Application Of Swarm Intelligence In Identification On Method Of Gas-Liquid Two-Phase Flow Regime

Gas-liquid Two-phase flow regimes are becoming more and more important in scientific field, human life and environmental protection. There is an important significance in the research of identification on gas-liquid two-phase flow regime. It will improve the recognition rate if selecting features from the whole feature sets reasonably. Therefore, gas-liquid two-phase flow pattern identification of the research not only has academic significance, but also provides a strong technical support for related industrial production equipment, security, economic design and operation.Gas - liquid two-phase flow of the comprehensive experimental stage was tested. Because the differential pressure fluctuation signals contain a lot of information, we achieve the level of tubes collected air - water two-phase flow of the differential pressure fluctuation signals that were analyzed. In order to overcome many issues irrelevant to extracted features of gas-liquid two-phase flow characteristics, we use the de-noising filter firstly and then empirical mode decomposition (EMD), at the end wavelet packet techniques. After the above methods have become a fusion feature set. We use swarm intelligence algorithm to select the useless and redundant features. The typical swarm intelligence algorithms are Genetic Algorithms (GA), Ant Colony Optimization (ACO) and discrete binary version of particle swarm optimization (BPSO). BPSO was used to search feature subsets in combined features with in-inter classes variance and accuracy to be fitness function of the BPSO. Furthermore, the least square support vector machine (LSSVM) was used as an evaluator to estimate the performance of the selected feature subsets.Horizontal tube gas-liquid two-phase flow of the four kinds of typical flow pattern identification results show that swarm intelligence algorithm can eliminate the useless features, reduce the computation and improve the correct classification rate greatly by repeated calculation. We compare swarm intelligence algorithms of applications in identification on gas-liquid two-phase flow regime comprehensively. Discrete binary version of particle swarm optimization shows the best performance than the other two algorithms. And experiments also prove that the optimized LSSVM by PSO has a higher search speed and correct classification rate.