Register now After registration you will be able to apply for this opportunity online.
This opportunity is not published. No applications will be accepted.
Adaptive autotuning controllers for a heat pump system
Heat pumps are essential devices for the realization of energy efficient HVAC systems. However, controlling them is challenging due to their non-linear dynamics, highly changing disturbances, and time delays, which occur in a typical installation.
Keywords: adaptive control, PID control, control of dynamic systems, nonlinear control
Heat pumps are essential devices for the realization of energy efficient HVAC systems. However, controlling them is challenging due to their non-linear dynamics, highly changing disturbances, and time delays, which occur in a typical installation. Commonly, manually tuned PID controllers are implemented, but there are several drawbacks to this approach. There are no guarantees for finding optimal parameters, the performance is degraded when process conditions change (e.g. aging, replacing some part of the HVAC installation), and its implementation is expensive in personnel costs and time. Another approach to control a heat pump, that can achieve better performance, is model predictive control, but it requires complex modeling and system identification for each installation, which results in even higher personnel costs and longer implementation time. On the other hand, auto-tuning controllers based on classical control theory can outperform the best manually tuned PID controllers, while requiring substantially less implementation time and costs[1].
[1] Killingsworth N.J. and Krstic M., PID tuning using extremum seeking: online, model-free performance optimization, IEEE Control Systems Magazine, 2006.
Heat pumps are essential devices for the realization of energy efficient HVAC systems. However, controlling them is challenging due to their non-linear dynamics, highly changing disturbances, and time delays, which occur in a typical installation. Commonly, manually tuned PID controllers are implemented, but there are several drawbacks to this approach. There are no guarantees for finding optimal parameters, the performance is degraded when process conditions change (e.g. aging, replacing some part of the HVAC installation), and its implementation is expensive in personnel costs and time. Another approach to control a heat pump, that can achieve better performance, is model predictive control, but it requires complex modeling and system identification for each installation, which results in even higher personnel costs and longer implementation time. On the other hand, auto-tuning controllers based on classical control theory can outperform the best manually tuned PID controllers, while requiring substantially less implementation time and costs[1].
[1] Killingsworth N.J. and Krstic M., PID tuning using extremum seeking: online, model-free performance optimization, IEEE Control Systems Magazine, 2006.
The goal of this project is to implement an auto-tuning PID controller following, for example, extremum seeking methodology. The performance metric includes the overshoot magnitude, settling time, convergence time, and adaptation capacity to changes in operating conditions. Multiple heat pumps are available at the ehub demonstrator at Empa in Duebendorf for experimental testing (ehub.empa.ch).
Tasks:
• Implement Extremum seeking method for auto-tuning of a PID controller for controlling a heat pump.
• (optional) Implement one more auto-tuning method from classical control theory (e.g. iterative feedback tuning).
• Evaluate the performance of the auto-tuning controller according to the specified metric above for different operating conditions.
Requirements:
Background in control theory and previous experience with real-time control implementation on PLCs or microcontrollers is required. Willingness to acquire new software tools and methods.
The goal of this project is to implement an auto-tuning PID controller following, for example, extremum seeking methodology. The performance metric includes the overshoot magnitude, settling time, convergence time, and adaptation capacity to changes in operating conditions. Multiple heat pumps are available at the ehub demonstrator at Empa in Duebendorf for experimental testing (ehub.empa.ch).
Tasks: • Implement Extremum seeking method for auto-tuning of a PID controller for controlling a heat pump. • (optional) Implement one more auto-tuning method from classical control theory (e.g. iterative feedback tuning). • Evaluate the performance of the auto-tuning controller according to the specified metric above for different operating conditions.
Requirements: Background in control theory and previous experience with real-time control implementation on PLCs or microcontrollers is required. Willingness to acquire new software tools and methods.
Dr. Bratislav Svetozarevic, Urban Energy Systems Lab, Swiss Federal Laboratories for Materials Science and Technology - Empa, Überlandstrasse 129, 8600 Dübendorf, Switzerland
ehub group: https://www.empa.ch/web/s313/ehub-group
bratislav.svetozarevic@empa.ch
Dr. Bratislav Svetozarevic, Urban Energy Systems Lab, Swiss Federal Laboratories for Materials Science and Technology - Empa, Überlandstrasse 129, 8600 Dübendorf, Switzerland