小型多旋翼无人机发动机气路故障残差参数偏差值挖掘

发布时间：2018-05-09 08:05

本文选题：无人机发动机 + 改进支持向量机　；参考：《科技通报》2016年10期

【摘要】：随着发动机工作性能的不断提高,获取的数据信号种类和数量也随之增加,为判断发动机故障增加了较大的难度。提出基于改进支持向量机数据挖掘算法完成对发动机气路故障偏差值进行挖掘和分析。首先将发动机工作状态实时监测系统收集的数据信息输入到支持向量机训练集中,计算发动机气路参数偏差值并输出;然后将输出的数据作为初始样本进行分析后,判断此数据里包括的发动机故障信息属性,以此为基础,采用指数平滑算法对不属于异常点的突变数据进行平滑处理,输出结果即为小型多旋翼无人机发动机气路故障残差参数偏差值,完成气路故障残差参数偏差值的挖掘。实验证明利用改进后的支持向量机数据挖掘算法,可以准确地计算发动机偏差值数据,更加精确地反映发动机的气路故障,为后期数据平滑处理提供了数据支持。
[Abstract]:With the continuous improvement of engine performance, the types and the number of acquired data signals are increasing, which makes it more difficult to judge engine faults. An improved support vector machine (SVM) data mining algorithm is proposed to mine and analyze the engine gas fault deviation. Firstly, the data collected by the real time monitoring system of engine working condition are input into the training set of support vector machine, the deviation value of engine gas path parameters is calculated and outputted, and then the output data is analyzed as the initial sample. On the basis of judging the attributes of engine fault information included in the data, the exponential smoothing algorithm is used to smooth the abrupt data which does not belong to the abnormal point. The output result is the deviation value of the gas path fault residual parameter of the small multi-rotor UAV engine, and the mining of the gas path fault residual parameter deviation value is completed. The experimental results show that the improved support vector machine data mining algorithm can accurately calculate the engine deviation data, more accurately reflect the engine gas fault, and provide data support for the later data smoothing processing.
【作者单位】：郑州工业应用技术学院现代教育技术中心;
【基金】：河南省高等学校重点科研项目(No.15B520015)
【分类号】：V263.6;TP311.13

【相似文献】