Control Performance Assessment And Its Industrial Application

Control system is widely-used in industrial production. The real-time monitoring of control loops and controller performance assessment become necessary when the performance is degraded after drifting of operation condition along the running time.Considering that the operating data hardly follow a uniform Gaussian distribution in complex industrial processes, the Gaussian mixture model (GMM) is utilized as the tool of control performance assessment (CPA) in this article. The consolidation operator is proposed and introduced to expectation maximization (EM) algorithm. The new approach can automatically optimize the number of Gaussian components on one hand, and solve the poor convergence problem of F-J method when Gaussian components overlapping too much during initialization on the other. With the obtained model, a criterion based on global probability is devised for real-time CPA with routine operating data. To evaluate the difference of control loops, a novel approach is proposed for controller performance assessment in SISO system which is called performance vector based assessment (PVA). It gives a numerical benchmark of performance current and expected, through which an improvement potential index can be obtained providing a reference for plant engineers to determine if the controller deserves retuning. On the basic of theoretical study, the ethylene cracking furnace is taken as the research object for control loop monitoring and performance analysis. The control platform is created in TPS 3000 which is the DCS of Honeywell. Then both PVA and INTUNE are used to make assessment.