基于熵方法的柴油机试验数据粗大误差分析.pdfVIP

基于熵方法的柴油机试验数据 粗大误差分析 刘磊，王铁，石晋宏，杨甜甜，曹贻森 （太原理工大学车辆工程系 太原 030024） 摘要：在四缸增压中冷柴油机上进行国标柴油试验，研究了不同批次国标柴油在柴油机上的经济性及排放 性，同时引入测量信息论中熵判别方法计算测试样本的熵及不确定度，判别试验数据在置信区间内是否含 有粗大误差。研究表明：在处理小样本数据时，熵判别方法可以对数据进行有效评价。本研究中，在1600rpm 且置信度为 95%时，各工况燃油消耗率的置信区间分别为896.25±8.97、324.58±3.68、255.17±2.36、 234.58±2.40、224.24±2.74；NOx 排放的置信区间分别为232.00±22.81、492.86±32.24、844.43±66.86、 1075.00±60.88、1371.71±78.37；排气温度的置信区间分别为161.00±8.48、180.29±8.31、231.43± 8.27、300.86±9.83、364.71±10.47。 关键词：柴油机试验数据；小样本；熵方法；粗大误差 Abstract: Tests had been done on a four-cylinder turbocharged and intercooled diesel engine burned different batches of national standard diesel for researching engine emissions and economy characteristic , while measurement information theory entropy was introduced for calculating entropy and uncertainty of test samples by entropy method, discriminated if there is gross error exists, calculated the confidence interval by 95% confidence level. Studies showed that: in dealing with small sample data, entropy discrimination method the data can be effectively evaluated. In this paper, each operating conditions at 1600rpm, the confidence intervals of fuel consumption rate were 896.25±8.97, 324.58±3.68, 255.17±2.36, 234.58±2.40 and 224.24±2.74; while the confidence intervals of NOx emission were 232.00±22.81, 492.86±32.24, 844.43±66.86, 1075.00±60.88 and 1371.71±78.37; the confidence intervals of exhaust temperature were 161.00±8.48, 180.29±8.31, 231.43±8.27, 300.86±9.83 and 364.71±10.47. Keywords: Engine test data; Small sample; Entropy method; Gross error 随着我国综合国力的提升，汽车行业 1 试验部分 得到了飞速发展，机动车保有量也逐年增 加，在城市，汽车尾气污染日益严重，已 1.1 试验设备 经成为自然环境和人类健康的最大威胁之 [1-2]