EnergyPlus and Python Co-simulation Model to Support Machine Learning-Based Control of Ground-Source Heat Pump System

As the world pays more and more attention to energy conservation and environmental protection, the ground-source heat pump system has been rapidly popularized in China. The machine learning-based automatic control technology can help to improve the energy efficiency of the ground-source heat pump system. However, the quality of the measured data is often not clean enough to train the machine learning algorithms. On the other hand, building energy simulation can generate clean data to support machine learning-based control techniques. This study presents a framework to generate EnergyPlus Functional Mockup Units (FMUs) for co-simulating with Python environment. A case study was presented to create an EnergyPlus FMU of an office building with a ground-source heat pump system in Beijing. The basic EnergyPlus model was generated using the Commercial Building Energy Saver developed by Lawrence Berkeley National Laboratory and modified to include the ground-source heat pump system. The model was calibrated using the measured sub-metering data to meet the ASHRAE 14 requirement. To support the machine learning technique, an EnergyPlus FMU was developed to co-simulate with Python environment. EnergyPlus outputs the current room and system status parameters to Python and obtains the control signals from Python. In the future, we will work with machine learning experts to train and evaluate their control algorithms using the developed co-simulation platform.