Linear Oscillation Diagnosis of Process Variable in Control Loop Based on Variational Mode Decomposition

Abstract

Controller tuning fall and harmonic external disturbances are big problems in industrial automation and control systems. One of the indicators that represents the problems is the presence of linear oscillations in the process variables. Diagnosis of linear oscillations in the process variables becomes an important step before increasing performance of control loops. On the other hand, the industry is being stretched towards Industry 4.0 era. The evolution of industry has had a significant impact on various aspects including the diagnosis of linear oscillations. Diagnosis of linear oscillations which were originally intended only to find out the presence of linear oscillations changed to a diagnosis method with the Industry 4.0 paradigms. This paper presents a diagnosis method for linear oscillation of process variable in a control loop based on variational mode decomposition. The proposed method will be implemented on a set of industry data to provide validation. The results prove that the proposed method can diagnose the multiple linear oscillation of process variable into 3 (three) different problems, i.e. controller tuning fall, harmonic external disturbances and nonlinear oscillation.