基于混合算法的柴油机缸内工作过程评估  

Evaluation of the Working Process of Diesel Engine Based on Hybrid Algorithm

作　　者：蔡建邦 CAI Jian-bang(School of Marine Engineering Guangzhou Maritime University,Guangzhou Guangdong 510725,China)

机构地区：[1]广州航海学院轮机工程学院,广东广州510725

出　　处：《广州航海学院学报》2018年第4期29-35,共7页Journal of Guangzhou Maritime University

摘　　要：针对船舶柴油机故障数据少,而正常状态数据丰富的特点,建立粒子群支持向量数据描述的混合算法,对柴油机缸内工作过程进行状态评估.选取并处理缸内气体压力变化数据,并将其作为基准数据,对示功图特征参数进行提取与计算,运用粒子群优化算法获得最优参数,并代入支持向量数据描述算法模型,以获得最小封闭超球体理论半径,将状态点距离正常工作过程理论球心位置的数值与理论半径的数值作对比,以评估柴油机缸内工作过程,并采用待测半径占理论半径相对长度的方式对评估结果进行打分.通过实例验证,表明该方法具有良好的实船应用价值.There are few fault data of Marine diesel engine,but normal data is rich.The hybrid algorithm of PSO-SVDD is established to evaluate the working process of the diesel engine cylinder.Selecting and processing the gas pressure change data in the cylinder and serving as the benchmark data.The characteristic parameters of the indicator diagram are extracted and calculated.The optimal parameters are obtained by using the PSO,then using the SVDD to obtain the least closed hypersphere theoretical radius.The numerical value of the theoretical sphere is compared with the theoretical radius,the result uses to evaluate the working process of the diesel engine cylinder.The evaluation result is graded by means of the relative length of the theoretical radius.The result shows that the method has good practical value.

关 键 词：状态评估 混合算法 示功图 粒子群优化算法 

分 类 号：TB421.2[一般工业技术]