Colloquia

Samenvatting

As part of the ongoing energy transition, energy systems are becoming decentralized, dynamic and complex. To manage this complexity, Energy Management Systems are being developed to predict, plan and control energy assets. This master thesis addresses the prediction and planning of a decentralized heating system at Sparkling Projects, located within the Ecofactorij industrial site in the Netherlands.

The system under study consists of an electric boiler, a heat pump, a water buffer and an office building that acts as the heat consumer. The heat pump’s coefficient of performance is modeled as a function of outdoor temperature using polynomial regression and indoor temperature is predicted with a reduced-order multiple linear regression model that accounts for building thermal inertia, heat input, ambient conditions and unmodeled disturbances such as occupancy and solar gains.

These models are integrated into a two-stage scheduling framework. The first stage forecasts the heat demand required to meet comfort setpoints from predicted outdoor temperature, while the second stage solves a mixed-integer optimization problem that respects system constraints, power and energy balances and an objective function with adjustable weighting factors for economic performance, sustainability and user comfort. Several combinations of these weightings define different operating strategies.

Simulation scenarios vary outdoor temperature, PV capacity, electricity prices and emission factors and key performance indicators such as total energy cost, CO₂ emissions, comfort violations, maximum grid import/export and PV utilization are calculated for each case. The baseline strategies, named Comfort, Economic and Sustainable, validate the algorithm, while hybrid strategies illustrate trade-offs. Results show that the multi-objective strategy “Sustainable-Economic,” which excludes a comfort weight, performs worse than both the “Balanced” and “Sustainable-Focused” objectives, which yield identical outcomes.

A sensitivity analysis examines how indoor temperature setpoints and flexibility levels affect performance. A setpoint of 21 °C provides the best compromise between comfort, costs, productivity and carbon emissions, while a flexibility factor of 50 % offers the most favorable trade-off in energy cost and emissions.