Optimization of support concepts for piping systems using evolutionary algorithms

For the design or backfitting of piping systems a support concept has to be found so that for all loading conditions the stresses and loads on the components, fittings and connections are as low as possible. The problems result from the fact that for different load cases different support concepts are optimal. The load from thermal strains need soft supports in order to avoid constraints, while for seismic loads the optimum support concept needs rigid fittings. Thus, the resulting concept is always a compromise for the load transfer of expected load cases. Based on the example of a bypass cooling loop in a nuclear power plant the authors show that genetic algorithms are adequate tools to solve the optimization problem.