Solving the Pipe Network Analysis Problem Using Optimization Techniques

Abstract

For more than forty years, approximate solutions for the classical pipe network analysis problem have been obtained by direct solution of the nonlinear stationary point conditions. We propose a revolutionary new approach involving optimization techniques for solving this well-known engineering problem. It is shown that the pipe network analysis problem may be described mathematically in terms of a nonlinear convex cost network flow problem. Three mathematical programming algorithms for solving this problem have been coded and are computationally compared with a code using the traditional Newton-Raphson technique. The computational experience demonstrates that this new approach provides an attractive alternative for solving this important problem.