基于AIC-K-means的载荷分段混合分布估计

Estimation of piecewise mixed load distribution based on AIC-K-means

针对工程机械作业载荷分布估计的问题,基于AIC准则及K-means算法,提出了一种载荷分段混合分布估计方法。根据Luise分组方法,得到正态分布、对数正态分布及威布尔分布的卡方检验值。选取最小卡方检验值对应的分布函数作为分段混合分布各段的拟合函数,通过极大似然估计算法获得分段混合分布的表达式。结合挖掘机土石挖掘工况,将本文方法应用到主泵出口压力载荷数据中,并与现有方法进行对比分析,本文方法得到的结果通过了卡方检验。同时,用1000组仿真数据来验证方法的可行性,得到了参数估计的误差仅为0~18%。

In order to solve the problem of load distribution in construction machines,an estimation method of piecewise mixed load distribution based on Akaike Information Criterion（AIC）K-means is proposed.First,according to Luise method,the chi-square test values of the normal distribution,logarithm normal distribution and Weibull distribution are calculated.Then,the corresponding distribution of the minimum test value is chosen as the fitting function of the piecewise mixed distribution.Finally,the formula expression of the distribution is obtained through the maximum likelihood estimation.The proposed method is applied to the outlet pressure load time history of the excavation operation the excavator,and compared with the existing method.The result obtained by the proposed method passes the chi-square test.The simulated data of 1000 groups are used to conduct feasibility analysis and the estimated parameter error range is 0~18%.