Application Of Information Fusion Technology On Fault Diagnosis Of Emulsifier

Emulsion explosive is the product that appeared in the 1970s. It becomes majorproduct of the industrial explosive in our country, owing to its characteristic which isstrong density, fast bursting speed, low cost and detonator sensitivity, etc. It is sowidely used in the field of traffic, electric power, oil exploration, mining and urbanconstruction that has done a great deal for the construction of national economy. Inthe production process of emulsion explosive, emulsifier, which is the key equipmentdeciding the performance quality of emulsion explosive in the production line, oftenattracts broad attention and research on its security problem at home and abroad. Withthe constant improvement of emulsion explosive’s production process, the security inits production process is being strengthened constantly too. But the emulsifier whichis used to product emulsion matrix contacts with raw materials of emulsion explosivefor a long time and continuous friction stir between each other. So the emulsifier’ssecurity problem is particularly conspicuous which concerns both security andreliability of the whole production line. Therefore, the research on the securityproblem of emulsifier has been a very essential and significant subject in the field ofemulsion explosive’s production all the time. With the development and innovation ofmathematics theory and analytical method in engineering application, new theoriesand techniques are constantly emerging in the application of the fault diagnosis forrotary machines. This offered new solution for condition monitor and fault diagnosisof machinery device. On the basis of elaborating the present status and developmenttrend and feasibility study on the application of information fusion technique on faultdiagnosis, this paper put forward emulsifier fault diagnosis system based oninformation fusion for the first time. In this paper some studies are made as below:(1) On the basis of research about the process of emulsified explosive productionand the importance and security of emulsifier during the production process,analyzing major faults of emulsifier starting from the emulsifier structure. The majorfault includes rotor rub, rotor imbalance, rotor misalignment, roller bearing fault andcatastrophic fault caused by cooling water or oil phase and water phase’s cutoff. Andthen this paper work deep in the reason about these major faults, fault mechanism andfault feature of these major faults. According to the result of analysis, the complexity and difficulty of emulsifier fault diagnosis are showed and total design solution of theemulsifier fault diagnosis system based on information fusion is proposed.(2) On the basis of the research about information fusion, this paper confirmsfeature layer-decision layer information fusion structure model according to structurecharacteristics and fault features, and determines analysis method of fault informationthat D-S evidence theory is combined and supplemented with improved BP neuralnetwork by contrasting many kinds of algorithms.(3) On the basis of the existing software and hardware of the emulsifiedexplosive production, the structure and the working principle of emulsifier faultdiagnosis system’s hardware components are confirmed according to both thestructure of the designed information fusion model and the production characteristicsof scene. Judging by the research of system signals’collection, conditioning,processing and different function of diagnosis, suitable hardware modules are choseby comparing with certain hardware design principles. And then the safety andpreventive measures of the on-the-spot hardware are proposed ensuring safe andeffective operation of the equipments.(4) Utilizing the hardware platform of emulsifier fault diagnosis system, thispaper proposes and realizes the software design of this system. Based on proposingstructural design of software, this paper realizes the function of communicating andcalling between different software by ODBC, ADO and ActiveX. Monitoring theproduction line through Kingview and transmitting the information data which arefrom Kingview to the diagnosis system consists of Visual Basic and Matlabcooperation together by database. This system processes the information data which isthe temperature and vibration characteristics of emulsifier by improved BP neuralnetwork trained by Matlab, and sends normalized data which are processed intouniform form as support degree of some kind of fault to algorithm of D-S evidencetheory fusing these correspondent support degrees of different faults in VB for thepurpose of acquiring more clearly recognizable support degrees. Thus this faultdiagnosis system can draw conclusions and report fault cause.The fault diagnosis system based on information fusion has a certain betterability of fault diagnosis though a test example diagnosing the structural model ofinformation fusion and a simulation test against the whole emulsifier fault diagnosissystem. This system is proven to have certain capacity on fault diagnosis, and canguarantee the safe operation of emulsifier. This system further offers a practical reference value to the development direction of safety monitoring and fault diagnosisfor emulsifier.