A practice in optimal cost design of water distribution networks

This study proposed more practical and realistic optimal pipe design approaches, using metaheuristic optimization algorithms. The first method considered the continuity of the pipe size in a pipeline to establish an effective design and construction, using Revised Harmony Search (ReHS). The proposed method is applied to Hanoi network, which is often used as a benchmark network in optimal design problems. Although slightly more expensive compared to previous studies the proposed method brings about more practical, realistic, and still close to optimal results. The second approach applied the optimal cost design of parallel main pipes. In general, diameter of parallel main pipes is decided with very ambiguous guideline that suggests a diameter of about 65% of that of the pipes used in practice. However, the instruction is not clear, so decision of the optimal diameter is necessary. Therefore, this study focused on how to decide the optimal diameter, using Multi-objective Genetic Algorithms (MOGA) to strengthen the CW Industrial water supply system by installing dual and additional main pipes. The objective functions are to minimize the amount of the expected water shortage and pipe construction cost. As a result, optimal design can reduce the costs by 3.0% and the expected water shortage by about 3.8%, compared to the costs when applying the existing guideline. These two approaches show that meta-heuristic techniques can be applied to practical optimal design of water distribution systems and achieve successful results.