摘要
采用偏最小二乘回归方法,经交叉有效性检验,选取2个成分建立了某中型坦克变速箱的润滑油油液污染度与磨粒浓度、磨损烈度指数和大磨粒百分比的关系模型,该模型对各自变量的累计解释能力均达到97%以上,对因变量的累计解释能力超过98%,通过该模型计算得到油液污染度的拟合值与实测值的最大相对误差不超过7%,可以对油液污染度进行有效预测,从而为设备的正确维护提供有效的依据,实现预防性维修。
Through cross validation test,two ingredients are choosen to establish the relation model of abrasive grain concentration,abrasion intensity index,big abrasive grain and oil pollution of transmission case in some medium sized tank.The cumulative explained abilities of this model to various independent variables exceed 97% and the cumulative explain ability to the dependent variable exceeds 98%.The maximum relative error of the fitted value and the real value of the oil pollution got by calculation of this model less than 7%.The oil pollution can be forecast effectively and then the effective base for the correct maintaining are provided and the precaution maintaining is realized.
出处
《科学技术与工程》
2010年第31期7735-7739,共5页
Science Technology and Engineering
基金
军队科研计划项目(编号)资助
关键词
装甲车辆
润滑油监测
偏最小二乘回归
armored vehicles lubrication oil monitoring partial least square regression
