Fuzzy predictive control of district heating network

This paper presents a concept for controlling the supply temperature in district heating networks (DHNs) using model predictive control. Due to the inherent non-linearity in the response characteristics caused by varying flow rates the use of fuzzy dynamic matrix control (DMC) is proposed. The fuzzy partitions of the local finite impulse response (FIR) models are constructed by an axis-orthogonal, incremental partitioning scheme. Furthermore, a novel approach for determining future fuzzy trajectory based on heat load forecasts is implemented. It is demonstrated that the fuzzy DMC performs well for the case study considered. In addition, different set point strategies are applied and the results are evaluated with respect to operational costs. In this context it is shown that the trade-off between pumping and heat loss cost plays an important role in minimising overall costs.