Dynamic data reconciliation to improve the result of controller performance assessment based on GMVC

Wangwang Zhu, Zhengjiang Zhang, Antonios Armaou, Guiting Hu, Sheng Zhao, Shipei Huang

Research output: Contribution to journalArticlepeer-review

Abstract

Due to the complexity of the industrial working environment, controllers are susceptible to various disturbance signals, resulting in unsatisfactory control performance. Therefore, it is especially important to assess the controller performance. Considering the harmful effect of measurement noise on controller performance assessment (CPA) based on generalized minimum variance control (GMVC), this paper proposes dynamic data reconciliation (DDR) to improve the accuracy of CPA based on GMVC. The paper first introduces CPA based on GMVC, and then analyzes the influence of measurement noise on GMVC based CPA index. DDR combined with GMVC based CPA is then proposed and analyzed in both SISO and MIMO systems to weaken the impact of measurement noise on CPA index. For both Gaussian distributed noise and non-Gaussian distributed noise, the formulation of DDR is derived from the Bayesian formula and maximum likelihood estimate. The effectiveness of the proposed method is verified in different case studies (involving both SISO and MIMO systems), and further verified by the control process of DC–AC converter. The simulation and experiment results demonstrate that the results of CPA based on GMVC can be obviously improved by using DDR.

Original languageEnglish (US)
JournalISA Transactions
DOIs
StateAccepted/In press - 2021

All Science Journal Classification (ASJC) codes

  • Control and Systems Engineering
  • Instrumentation
  • Computer Science Applications
  • Electrical and Electronic Engineering
  • Applied Mathematics

Fingerprint Dive into the research topics of 'Dynamic data reconciliation to improve the result of controller performance assessment based on GMVC'. Together they form a unique fingerprint.

Cite this