改进欧拉算法在油液光谱分析趋势预测中的应用

Improved Euler Algorithm for Trend Forecast Model and Its Application to Oil Spectrum Analysis

基于原子发射光谱油液分析是大型机械设备磨损状态监测与故障诊断的重要技术,由于灰预测理论在趋势预测方面具有明显的优势,文章利用油液原子发射光谱分析结果,结合灰预测理论,建立了某综合传动油液中金属元素Fe趋势变化的灰预测模型。在模型参数辨识求解上首次引入了改进欧拉算法,解决了避免原灰预测模型在实际应用过程中出现的预测结果主要依赖于第一个实测值的问题,使得预测结果更准确。将该算法结合原子发射光谱分析Fe元素浓度的阈值制定,有效地捕捉到综合传动发生故障的征兆信息,及时采取措施防止综合传动的故障,具有很好的推广应用价值。

The oil atomic spectrometric analysis technology is one of the most important methods for fault diagnosis and state monitoring of large machine equipment. The gray method is preponderant in the trend forecast at the same time. With the use of oil atomic spectrometric analysis result and combining the gray forecast theory, the present paper established a gray forecast model of the Fe/Cu concentration trend in the power-shift steering transmission. Aiming at the shortage of the gray method used in the trend forecast, the improved Euler algorithm was put forward for the first time to resolve the problem of the gray model and avoid the non-precision that the old gray model＇s forecast value depends on the first test value. This new method can make the forecast value more precision as shown in the example. Combined with the threshold value of the oil atomic spectrometric analysis, the new method was applied on the Fe/Cu concentration forecast and the premonition of fault information was ob- tained. So we can take steps to prevent the fault and this algorithm can he popularized to the state monitoring in the industry.