Optimization Of Tree Pipe Networks Layout And Size Using Particle Swarm Optimization

Optimal design of the irrigation pipe network (IPN) is defined as selection of combination of discrete pipe size at minimum possible cost without violating the specified constraints. An IPN is a complex hydraulic system consisting of valves, reservoirs, pipes, sprinklers, hydrants and pumps. These elements are designed to convey water to the irrigable farms at sufficient pressure head and demand.In the design of IPN, the determination of the layout of the links is an essential step while determining the optimum size of components is then another step. However, it is very important to design a network that is not only cost effective but also satisfies the system constraints such as adequate pressure and velocity. In recent years, many researches have been done using computer software programmes as well as optimization techniques to investigate optimal solution to piped networks. A number of these techniques are heuristic optimization techniques, dynamic programming, linear and non-linear programming.This research provides an explanation about the computer code development that presents particle swarm optimization (PSO) to the least-cost design problem of irrigation pipe network (IPN). Given that, a manual calculation in designing for IPN layout and size is based on trial and error, that makes it difficult to minimize capital investment and energy cost especially if the network is large.To design a system with least-cost of pipes is the main aim of many irrigation and hydraulic engineers, which is normally considered as the optimal solution. However, for a number of reasons the optimal solution may not be feasible. In this research, the objective was to optimize simultaneously size and layout of the irrigation pipe networks using PSO technique, linked to the MATLAB software, to reduce the pipeline investment cost in irrigation projects. The Pipe layout and size optimization model for a tree irrigation pipe networks are presented. The PSO technique performance was tested and results are compared against non-optimized (Step-by-step) and genetic algorithm optimization methods. The non-optimized (Step-by-step) optimization method yielded the least cost CNY33,256. Nevertheless, the least cost yielded by non-optimized (St^p-by-step) did not necessarily represent the optimal solution because constraint like maximum pressure variation which is very important was not met. The PSO and Genetic Algorithm (GA) offered the global optimal solution of CNY15,712and21,119; this solution is less than the total investment amount of the arrangement by25.63%. The proposed PSO technique with an increase in the search space showed a quick response in the size of the swarm and the initial swarm compared to the non-optimized (Step-by-step) design method and GA.The results of this research proved that the use of PSO algorithm technique could be effectively used for solving pipe irrigation design problems. PSO is a quick search algorithm if compared to non-optimized (Step-by-step) and GA design method. However, when exploring complex functions PSO faced a premature convergence, therefore, further researches for using PSO technique are required within the field of hydraulic networks.