摘要
针对常规的可靠性预测方法无法实现对数控珩磨机液压系统可靠性进行有效预测的缺点,提出了一种基于运行状态信息及支持向量回归(SVR)的数控珩磨机液压系统可靠性预测方法.该方法主要包括状态特征指标的选取、瞬时可靠度的计算以及SVR预测模型的建立.为实现对SVR预测模型的有效解算,分别采用遗传算法(GA)、粒子群算法(PSO)和混合算法实现对SVR模型的核参数的寻优计算,并比较了3种方法下SVR模型的瞬时可靠度预测精度.实例仿真结果表明,与GA及混合算法相比,采用PSO算法来解算SVR预测模型能够得到更优的数控珩磨机液压系统的可靠性预测精度.
To overcome the shortcoming of the traditional reliability predication method which cannot effectively predict the reliability of the hydraulic system of a CNC honing machine, a new predicting method based on operational status information and support vector regression theory is proposed, which includes the selection of status characteristic index, calculations of instantaneous reliability and establishment of SVR-based predicting model. To solve the SVR-based predicting model effectively, three optimization algorithms such as genetic algorithm （GA）, particle swarm optimization （PSO） and hybrid algorithm are adopted to obtain the nuclear parameters of SVR model respectively, and the predicting accuracy of instantaneous reliability with the three optimization algorithms are compared. The results of simulation instance demonstrate that more reliability predicting accuracy for the hydraulic system of CNC honing machine can be obtained by adopting PSO other than GA and hybrid algorithm to solve the presented SVR-based model.
出处
《应用科技》
CAS
2012年第6期30-33,共4页
Applied Science and Technology
基金
国家科技重大专项基金资助项目(2011ZX04002-121)
关键词
数控珩磨机
液压系统
运行状态信息
SVR
智能算法
可靠性预测
瞬时可靠度
CVC honing machine
hydraulic system
perational status information
SVR
intelligent algorithm
reliability prediction
instantaneous reliability
