On-condition Maintenance Decision Research Of Aero-engine Based On Cost Optimization

Facing the serious marketing competition, it is vitally important for the airlines to achieve any possible profit and minimize the service cost with the prerequisite of safety. According to the statistics data, the maintenance cost of the aero-engine accounts for more than 30 percent of the total operating cost of airlines, which reveals much controllable margin. Based on this, the practical investigation is conducted to the maintenance cost decision-making system of aero-engine in this thesis. Further the optimal models are built. After the substantive study and contrast analysis of the existing aero-engine maintenance condition, an optimal decision-making system is built and verified by the industrial case. This thesis includes the following 4 parts. (1) Building an on-condition maintenance optimal model for a gradually deteriorating single-unit system. The proposed decision model enables optimal inspection and replacement to balance the cost engaged by failure and unavailability on an infinite horizon. Two maintenance decision variables are considered: the preventive threshold and the inspection schedule based on the system state. Using the regenerative and semi-regenerative process theory a mathematical mode is developed to assess the performance of the proposed maintenance structure. To solve the optimal decision variables, an innovative convergence algorithm for the renewal function is proposed based on the characteristic of Weibull distribution function. The algorithm is proved to be convergent theoretically and coded for numerical calculation. (2) Determining the aero-engine shop visit maintenance level The key objective of this section is to propose a method to determine the proper shop visit maintenance level before the aero-engine is disassembled for inspection. First of all, the current determining method and work flow of aero-engine maintenance are simply introduced. Then classical and variable precision rough set theory are adopted to conduct Knowledge discovery in this thesis. The uncertainty of acquired knowledge is decreased by the employment of the simplified knowledge entropy algorithm based on the variant precision rough set theory. An increment learning algorithm is also proposed for the knowledge update case, which is beneficial to maintain the rule database. (3) Forecasting the aero-engine shop visit cost The brief introduction of general procedure, condition and method about cost forecasting is proposed first. Then, based on the plenty analysis of the fundamental aero-engine maintenance task, the breakdown structure of the aero-engine maintenance cost is put forward. Integrated with the characteristic of aero-engine maintenance policy and Grey Model Theory a new type of condition-based cost forecasting model is built. Also the attribute significant method, which is induced from rough set theory, is employed to calculate the weigh coefficient for the improvement the prediction precision and simulating stability. Finally, a numerical example is illustrated to validate the built model. (4)The design of fleet maintenance decision support system The decision-supporting system of fleet maintenance is designed based on the practical requirement of the airlines and the analysis of the operating logic of aero-engine team maintenance. With the theory of decision-making system, an aero-Engine Fleet Maintenance Decision Support System (EFM-DSS) with 3-tie network structure is built up on the basis of Oracle database management system.