Modeling of Hydraulic Pipeline Transients Accompanied With Cavitation and Gas Bubbles Using Parallel Genetic Algorithms

Abstract

Mathematical models of pressure transients accompanied with cavitation and gas bubbles are studied in this paper to describe the flow behavior in a hydraulic pipeline. The reasonable prediction for pressure transients in a low pressure hydraulic pipeline largely depends on several unknown parameters involved in the mathematical models, including the initial gas bubble volumes in hydraulic oils, gas releasing and resolving time constants. In order to identify the parameters in the mathematical models and to shorten the computation time of the identification, a new method—parallel genetic algorithm (PGA)—is applied in this paper. Based on the least-square errors between the experimental data and simulation results, the fitness function of parallel genetic algorithms is programed and implemented. The global optimal parameters for hydraulic pipeline pressure transient models are obtained. The computation time of parallel genetic algorithms is much shorter than that of serial genetic algorithms. By using PGAs, the executing time is 20h. However, it takes about 204h by using GAs. Simulation results with identified parameters obtained by parallel genetic algorithms agree well with the experimental data. The comparison between simulation results and the experimental data indicates that parallel genetic algorithms are feasible and efficient to estimate the unknown parameters in hydraulic pipeline transient models accompanied with cavitation and gas bubbles.