed the following information in a formalized manner:the assembly relations, affiliations and disassembly priorities between parts/components, the disassembly depth and recovery strategies of electromechanical product, the inherent attributes, residual life and market price of part/component. The product lifecycle information provided by RFID could be transmitted into a semantic model (which is understandable by computers) by ontology mapping, and the ontology could eliminate the ambiguity of the information expression. The semantic model could provide the semantic foundation for the automated disassembly decision-making.(3) A Case-Based Reasoning (CBR) automated disassembly decision-making method based on ontology was proposed to quickly obtain the optimal disassembly plan for electromechanical product. The ontology based case expression and storage methods that comprise’semantic case description’and case solution’ were discussed; the automated case retrieval and matching methods based on semantic index and semantic similarity were studied; a’multi-granularity decomposition-recomposition’ method was used to realize the automated modification of case; a case base maintenance plan similar to human memory mechanism was designed.(4) A Rule-Based Reasoning (RBR) automated disassembly decision-making method based on ontology was studied to make up for the deficiency of the CBR method (i.e. CBR method will be invalid due to lack of proper cases). Two semantic assembly relations,’fix’and ’direct cover’, were proposed; SWRL/SQWRL, as well as SPARQL and Jena Ontology API, were used to construct the semantic rules for disassembly; an iteration method corresponding to the disassembly rules was designed for the automated generation of feasible disassembly sequences.(5) In order to determine the optimal disassembly depth of electromechanical product and solve the NP-hard problem that may occur during the calculation of the optimal disassembly sequence, a’calculation model for total recovery profit of electromechanical product’was constructed by considering the disassassembly costs and recovery profits corresponding to different recovery methods; genetic algorithm was used to study the’optimization problem in the searching of disassembly sequences’(6) Aiming at implementing the automated disassembly decision-making in view of incomplete product lifecycle information, research was conducted to use product historical recovery data to guide the disassembly decision-making. A bayesian prediction algorithm was proposed to realize the assessment for part/component damage condition; a fuzzy logic controller that estimates the product recovery value was designed; a calculation model for the recovery net profit based on the product historical data was constructed.With the combination of above researches, an RFID and ontology based automated disassembly decision-making system was developed in this thesis; the above theoretical methods were verified by applying them to typical electromechanical products, which proved the validity of the RFID based information transfer method and the automated disassembly decision-making method proposed in this thesis. The researches in this thesis have significant importance in improving disassembly decision-making efficiency and reducing disassembly decision-making costs.
|
PDF Full Text Request