User-Friendly Robust MPC Tuning for Uncertain Paper-Making Processes

• Author / Creator

  He, Ning

• Paper-making processes are large-scale two-dimensional systems. For these systems, the performance along both the machine direction (MD) and cross direction (CD) needs to be ensured for successful paper production. This thesis studies the robust tuning problem of MD and CD Model Predictive Control (MPC) systems, and focuses on paper-making processes with uncertain model parameters. Four problems are solved in the work. First, the robust MD-MPC tuning for uncertain single-input, single-output (SISO) paper-making processes is explored. A sufficient condition is first derived for robust stability of MD-MPC, and then an automatic tuning procedure is proposed to achieve satisfactory closed-loop responses, as measured by overshoots, settling times and output oscillations with user-specified parametric uncertainties. Second, the robust MD-MPC tuning for uncertain multiple-input, multiple-output (MIMO) paper-making processes is studied. An efficient visualization technique is developed to characterize the set of all possible step responses for all outputs given the parametric uncertainty. An automatic tuning algorithm is then developed to achieve the desired time domain performance. In addition, a technique to predict the computation time of the tuning algorithm is proposed. Third, the robust spatial tuning of CD-MPC is investigated. The weighting matrix Sb is first appropriately designed via the real valued Fourier matrix approach to suppress high frequency components in the actuator profile. A systematic tuning procedure is then developed to adjust the corresponding multipliers to guarantee the robust stability and to reduce the variability of the measurement profile given the pre-specified parametric uncertainties. Lastly, the robust temporal tuning of CD-MPC is studied. A performance visualization technique is proposed to evaluate all possible 2-sigma (two times of the standard deviation) spreads of the measurement and actuator profiles given the pre-specified parametric uncertainties. Then, a temporal filter is adopted to smooth the MPC reference trajectory, and a systematic procedure is developed to tune the parameter in the temporal filter for robust stability and satisfactory 2-sigma performance. The effectiveness of the proposed tuning algorithms is verified through industrial examples extracted from the pulp and paper industry. By utilizing the proposed techniques, the MPC tuning parameters can be automatically determined to meet intuitive robust performance specifications for paper-making processes with easy-to-understand parametric uncertainties.

• Subjects / Keywords

  • Model predictive control
  • Controller tuning
  • Pulp and paper
  • Robust performance

• Graduation date

  Spring 2017

• Type of Item

  Thesis

• Degree

  Doctor of Philosophy

• DOI

  https://doi.org/10.7939/R34Q7R27C

• License

  This thesis is made available by the University of Alberta Libraries with permission of the copyright owner solely for non-commercial purposes. This thesis, or any portion thereof, may not otherwise be copied or reproduced without the written consent of the copyright owner, except to the extent permitted by Canadian copyright law.