Key Technologies Research On Knowledge-based Structural Optimization Design Of Excavator Boom

In allusion to the deficiencies existing in current structural optimization algorithm ofhydraulic excavator arm such as the aptly occurring malformed structure in optimizingresult, the inconsequent selection of structure driving variables and the regions to map theboom structure stresses, the low efficiency in optimization and the easily falling into localoptimal solution and so on, the thesis considers the hydraulic excavator boom as theresearch object, mainly studies the following aspects:As the existing numerical optimization of boom structural hard to completely expressand deal with domain knowledge, experience knowledge and normative knowledge, whichis likely to cause malformed boom structure, a method integrating the numericaloptimization algorithm and the logical-inference technology called knowledge-basedstructure optimization design of boom is presented. The optimization design model undermultiple loading conditions with the object of minimizing the volume of the boom isestablished based on the numerical calculation completed by Genetic Algorithm (GA)while the logical-inference technology realized by KBS (Knowledge-Based System). Theintegrated method between GA and KBS and the solving algorithm of the optimizationdesign model are discussed. Then the method to represent and handle knowledge for boomstructure design based on layered coding and layered classifying is propounded. Theknowledge of boom structure design is obtained, represented and handled by establishingthe boom structure classifying and coding system and various types of boom structuresimilarity eigenmatrix. Thus, the malformed boom structure occurred in theknowledge-based structure optimization design of boom is averted with the technology ofadjusting the malformed boom to become an individual of the class that has the shortest“shape-distance” from the malformed boom guiding by the experience knowledge.In view of the irrational choice of structural driving variables which is apt to result inlocal minimization appearing in existing excavator arm structural optimization, the boomstructure optimization driving variables selecting method based on Spearman rankcorrelation analysis of boom structural parameter with volume response and stressresponse completed by Monte Carlo Simulation is proposed. And, the divided-phase LatinHypercube Sampling method that could effectively ensure a normal and distribution-equal boom structure sample as well as the method of automatic modeling and finite-elementanalysis for boom structure is offered. In the light of the low convergence rate occurs inexisting boom structure optimization, the experience knowledge of boom structure designis used for guiding the optimization computing process by applying stress strengthequalization.Furthermore, the frame for the knowledge-based structure optimization of boomconsidering its functional requirement is presented. Based on the document organizationwith project file and project folder and the technology of integrated tools such as CLIPS,VC++6.0, Pro/Toolkit and APDL, the knowledge-based structure optimization software ofboom is developed. And the feasibility and the effectiveness of the knowledge-basedstructure optimization software of boom in finding the global optimum solution, ensuringthe quality of optimization design is corroborated by a case.Finally, the result of research demonstrates that the knowledge-based structureoptimization software of boom can effectively avert the malformed boom structureoccurred. And the method of layered coding and layered classifying the boom structure canbe very effective in obtaining, representing and handling the experience knowledge and thenormative knowledge of boom structure design. The divided-phase Latin HypercubeSampling method provides high quality boom samples with normal and distribution-equalstructure for the Monte Carlo Simulation and the neural network modeling. The CAD/CAEmodule supports the automatic modeling and finite-element analysis for boom structureand the results visualization and output well.