Model Predictive Control Applied to Ground Source Heat Pumps

Building heating is one of the most important sources of energy consumption and Ground Source Heat Pump (GSHP) is an efficient alternative particularly used in the Nordic countries. However, the GSHPs available for the consumer market today typically only utilize basic control schemes that are relatively inflexible. Model Predictive Control (MPC) appears as a promising approach to combine both comfortable indoor temperature and reduced consumption, notably using weather forecasts. The present thesis considers a typical user case of a single family house heated by a ground source heat pump willing to reduce its energy consumption and environmental impact. Our perspective is to create a control strategy to be applied on top of the existing heat pump system and with almost no additional hardware needed. Specific attention is dedicated to the system’s efficiency in order to reflect the real working performances of a ground source heat pump. Different scenarios are developed, including varying electricity prices using Economic MPC. Simulations suggest MPC approach is most beneficial when considering both time varying comfort and prices with savings up to 8% despite the loss of efficiency of the heat pump. Practical applications are considered with system identification and experimental testing implementing a running MPC controller.