The Method Of Rapid Finite Element Analysis Based On KBE

Finite Element Analysis (FEA) is widely used in various engineering domains such as machinery, electronics, construction and water conservancy etc. However, in the process of enterprise implementation, such issues like lacking of the design/analysis knowledge, wasting effort in design process and inadequate knowledge reuse still be exist. This thesis researches on the method of Rapid Finite Element Analysis (RFEA) based on KBE and performs several key techniques on engineering semantics markup, intelligent analysis solution generation and design/analysis information integration etc. The contents are listed below:The theoretical framework of RFEA methods is established based on KBE. The design/analysis integrated production developing process is discussed thoroughly and production data type is distinguished in every phase. A solution which is suitable for the multiple transmit of production data, information and knowledge is proposed in the design/analysis process and an intelligent processing technological system to support this is built.An ontology based method which intends to semantically annotate the CAD models is proposed and the engineering semantic and its representing method is discussed in detail. The design/analysis ontology is built and the same terminology and ontology for annotation is provided. The issues of annotation method, implementation technology and annotation information saving etc are researched. This method allows the academic team to semantically annotate the CAD models, give the CAD models engineering ontology and implement the expansion from production model design phase to analysis phase in various angles.According to the feature of knowledge diversity in FEA solution producing analysis phase, the multi-modal knowledge support (MMKS) FEA solution producing strategy is proposed. The concept of ontology object diagram is introduced. The case matching is turn into the diagram matching. A mapped node matching algorithm is built and the object diagram is simplified into the case diagram. A CBR algorithm is introduced and the ontology similarity numeration method is proposed. Based on those, ontology based RBR case amendment algorithm is built by using the RBR to guild the case amendment. Meanwhile, considering adequately on the intelligent activity of humankind, an ontology based retrieving mechanism is set up to help the designer to locate the necessary knowledge source for completing the FEA solution. This kind of human-machine combined method can be more flexible and suitable for the multiple modal knowledge support method.The design/analysis information transmits method based on multi-domain feature mapping (MDFM) theory is proposed. A framework for this method is built and the corresponding implementing mechanism is analyzed. By analyzing the XML based information mapping between annotated files and reasoning generated analysis solution, the FEA template in the previous phase is instantiated into the concrete FEA solution and turns into executable command flow after. After the execution, the forward information from design domain to analysis domain can be completed. Through the description of key features and rules based reasoning, the intelligent post processing mechanism can be built and the feedback transmission from analysis result information to design domain can be completed. Advice for design modification is proposed.With methods proposed above as unit techniques, the prototype is built and the function modules of the system are planned. The effectiveness of the RFEA is proved by the case of static analysis of wind power gearbox. The result shows that the theory in this paper can improve the efficiency and productivity of traditional FEA.Finally, the research is summarized and the innovations are pointed out. Suggestions for further study are raised.