基于DNA遗传技术的气液两相流型辨识研究

Study of Two Phase Gas-liquid Flow Regime Identification Technology Based on DNA Genetic Algorithm

利用DNA分子的复制、储存和传递信息的功能,采用基于DNA编码技术的遗传算法优化BP神经网络,建立了具有更高辨识能力的气液两相流流型辨识模型,并就若干参数对辨识模型计算效率的影响规律进行了研究。研究结果表明,基于DNA编码遗传技术对BP神经网络进行优化处理而形成的气液两相流型辨识模型比已有其它辨识模型具有更高的辨识能力;通过轴承腔气液两相流动数值仿真流型模式的辨识分析,表明这一辨识模型可应用于航空发动机轴承腔气液两相流的研究。

By using the copy, storage and information transition of DNA molecules, adopting the optimized neural network with genetic algorithm based on DNA coding, the model with higher identifying ability was built and the discipline was studied how a few parameters influence on the calculation efficiency of identification model. The results show that the identification method,in which the BP neural network is optimized by genetic method with DNA coding, is more powerful. The analysis of two phase gas-liquid regime simulation in bearing chamber proves the method can apply to two phase flow research in aero-engine bearing chamber.