Application of PSO and cultural algorithms for transient analysis of natural gas pipeline

Abstract

System simulation is an essential step for the optimization and control of natural gas transmission networks. Transient simulation of natural gas networks is more accurate than steady state simulation, but it imposes the complexity of solving nonlinear PDE flow equations.In this paper, a novel approach based on using two meta-heuristic algorithms: (a) particle swarm optimization (PSO), (b) cultural algorithm (CA) is presented to simplify transient simulation of gas networks with known inlet and outlet pressures.PSO or CA estimates different values for the network inlet flow rates. Using each of these estimated values, both boundary conditions will be known at the network inlets and discretized flow equations can be linearized. These linear flow equations will be solved for the network inlet nodes. This procedure will be continued for the next nodes until reaching network outlets. Then the differences of calculated and actual network outlet pressures will be defined as a cost function or error. Eventually, these algorithms will obtain the optimum inlet flow rates which minimize the cost function. Thus, the calculated pressures and flow rates at different gas network nodes which are obtained by using these optimum inlet flow rates will be the true values. Our proposed approach reduces the complexity of solving network transient flow equations with the mentioned boundary conditions while simulation results confirm its accuracy and efficiency.