All news

KU LEUVEN BEING BOLD AT BOULDER

Posted in: News

KU Leuven will be presenting at the Intelligent Buildings Operation Workshop in Boulder Colorado, Aug 7-9 2019

'Development and impact of a borefield controller model for Model Predictive Control’

Iago Cupeiro Figueroa, Damien Picard, Lieve Helsen

This paper is result of collaboration within IBPSA Project 1, also supported by the hybridGEOTABS project

‘The strength of white-box model predictive control in buildings’

Filip Jorissen, Damien Picard, Lieve Helsen

 

‘State estimation of control-oriented white-box models for buildings’

Jan Drgona, Iago Cupeiro Figueroa, Lieve Helsen

 

 

 

 

 

Title of proposed presentation:

 

Development and impact of a borefield controller model for Model Predictive Control

 

Name of presenting author: Iago Cupeiro Figueroa

Affiliation of presenting author: KU Leuven, EnergyVille

Supporting author(s) and affiliation(s): Damien Picard(1) and Lieve Helsen(1,2), KU Leuven; KU Leuven (1) and EnergyVille (2)

Contact email: iago.cupeirofigueroa@kuleuven.be

 

Body of abstract:

 

Model Predictive Control (MPC) is an optimal model-based control strategy that has shown significant energy savings potential in the operation of building energy systems. However, one of its main current challenges is to get accurate controller models that describe the behaviour of the system components well, but also simple enough to allow tractable computation times during optimizations. For buildings equipped with hybrid systems that include geothermal borefields, the borefield controller model is especially critical:

the prediction of the return fluid temperature of the borefield will affect not only the efficiency of the heat pump and passive cooling heat exchanger, but also its operation feasibility. For example, applying a sudden peak load to the borefield could cause freezing of the immediate surrounding ground, consequently having to resort to a less-efficient secondary system, while this could be avoided by spreading the load over a longer period instead, and thus applying smaller thermal loads. A short-term dynamic model is therefore a high added value to the controller. However, it is also important to consider the long-term scales, due to the thermal interaction between boreholes in a borefield and possible load unbalances, in order to guarantee a sustainable use of the borefield over the design period. This is especially challenging in optimal control of buildings, where the prediction horizons are in the range of a few days while the design period is in the range of decades.

 

In this presentation, we present a borefield controller model that tracks its return temperature and we evaluate the impact of its use on MPC results.

First, we show the modeling approach to construct a borefield controller model based on the simulation borefield model (1) from the IBPSA Modelica library (2). The ground response in the proposed controller model is approximated by a physics-based RC network and the borefield model is non-linear due to advection heat transfer. This controller model is evaluated in a virtual-test bed building that is equipped with a hybrid geothermal system. The virtual-test bed is constructed in the Modelica language using the IDEAS (3) library. Two approaches are compared: a controller model of the building and its heating and cooling system is set-up (i) approximating the borefield as an ideal source and (ii) including the proposed borefield controller model. A closed-loop MPC simulation is carried out using these two approaches in the Toolchain for Automated Control and Optimization (TACO (4)). Perfect state update is applied by using the emulator, as such capturing the long-term dynamics of the borefield: the ground temperatures at different locations in the borefield can be easily updated by using the emulator and communicating the result to the controller model at each optimization time-step. The benefits of including the proposed borefield model are evaluated by quantifying the energy savings, global system efficiency, variation of the fluid temperature over the whole year and computation time.

 

Acknowledgements

 

The authors acknowledge the financial support by the European Union through the EU- H2020-EE-2016-RIA-IA programme for the project ‘Model Predictive Control and Innovative System Integration of GEOTABS;-) in Hybrid Low Grade Thermal Energy Systems - Hybrid MPC GEOTABS’ (grant number 723649 - MPC-; GT).

 

References

 

1 A. Laferriere, M. Cimmino, D. Picard & L. Helsen (2019). Development and validation of a full-scale semi-analytical model for the short- and long-term simulation of vertical geothermal borefields. Submitted to Geothermics.

2 IBPSA Project 1: https://ibpsa.github.io/project1/

3 IDEAS library: https://github.com/open-ideas/IDEAS

4 F. Jorissen, W. Boydens & L. Helsen (2019). TACO, an automated toolchain for model predictive control of building systems: implementation and verification, Journal of Building Performance Simulation, 12:2, 180-192, DOI:

10.1080/19401493.2018.1498537