Sequential simulation-optimization of a borehole heat exchanger field considering variability in heating demand and subsurface thermal responses

In installations with multiple borehole heat exchangers (BHEs), neighboring BHEs commonly influence each other. Imbalanced heat extraction and operation over many years thus causes growing interference and a pronounced thermal anomaly in the ground. This may not only have detrimental environmental effects but hamper the long-term efficiency of the entire system. As a remedy, concepts for adjusting the geometric layout and individual heat extraction loads of BHEs to a given site condition and heat demand have been presented. However, the more critical question is how to obtain an optimal and well-controlled system during the entire lifespan that often covers decades. In most of the optimization concepts that have been developed so far, the optimal load distribution is calculated on the basis of an estimated fixed heating demand. It is clear that this is not a realistic assumption due to various reasons such as changing weather conditions or a shift in consumption behavior. In our presentation, a more flexible combined simulation-optimization is introduced that takes into account a fluctuating heat demand at a monthly resolution. Here, in case of a deviation from the predefined heating demand, the BHE loads are revised and adjusted to the new conditions. Another aspect that can be considered in this sequential updating procedure is the inaccuracy of the model predicting the heat transport mechanisms in the subsurface. The new sequential optimization allows accepting a margin of uncertainty in subsurface thermal responses and taking into account thermal conditions in the ground that differ from what has been initially predicted. In our work, for demonstration, a heat demand profile with a series of monthly variations as well as a range of uncertainties in the measured temperatures of some individual boreholes in a BHE field is chosen. The optimization problem is solved efficiently by linear programming. To demonstrate the capability of the sequential procedure, fields with different densities and configurations of BHEs are studied.