CAJ | 학술논문

为了提高非线性特征诊断的效率，提出了基于数据驱动的有向图和高阶统计的控制系统非线性故障检测方法。所提出方法以解释结构模型为基础，采用偏相关系数确定邻接矩阵，并利用基于工艺流程的先验知识，确定可达过程控制回路，合理选择数据的采样周期以及数据窗口，并采用高阶统计动态计算完全非线性指数，非高斯指数和非线性指数来提取故障特征，从而消除非线性特征诊断方法由于数据选择带来的不确定性，提高了控制系统回路故障诊断的效率。以TE过程为对象，验证了提出方法的有效性，并取得了良好的效果。
위료제고비선성특정진단적효솔，제출료기우수거구동적유향도화고계통계적공제계통비선성고장검측방법。소제출방법이해석결구모형위기출，채용편상관계수학정린접구진，병이용기우공예류정적선험지식，학정가체과정공제회로，합리선택수거적채양주기이급수거창구，병채용고계통계동태계산완전비선성지수，비고사지수화비선성지수래제취고장특정，종이소제비선성특정진단방법유우수거선택대래적불학정성，제고료공제계통회로고장진단적효솔。이TE과정위대상，험증료제출방법적유효성，병취득료량호적효과。
To improve efficiency of Nonlinear Characteristics Diagnosis, the method which combined data-driven SDG and Higher-order statistical is proposed. Based on interpretative structural model, this approach determined adjacency matrix using partial correlation coefficient, then defined the control loops of reachability with prior knowledge of process and reasonably chose the sampling period of the data as well as data window. Then, Non-Gaussianity Index, Nonlinearity Index and Total Nonlinearity Index were calculated to effectively eliminate uncertainty of Nonlinear Characteristic Diagnostic Method which results from data selection. This method increased efficiency of diagnosis of control system. At last, the proposed method is applied to TE process effectively and obtained some good performances.