Dynamic Data Reconciliation
for Enhancing Performance of Minimum Variance Control
in Univariate and Multivariate Systems
Zhengjiang Zhang
Junghui Chen
10.1021/acs.iecr.6b02532.s001
https://acs.figshare.com/articles/journal_contribution/Dynamic_Data_Reconciliation_for_Enhancing_Performance_of_Minimum_Variance_Control_in_Univariate_and_Multivariate_Systems/3980694
Healthy
controllers are required in order for the control systems
to maintain a high level of performance. In past research, minimum
variance control (MVC) played a crucial role as a benchmark in performance
monitoring because of the attractive theoretical and computational
properties associated with it. Since the influence of measurement
errors has not been explicitly considered in the MVC theory in stochastic
control systems, this paper first analyzes the influence of measurement
errors on the control performance of MVC in both univariate and multivariate
systems. And then the dynamic data reconciliation (DDR) methods are
proposed and combined in the procedure of MVC/multivariate MVC (MMVC)
to decrease the influence of measurement errors and to enhance the
control performance. Considering both random measurement errors and
gross errors, the effectiveness of MVC/MMVC combined with DDR on the
variances of process outputs is illustrated by two simulated cases,
including both univariate and multivariate systems. An actual pilot
scale experiment with two inputs and two outputs is used to demonstrate
the effectiveness of MMVC combined with DDR. Results of simulation
and experiment show that MVC/MMVC combined with DDR can efficiently
enhance control performance.
2016-09-08 00:00:00
multivariate systems
influence
pilot scale experiment
DDR
control performance
MVC
measurement errors
Dynamic Data Reconciliation
control systems
Minimum Variance Control
MMVC