Study Of The Identification Methods Of Gas-Liquid Two-Phase Flow Regime And The Analysis Realization In Bearing Chamber

When aero-engine bearing chamber rotates with high speed, there is gas-liquid two-phase flow in it. Gas-liquid two-phase flow regime influents the design of the bearing chamber lubrication and its thermal analysis greatly. So the study of the identification methods of gas-liquid two-phase flow regime is important. The experiment of gas-liquid two-phase flow regime identification costs too much and it's difficult to identify the flow regime correctly, so the paper aims to build a theoretical model, from which we can identify the gas-liquid two-phase flow regime in bearing chamber.The prototype of the study comes from the gas-liquid two-phase flow in pipes. We build the model on the basis of theoretical method of using neural network. By comparing the recognition possibility of three types of network, BP network is used to identify the flow regime more precisely. Because of the weakness of BP network, it can't identify some flow regime correctly. So we further use genetic algorithm to train BP neural network and optimize the weights of the network. Based on it we build the model from which we can identify the flow regime. Because DNA genetic algorithm improves greatly in coding, we make further study in using DNA genetic algorithm to optimize BP network and building a more accurate model. Then a research with finite element method(FEM)has been conducted for the two-phase flow in bearing chamber and the analysis data are collected. We use the model to identify the flow regime in bearing chamber and validate the method of using DNA genetic to train BP network.