Research On The Diagnosis Decision-making Of Centrifugal Blower Based On Ontology And Fault Tree

Reasonable equipment knowledge management can not only make the enterprise’s equipment management more orderliness, but also reduce the cost of equipment management greatly, thus, it is very important for ensuring the successful operation of an enterprise. On the other hand, too long downtime maintenance of the key equipments means enormous economic losses of the continuous operation enterprise. Thus, how to carry on an effective management of fault knowledge and how to find out the malfunction as soon as possible are the important research topics in the field of fault diagnosis.This thesis chooses the key equipments of some petrochemical enterprise as the research object. The ontology based fault knowledge management and the fault tree construction problems are investigated deeply. At present, the fault information of enterprise’s equipments is mostly heterogeneous. There is no semantic relation in fault knowledge. It is too hard to share and reuse the fault knowledge. This makes the management of the fault knowledge difficult. The ontology has obvious advantages in knowledge expression, knowledge sharing and knowledge reusing. Thus, this thesis uses ontology to rebuild the fault knowledge of enterprise’s centrifugal blower. On the other hand, the fault tree is generally hierarchical constructed. This is in consistent with the construction of ontology. Thus, a fault tree construction method based on ontology is proposed. It builds the equipment malfunction fault tree through the fault information ontology of the centrifugal blower and the ontology rules. Then the shortest path of the fault diagnosis can be found through the qualitative and quantitative analysis of the fault tree.The main works and innovative achievements in this thesis are as follows(1)The working principle and basic structure of centrifugal blowers is introduced. On this basis, the common failures and failure mechanisms of centrifugal blowers are analyzed. It provides the basis for the research of the fault knowledge management of centrifugal blowers.(2) The comparation of the commonly used methods of ontology construction is given. As the "skeleton method" has good scalability and standard semantic model, and it is mainly used in the development of enterprise ontology, this thesis chooses the "skeleton method" to build the equipment failure ontology. However, the traditional "skeleton method" has shortcomings in dynamics, knowledge sharing and knowledge reusing. Thus, this thesis proposes an improved "skeleton method" based on the traditional "skeleton method". The building process of the equipment failure knowledge is researched specifically based on the improved "skeleton method"(3)A fault diagnosis decision method based on ontology and fault tree is proposed. This method constructs equipment fault tree based on ontology and ontology rules. The minimal cut-sets, the probability importance degrees of the bottom events, and the key importance degrees of the system are calculated. At last, the shortest fault diagnosis path of the fault diagnosis can be obtained according the descending order of the ratio of efficiency to time. This method combines the advantages of ontology with fault tree. It realizes rapid fault locating based on knowledge sharing and knowledge reusing.(4) By choosing a centrifugal blower of some petrochemical enterprise as the research object, a diagnosis decision support and application system based on ontology and fault tree is designed and developed. Then the practicality and effectiveness of the system is verified.In this thesis, the common failures and failure mechanisms of centrifugal blowers are investigated. Then the fault knowledge ontology is built by using of the improved "skeleton method". A diagnosis decision method based on ontology and fault tree is proposed. A diagnosis decision support and application system of centrifugal blowers based on ontology and fault tree is developed, and is applied in practice successfully. The research results in thesis have certain theoretical significance and engineering application value.