Abstract:Current generation of industrial MPC technology has two main disadvantages: 1) it is very costly to implement and to maintain an MPC system; 2) the performance in disturbance reduction is low. To solve these two problems, an adaptive MPC technology is introduced. It consists of three modules, a Control Module, an Identification Module and a Monitor Module. It can perform various steps of an MPC project automatically and in a parallel manner. Thus, the efficiency of MPC commissioning and maintenance can be increased by a factor of 3 or more. In the MPC Control Module, an adaptive disturbance model is used to improve control performance and robustness. A prototype of the adaptive MPC system has been applied successfully to a PTA unit. Due to its simplicity and user friendliness, this technology may change the way MPC is applied and can make MPC feasible for all process industries, not just the refining/petrochemical industry. Recent tests have shown that the adaptive disturbance model can increase control performance considerably. A traditional adaptive MPC controller performs sample-wise identification and adaptation of both process models and disturbance models without monitoring the test condition and model quality. The new adaptive MPC system is very different from the traditional adaptive control: the process model is adapted on demand (when activated by the Monitor Module) where test signals are used in model identification; the disturbance model is adapted sample-wise without using test signals. We believe that this new scheme is more adequate for the control of large scale industrial processes than the traditional adaptive controller.

Extended Prediction Error Approach for MPC Performance Monitoring and Industrial Applications

Multi-step Prediction Error Approach for Controller Performance Monitoring

A Pragmatic Approach for Assessing the Economic Performance of Model Predictive Control Systems and Its Industrial Application

Design and implementation of an MPC system with model error detection

Get the design right, the rest is automatic: Towards an adaptive MPC controller

A Pragmatic Approach For Selecting Weight Matrix Coefficients In Model Predictive Control Algorithm And Its Application

Toward a low cost and high performance MPC: The role of system identification

A new adaptive MPC system

Some notes on MPC relevant identification

Identification of k-step-ahead prediction error model and MPC control

A multi-iteration pseudo-linear regression method and an adaptive disturbance model for MPC

Economic Performance Assessment of Multivariable Model Predictive Control Based on Linear Quadratic Gaussian Benchmark

An Improved Model Predictive Control Approach Based on Extended Non-Minimal State Space Formulation

A method of MPC model error detection

MPC Economic Performance Assessment Based on Equal-Grid LQG Benchmark

Self-Evolving Offset-Free Model Predictive Control with Model-Plant Mismatch for Dynamic Working-Point Change Tasks in Industrial Processes

Next-generation Adaptive MPC Controller

A novel robust tuning strategy for model predictive control

Learning MPC for Process Dynamic Working Condition Change Tasks under Model Mismatch

MPC Performance Assessment Based on Priority Strategy

Neural-Network-Based Predictive Control for PMSM Drives