Abstract
D.Ing. (Engineering Management)
When the complexity of the controller plant is a non-linear system, the production process cannot be fulfilled by the control effect and it is not easy to manipulate the parameters on Proportional-Integral-Derivative Controller. The major objective under study is simulating and modelling machinery then come up with a suitable intelligent Condition Based Maintenance system (i.e to develop a suitable intelligent system using fuzzy logic and artificial intelligence). The area under study of Fuzzy Logic came as a concern of nonlinear systems which fail to give correct information since machines use conventional Proportional-Integral-Derivative and hence this is for linear systems. Considering vibrations and continuous failure of machines, this will cause incorrect data portraying. Three companies were done in this thesis and all were analysed using fuzzy logic and some finite element analysis of components to monitor the parameters. Root cause analysis was also done to determine what was failing on different equipments. Three companies in a developing country were used as case studies for testing fuzzy logic approaches and models to machinery failure analysis. The three companies are designated as follows: Beverages Production Company, Hydro Power Generating Company and Water Distributing Company. The overall effectiveness of the companies were heavily affected due to many breakdowns. The main problem was found to be resulting from failure to monitor nonlinear systems that naturally exist in such environments. On the bottle washer at Beverages Production Company, fuzzy logic with Model Reference Adaptive Control to prevent failure on the bottle washer and introduce an intelligent Condition Based Maintenance program was done, a solution to come up with a control system that monitors water in the dam as well as protecting the equipment from failure and do stability control of machinery for Water Distributing Company as well as apply the concept of Maintenance Free Operating Period reliability metric to determine maintenance intervals as opposed to Mean Time Between Failure to the Hydro-Power Generation Company and also do a Simulink control on the governor. The results of the simulation of the control strategy of Fuzzy Logic Proportional-Integral-Derivative Controller have much preferred performance as compared to the general Proportional-Integral-Derivative Controller. This was done in the Simulink/Matlab simulation environment. The innovations behind this thesis at Beverages Production Company, the pneumatic valve of the bottle washer which controls the discharge of clean bottles was occasionally sticking or failing resulting in significant loss of production in the plant since all the other processes which follow depend on the bottle washer. The main causes of failure were caused by poor control of temperature and pressure. Excessive moisture and abrasive particles also caused failures. In order to solve this problem a Model Reference Adaptive Fuzzy Controller was designed for the pneumatic valve using the Matlab software. The error resulting from the difference between the actual system output and that of the reference model was executed by the Fuzzy Logic Controller. At company Hydro-Power Generation Company, the administering system of a 150 Megawatts Francis Turbine was developed using fuzzy logic. The representative framework parameters were mirrored with the real information accessible in the power plant. At Water Distributing Company the problem of corroding gates and failure was solved using solid works, Matlab, Programmable Logic Controller and Simulink. The sluice gate valve was controlled and done stability by Matlab to avoid failure. After fuzzy logic is applied in each and every company under study, the breakdowns will be less and machinery can be diagnosed and prognosis is easy to foresee the failures. A new gate was designed by the researcher. The research has shown that fuzzy logic being intelligent, machinery can last longer and production is high. The limitations are that although it is good in monitoring but in some machinery and companies may not work well because of environment like in the water gates. It is also expensive to implement and difficult to convince management to do the research for a start.