基于人工神经网络的燃速相关性研究

Burning rate relativity investigation using artificial neural network

关键词:  固体推进剂  推进剂燃速  相关特性  尺度效应  人工神经元网络

Key words:  Solid propellant  Propellant burning rate  Scale effect  Correlation property  Artificial neural network

摘要:

采用基于误差反向传播算法(BP算法)的人工神经网络技术,通过一维模型的分析计算,表明了该方法应用于燃速相关性研究是可行的,且具有更好的稳定性;初步建立了的燃速相关性二维模型,考虑了发动机尺寸效应对燃速的影响,并利用试验数据进行了验证,结果表明二维模型有较高的燃速预示精度。可以利用该方法开展燃速相关性分析。

Abstract:

Artificial Neural Network (ANN) theory using BP arithmetic was used to investigate the burning rate relativity of solid propellant. The result of 1D model proved the feasibility and stability of this method. A two-dimension model was also established. Experimental data was adopted to verify the model. The calculating results indicate the precision of predicted burning rate is higher than the 1D model. This method can be used to calculate the burning rate relativity of solid propellant.