论文部分内容阅读
提出敏感稀疏主元分析(SSPCA)算法用于监测复杂的化工过程.根据主元分析与数据矩阵奇异值分解之间的关系,通过将L_(2,1)范数作为目标函数和惩罚项得到一个获取稀疏主元负载的凸优化问题,并通过一个迭代算法进行求解.SSPCA算法能同时兼顾大得分主元与小得分主元在监测算法中的作用,提高了其对故障的敏感度.证明了SSPCA算法的单调性和全局收敛性,对田纳西伊斯曼过程一个算例的监测结果表明了SSPCA算法的有效性.
A sensitive sparse principal component analysis (SSPCA) algorithm is proposed to monitor complex chemical processes.According to the relationship between principal component analysis and singular value decomposition of data matrix, the L_ (2,1) norm is obtained as the objective function and penalty term A convex optimization problem for obtaining sparse principal component loads and solving it by an iterative algorithm.SSPCA algorithm can take the role of both large-scoring principal component and small-scoring principal component into account in monitoring algorithm, and improve its sensitivity to failure. Monotonicity and global convergence of the SSPCA algorithm, the monitoring results of an example of the Tennessee Eastman process show the effectiveness of the SSPCA algorithm.