Research On Elevator Fault Diagnosis Based On Rough Sets

It is a hot issue that how can we detect quickly and effectively to isolate and repair the located elevator malfunction.Fault diagnosis expert system has been successfully applied to the elevator fault diagnosis.The elevator fault diagnosis knowledge has some features such as uncerainty, inaccurate and redundancy.How to get the accurate and simplified diagnosis knowledge is a key problem which needed to be solved to the elevator fault diagnosis system. Rough Set theory is employed to knowledge acquisition of the elevator fault diagnosis expert system. A new approach for building fault diagnosis knowledge tree was proposed.The main works are as follows:This paper is based on the frequency of the distinguishing the matrix attribute reduction method and establishing the fault tree which is based on the simplest rule and attribution importance. The paper applies the rough set theory to the knowledge acquisition of the elevator fault diagnosis expert system, through classifying and simplifying the elevator fault diagnosis sample information obtains the simplest diagnosis rules and establish the tree of fault diagnosis.The fault diagnosis expert system uses the tree of fault diagnosis as its knowledge,finds the fault reason by gradual derivation of the symptom. We applied this method to establish the unit fault tree which is based on the method of unit small fault tree establishment,eventually establish the whole fault diagnosis tree of the elevator.This paper takes the safey circuit part of the elevator control system as an example,simplifies the fault decision-making information,gets the simplest diagnosis rules and generates the fault tree of that part.We can conclude from the safety circuit schematic that the fault tree can effectively derive the fault and it is more simplified than that by human work.It increases the diagnosis rate and reduces the complexity of the fault diagnosis expert system knowledge base.Rough set was applied to knowledge acquisition in the paper. It analyzes and processes the fault information, reduce the fault diagnosis rule and establishes the fault tree,then collect the fault tree of separate device parts by the given system structure to establish the entire fault tree.The generated fault tree not only remains the advantages of human-establishing which can fully understands the system structure,but also avoids the logic and pretermission problems that can be made easily during human establishing complex system fault tree.It reduced the time for estabishing trees and improved the rate of fault diagnosis information processing.