Dynamic Disturbance Prediction in Model Predictive Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing Model-based Predictive Control (MPC) systems in industrial processes face inefficiencies due to the constant additive disturbance assumption, leading to oscillatory behavior and high control effort, especially in high-frequency and pulse disturbances.
Innovation Solution
A dynamic state space model with a variable disturbance prediction module is used to segregate transient and steady-state disturbances, allowing for improved prediction of future disturbances through a Kalman filter, which is integrated into the MPC system to optimize process control.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a constant additive disturbance assumption is used in MPC, then the implementation is simple, but the regulatory performance deteriorates and oscillatory behavior occurs
Solution Approach 1:
The patent applies dynamics by transforming the static constant disturbance assumption into a dynamic disturbance model that evolves over time. The disturbance is modeled as a dynamic system with state variables that capture transient and steady-state components, allowing the disturbance prediction to adapt to changing process conditions rather than remaining fixed.
Solution Approach 2:
The patent segments the disturbance into distinct components: transient disturbance and steady-state disturbance. This segmentation allows each component to be modeled and predicted separately using appropriate dynamic models, improving overall disturbance prediction accuracy compared to treating the disturbance as a single constant value.
2Power
If a constant additive disturbance assumption is used in MPC, then the computational burden is reduced, but control effort increases significantly
Solution Approach 1:
The patent applies preliminary action by predicting future disturbance trajectories in advance over the prediction horizon before the control optimization is performed. The dynamic disturbance model generates predicted disturbance values at each future time step, which are then used as inputs to the MPC optimization, allowing the controller to proactively compensate for anticipated disturbances rather than reacting to them.
3Stability of the object's composition
If a constant additive disturbance assumption is used in MPC, then the model structure is simple, but oscillatory behavior occurs in high-frequency disturbances
Solution Approach 1:
The patent applies parameter changes by introducing time-varying disturbance parameters through the dynamic disturbance model. Instead of using a single constant disturbance parameter, the model uses multiple state variables that evolve over time according to dynamic equations, allowing the disturbance characteristics to change with process conditions and frequency content.
Data Source
AI summary
A system and method for predicting future disturbance in MPC applications by segregating a transient part and a steady state value associated with the disturbance. A dynamic state space model that includes a variable disturbance prediction module can be utilized for analyzing a dynamic behavior of a physical process associated with a process model. The process model represents a dynamic behavior of the physical process being controlled and the dynamic state space model represents current deviations from the process model and future deviations over a predetermined prediction horizon. A predicted trajectory can be calculated as a response to the initial conditions estimated by a Kalman Filter for the process model extended by a disturbance model. The output of the dynamic state space model utilized for the disturbance prediction can be further provided as an estimated input to a MPC.