Performance monitoring of model-predictive controllers via model residual assessment

Sun, Zhijie; Qin, S. Joe; Singhal, Anupam; Megan, Larry

doi:10.1016/j.jprocont.2013.01.004

Cited by 69 publications

(86 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Further works continued in various directions. Modelbased approaches [3,7,42] are accompanied with minimum variance methods [47,48] that also require some process knowledge. Statistical approach through correlation analysis of optimal and working controller was proposed in [3], while prediction error benchmarking was used in [49].…”

Section: Introductionmentioning

confidence: 99%

Assessment of predictive control performance using fractal measures

Domański

Ławryńczuk

2017

Nonlinear Dyn

View full text Add to dashboard Cite

This paper presents novel approach to the task of control performance assessment. Proposed approach does not require any a priori knowledge on process model and uses control error time series data using nonlinear dynamical fractal persistence measures. Notion of the rescaled range R/S plots with estimation of Hurst exponent is applied. Crossover phenomenon is observed in data being investigated and discussed. Paper starts with industrial engineering rationale. Review of the control error histogram is followed by statistical analysis of probabilistic distribution functions (PDFs). Lévy α-stable PDF parameters seem to be best fitted. They directly lead to the fractal analysis using Hurst exponents and R/S plot crossover points. The evaluation aims at performance of the generalized predictive control (GPC) and discusses freshly introduced loop performance quality sensitivity against design parameters of the GPC controller.Keywords Control performance assessment · GPC · Fractal nonlinear analysis · Control error nonlinear dynamics · Rescaled range plot · Hurst exponent · Crossover

show abstract

Section: Introductionmentioning

confidence: 99%

Assessment of predictive control performance using fractal measures

Domański

Ławryńczuk

2017

Nonlinear Dyn

View full text Add to dashboard Cite

show abstract

“…Since predictions from a poor model can lead to control action that is far from optimal, high model quality is key to good performance in an mpc loop. The most expensive and time-consuming part of mpc commissioning is frequently cited as modeling, with up to 80 % of the design effort often spent on obtaining a model suitable for mpc (Sun et al [1]). As the controlled process may change over time, the verisimilitude of the model may decrease and lead to lower performance.…”

Section: Introductionmentioning

confidence: 99%

“…As the controlled process may change over time, the verisimilitude of the model may decrease and lead to lower performance. There are many possible sources of performance deterioration in mpc loops, including inappropriate setup of constraints, inconsistencies in dynamic optimization in mpc and the higher-level optimization, and poor quality of the input-output and/or disturbance models; among these, the model quality is the most significant for control performance in mpc (Sun et al [1]). …”

Section: Introductionmentioning

confidence: 99%

MPC-based dual control with online experiment design

Heirung

Foss

Ydstie

2015

Journal of Process Control

View full text Add to dashboard Cite

“…The performance degradation can be caused by model plant mismatch, bad tuning , bad set of soft and hard constraints, and unmeasured disturbances (Sun et al, 2013). Among these many sources, the poor model quality is the most frequent and impactful one.…”

Section: Introductionmentioning

confidence: 99%

“…A class of the methods such as: Huang et al (2003), Conner et al (2005), Jiang et al (2012) are based on investigate the need of system re-identification. Other approaches (e.g., Badwe et al (2009), Kano et al (2010), Ji et al (2012) and Sun et al (2013)) are looking for the locations in the model (i.e. the pairs controlled-manipulated variables) responsible for the performance degradation.…”

Section: Introductionmentioning

confidence: 99%

Assessment of Model-Plant Mismatch by the Nominal Sensitivity Function for Unconstrained MPC

et al. 2015

View full text Add to dashboard Cite

Model Predictive Control (MPC) is a class of control systems which use a dynamic process model to predict the best future control actions based on past information. Thus, a representative process model is a key factor for its correct performance. Therefore, the investigation of model-plant-mismatch effect is very important issue for MPC performance assessment, monitoring, and diagnosis. This paper presents a method for model quality evaluation based on the investigation of closed-loop data and the nominal complementary sensitivity function. The proposed approach ensures that the MPC tuning is taken into account in the evaluation of the model quality. A SISO case study is analyzed and the results show the effectiveness of the method.

show abstract

Performance monitoring of model-predictive controllers via model residual assessment

Cited by 69 publications

References 13 publications

Assessment of predictive control performance using fractal measures

Assessment of predictive control performance using fractal measures

MPC-based dual control with online experiment design

Assessment of Model-Plant Mismatch by the Nominal Sensitivity Function for Unconstrained MPC

Contact Info

Product

Resources

About