原子发射光谱是分析油液中微小磨损颗粒元素浓度的重要方法。 作为一种非直接测量方法, 油液光谱数据是车辆综合传动装置可靠性评估中的系统性能劣化的重要监测指标, 可用于系统失效评估与剩余寿命预测。 针对油液光谱数据这类型的一元劣化失效, 随机过程尤其是Wiener过程模型具有良好的计算分析性质, 在基于性能劣化的可靠性分析中应用日趋广泛。 通过对车辆综合传动装置运行中的实时采样, 共取得50个油液光谱样本。 采用其中三种指示元素的线性回归方程来计算综合传动装置运行中每个瞬时的特征值与均值。 基于正漂移Wiener过程, 建立了综合传动装置的劣化失效预测模型, 并基于R语言环境进行了随机微分方程的仿真与求解。 得到了油液光谱中的Fe, Cu和Mo元素含量增长趋势的预测结果以及三种指示元素各自的首中时间。 经比较, 劣化失效周期的预测值较之条件维护时间延长了27 Mh(15.9%)。 维护时间的延长, 能够有效的减少全寿命周期内的维护次数, 并最终降低维护成本。 研究结果表明, 该方法适用于综合传动装置的磨损与失效预测、 全寿命周期费用与维护计划的优化。 同时, 也可推广至其他复杂机械系统的失效预测与评价等相关领域。

The most common methodology used in element concentration measurement and analyzing of wear particles is Atomic emission (AE) spectroscopy. As an indirect measuring method, the oil spectral data is introduced to indicate the performance degradation and the residual life prediction in the reliability evaluation of Power shift steering transmission (PSST). Stochastic methods especially the Wiener process is convenient in solving and analyzing the unitary degradation failure indicated by the oil spectral data. The oil data have been sampled in the real operating condition, and the data set has more than 50 samples taken from PSST. The mean values and time-dependent characteristics of three indicating elements are statistically obtained by the linear regression analysis. The model of the degradation and failure prediction has been proposed based on the Wiener process with the positive drift. For modeling and simulation the software R was used. Therefore, the trend curves of diffusion process with their First Hitting Time have been predicted. Through comparison, the time intervals of condition-based maintenance have been extended as 27 Mh (15.9%). This will save the cost of maintenances by eliminate the preventive maintained cycles. The advantage and novelty of the outcomes presented in the article are that the stochastic process might be applied for predicting the degradation failure occurrence and also for optimizing the maintenance intervals and the cost-benefit. As might be expected, the method can be extended to other cases of wear prediction and evaluation in complex mechanical system.