Research On Intelligent Methods For Aeroengine Pipe-Routing Design

The pipe-routing design is one of important steps to design the aeroengine, which directly impacts the performance, reliability, lifecycle as well as cost of the aeroengine. In practice, the design for aeroengine pipe-routing is mainly finished with the aid of computers by engineers with rich experience. Because there are a great number of pipes, numerous constraints and highly complex layout space involved in the design of the aeroengine pipe-routing, the procedures of this method are very complicated and time-consuming, which not only makes the aeroengine pipe-routing design very difficult to be optimized, but also makes the cycle, quality and efficiency of the design difficult to meet the requirements of the development of equipment manufacturing. Therefore, the automatic and optimal design for the pipe-routing has become a hot issue to be urgently solved.With the rapid development of life science, many intelligent algorithms to be used to solve complex problems emerged. These algorithms-based intelligent design methods for the pipe-routing can realize the automatic and optimal design for the pipe-routing which has important significance to improve the quality and efficiency of the pipe-routing design, alleviate the technician's burden, shorten the design cycle and reduce the manufacturing cost. So far, although lots of scholars and institutions at home and abroad did a good deal of significant research about intelligent design methods for the pipe-routing, there is currently not a set of mature theory and methods yet. In this context, based on a comprehensive analysis of the domestic and overseas research about intelligent design methods for the pipe-routing, the dissertation considering aeroengine pipes as the research object thoroughly studied intelligent design methods for the single pipe-routing, branch pipe-routing and multi-pipe-routing, and developed a prototype system. The main contents are described as follows:(1) The dissertation provides a comprehensive survey for the domestic and overseas research about intelligent design methods for the pipe-routing and analyzes existing problems. On this basis, the dissertation presents its research direction and ideas.(2) The mathematic model of the pipe-routing is built regarding the coordinates of endpoints as the variable and the shortest pipe-routing as the optimized object. Moreover, considering the issue that the pipe-routing layout space is highly complex and its digital model occupies a large number of computer resources, the filling-surrounding spatial information extraction method is proposed based on the regionalization, gridding and scanning thoughts. It solves the primary key issue of the intelligent pipe-routing. The case study illustrates the feasibility and effectiveness.(3) Based on spatial creation, spatial discretization and heuristic thoughts and taking engineering rules into consideration, the engineering rules and spatial discretization-based aeroengine pipe-routing algorithm is presented. In the algorithm, the creation of the local searching space narrows the searching scope and improves the searching efficiency; the discretization of the local searching space reduces the complexity of the pipe-routing search and avoids the disordered pipe-routing; the proposed trajectory equations and searching strategies meeting the requirements of the engineering improves the quality of the pipe-routing. The case study proves that the algorithm is feasible and effective to be applied for the pipe-routing of the single pipe.(4) On the basis of in-depth research on artificial fish swarm algorithm and chaos optimization algorithm, the improved chaos artificial fish swarm algorithm is put forward. For guaranteeing the convergence of the algorithm and effectively improving the population diversity, global searching ability and efficiency of the algorithm, on the one hand, the randomicity, ergodicity, sensitiveness characteristics of chaos motion is used to improve the global search and local search of the artificial fish swarm algorithm; on the other hand, the visual scope and moving step are set with gradient change. With the comparison between the performance of the artificial fish swarm algorithm and the improved chaos artificial fish swarm algorithm, the latter can more efficiently search the global optimal solution. It has been proven by the case study that the algorithm results in the satisfied pipe-routing for the single pipe.(5) Based on thorough research about artificial bee colony algorithm, the improved artificial bee colony algorithm applying to the pipe-routing is brought forward. First, the algorithm solves the problem that path node is not fixed by setting the maximum number of path nodes. Secondly, it adopts the grid division and engineering rules-guided search strategies to deal with the neighborhood construction problem. Thirdly, the tournament selection strategy is used to select employed bees for onlookers for overcoming the premature convergence and stagnation problem. It has been proven by the case study that the algorithm can find the global optimum and is effective for solving the pipe-routing for the single pipe.(6) According to the characteristics of branch pipe itself, the engineering rules-based branch pipe-routing algorithm is presented based on heuristic engineering rules searching thoughts. In the algorithm, firstly the layout space is transformed to two-dimensional layout space; secondly, the branch pipe is decomposed into multiple seperate pipes; thirdly, the pipe-routing of every separate pipe is searched in terms of the proposed pipe-routing searching strategies meeting the requirements of the engineering; finally, the optimal pipe-routing is determined by calculating the lengths of the overlapped pipe-routings and comparing them. It has been concluded that not considering the position of the branchpoint and the sequence of laid pipes, branch pipe-routing meeting the requirements of the engineering can be done through the algorithm.(7) Based on in-depth research about maze algorithm, the branch pipe-routing layout-oriented collaborative maze algorithm is proposed. In the algorithm, by introducing the thought of collaborative coevolution, the branch pipe is decomposed into multiple separate pipe, at the same time every separate pipe is regarded as a population and all populations constitute an ecological system; then the current optimal individual of every population is rapidly found by maze algorithm; at last, the optimal pipe-routing for the whole ecological system is determined by the calculation of the lengths of the overlapped pipe-routings and comparison among them. It turned out that not considering the position of the branchpoint and the sequence of laid pipes, the branch pipe-routing can be fast optimized by using the algorithm.(8) Based on the comprehensive analysis of the principle, characteristic and type of the co-evolutionary algorithm, the co-evolutionary algorithm and the improved chaos artificial fish swarm algorithm are combined. Further the improved chaos artificial fish swarm algorithm-based co-evolutionary algorithm appied to the multi-pipe-routing is put forward. In the algorithm, every pipe is considered as a population and all populations constitute an ecological system; further, on the one hand, every population is evolved independently by the improved chaos artificial fish swarm algorithm. On the other hand, choosing the representation to create the system model, the current optimal pipe-routing is achieved with populations mutually collaborating in the system model; finally, the optimal pipe-routing for the whole ecological system is gained by comparing the shortest total path length from every evolution. The algorithem not only avoids the combination explosion when the number of the pipes is increased, but also optimizes the pipe-routing on the whole not considering the layout sequence. The case study demonstrates that compared with the the improved chaos artificial fish swarm algorithm used for sequential layout, the proposed algorithm used for parallel layout could achieve more satisfactory pipe-routing in the design of the multi-pipe-routing.(9) In accordance with above methods and contents, the framework of aeroengine intelligent pipe-routing prototype system is designed and the prototype system is developed by using UG and MATLAB software. The prototype system includes the functions:the information extraction of the pipe-routing layout space, the identification of the initial pipe-routing information, the pipe-routing planning, the creation of pipe-routing model as well as the output of the pipe-routing engineering information, which makes the operation of the pipe-routing easy and the pipe-routing solution visualized.