基于模糊聚类对柴油机试验台架一致性的分析

Analysis of Diesel Engine Test Bench Consistency Based on Fuzzy Clustering

发动机试验台架的一致性对保证发动机出厂性能一致性至关重要。本文主要以燃油量数据为实例,采用模糊聚类的方法,对柴油机试车过程数据进行分析,选取每个台架上样本均值与标准样机均值的绝对偏差和每个台架的方差,两个特征变量作为聚类的指标,并使用F统计方法确定最佳分类,最终确定一致性相对较好的台架和一致性相对较差的台架,并对后者根据相应的特征变量进行排序,输出工艺关注和分析改进的先后顺序。

The consistency of engine test bench is very important to ensure the consistency of engine factory performance.This paper mainly takes the fuel quantity data as an example,adopts the method of fuzzy clustering to analyze the data of the diesel engine test process,and selects the absolute deviation between the sample mean on each bench and the standard prototype mean and the variance of each bench,two features variable as the index of clustering,and use the F statistical method to determine the best classification,and finally determine the benches with relatively good consistency and the benches with relatively poor consistency,and sort the latter according to the corresponding characteristic variables,and output prioritize process focus and analysis improvements.