Multi-Stage Optimization for Long-Term Building Climate Operation with Seasonal Thermal Storage

In an effort to minimize the environmental impact of space conditioning and water heating in buildings, seasonal heat storage has gained attention recently. We propose an operational strategy for a building in combination with a environmental friendly heat storage. To reduce the computational effort of the long-term optimization problem, we reduce and solve it in two steps. First, a simplified building model is used for annual input reference generation with an increased sampling time, minimizing energy cost. Subsequent, the input for the non simplified building is generated with a second, short-horizon optimization problem, minimizing cost and tracking the annual input reference with higher time resolution. The results show that deviations due to model and discretization errors occur, but the long-term reference is necessary to optimally operate the heat storage. Moreover, the long-term reference can be recomputed within one sampling interval when deviations are too high. The resulting operation principle is the charging of the thermal storage in times of low energy cost, and decharge in between. The storage is not filled once in summer and emptied in winter, but goes through multiple storage cycles during the year. The long-term reference optimized with the simplified model is a good approximation for the energy cost and consumption with the non simplified model.