20240425-PresentationOnForBitsMpc-R1.pptx

MPC

1. MODEL PREDICTIVE CONTROLLER PRESENTATION BY K SHRIRAM 2023PHXP0520H EEE Department

2. Introduction Brief overview of the importance of super heater temperature control in Thermal Power Plants. Introduction to Model Predictive Control (MPC) and its significance in advanced control strategies.

3. Typical Boiler Structure

4. Scheme with Conventional PID Controller

5. Challenges in Superheater Temperature Control Nonlinear Dynamics: The super heater temperature control system exhibits nonlinear dynamics due to the interaction between various factors such as fuel input, steam flow rate, boiler load, and other conditions. Predicting and controlling the behaviour of such a complex system can be challenging. Time Delays : There are inherent time delays in the response of the super heater temperature to changes in control inputs. These delays can result from factors such as the time taken for fuel combustion, heat transfer within the boiler, and transportation delays in the steam distribution system. Managing these delays is crucial for maintaining stable control . Uncertain Load Variations : Steam demand can vary significantly over time, leading to changes in boiler load. Sudden load changes can affect super heater temperature dynamics, requiring rapid adjustments in fuel and airflow to maintain temperature within desired limits. Transient Operation : Super heater temperature control must be effective during transient conditions, such as start-up, shutdown, and load changes. Transient response characteristics, including overshoot, settling time, and stability, need to be carefully considered in the design of control strategies.

6. Typical Scheme with Model Predictive Controller

7. Model Predictive Control ( MPC) An advanced control strategy used in industrial processes to optimize system performance while satisfying constraints. It is a control method that employs a dynamic model of the system to predict future behaviour and determine control actions that minimize a specified cost function over a finite time horizon. It is particularly effective for systems with complex dynamics, constraints, and uncertainties.

8. Components of Model Predictive Controller

9. Components of MPC Dynamic Model : MPC relies on a mathematical model that describes the dynamics of the controlled system. This model captures the relationships between control inputs, process variables, disturbances, and constraints. Prediction Horizon : MPC operates over a finite prediction horizon, which defines the time span over which future predictions are made. Control Horizon : Within the prediction horizon, MPC selects a finite sequence of control inputs known as the control horizon. Cost Function : MPC optimization is driven by a cost function that quantifies the performance of the system and reflects control objectives. Constraints : MPC incorporates constraints on control inputs, process variables, and operational limits to ensure that control actions remain within safe and feasible bounds.

10. Principle of Receding Horizon

11. Comparison of MPC to PID wrt Super-Heater Temperature Control PID Handling Nonlinear Dynamics : PID controllers are linear controllers that struggle to handle the nonlinear dynamics inherent in steam boilers, particularly during transient conditions and when dealing with large load variations . Management of Time Delays : PID controllers do not explicitly account for time delays in the system, which can lead to performance degradation and instability, especially in systems with significant transportation delays. MPC Handling Nonlinear Dynamics : MPC utilizes a dynamic model of the system, which can better capture and address nonlinearities. By considering the dynamic behaviour of the super heater and boiler system, MPC can provide more accurate predictions and control actions, especially during dynamic operating conditions . Management of Time Delays : MPC incorporates time-delay compensation through the prediction horizon, allowing it to anticipate future behavior and account for time delays in the system. This enables MPC to provide more effective control, even in the presence of delays in steam distribution and heat transfer processes.

12. Comparison of MPC to PID wrt Super-Heater Temperature Control PID Dealing with Uncertain Load Variations : PID controllers may struggle to adapt to sudden changes in steam demand and boiler load, leading to oscillations, overshoot, or slow response times . Robustness to Modeling Complexity : PID controllers do not require detailed dynamic models of the system, making them easier to implement and tune in practice. MPC Dealing with Uncertain Load Variations : MPC can handle uncertain load variations more effectively by continuously updating its predictions and optimizing control actions based on real-time measurements. By considering future system behaviour and constraints, MPC can proactively adjust control inputs to minimize deviations from set points and constraints . Robustness to Modeling Complexity : MPC relies on dynamic models, which may require more effort for development and validation. However, MPC's reliance on models also enables it to handle complex system dynamics and uncertainties more effectively, leading to improved control performance and robustness.

13. Signal flow between MPC and maxDNA

14. Switching scheme for MPC During training of MPC During running of MPC

15. Predicted value of SH steam temperature by MPC when tested with simulator

16. V ariation of SH steam temperature (Left) in MPC mode when tested with simulator at R&D

17. V ariation of SH steam temperature (Right) in MPC mode when tested with simulator at R&D

18. Prediction error in SH steam temperature (Left ) by MPC at NTPL 2x500 MW plant, Tuticorin

19. Prediction error in SH steam temperature (Right) by MPC at NTPL 2x500 MW plant, Tuticorin

20. Prediction error in SH steam temperature by MPC at NTPL 2x500 MW plant, Tuticorin

21. SH steam temperature (Left) by MPC at MAHAGENCO 2x500 MW plant, Chandrapur

22. SH steam temperature (Left) by MPC at GSECL 1x800 MW plant, Wanakbori