采用傅里叶变换红外光谱法测定了航空润滑油中的水分,通过遗传算法(GA)优化选取有效波数点,用误差反向传播神经网络(BP-ANN)进行水分预测计算。模型的预测相关系数为0.957,预测标准偏差为0.022。随机抽取某型航空润滑油样品进行预测并对预测结果进行配对t检验,结果表明:红外光谱定量分析结果与标准方法测定值没有显著性差异,模型可以用于该型在用航空润滑油水分含量现场快速检测。

In fourier transform infrared spectrometric (FTIR) determination of content of moisture in aviation lubricants, the algorithm of genetic algorithms (GA) was applied to the select of a few points of significant wave and the algorithm of back propagation-artificial network (BP-ANN) was applied to presumptive calculation of moisture. Value of correlation coefficients of the model was achieved to 0.957, and the square errors of prediction (SEP) was 0.022. The proposed method was used in the forecast of a certain type of aviation lubricants on randomly selected, and the paired t-test was used in the test of predicted values. It was found that there was no significant difference between the standard method and IR quantitative analysis. The model can be used for the determination of moisture in aviation lubricants.