Nonlinear and dynamic optimization for short-term production planning in district heating networks

Abstract:

Production planning in district heating networks aims at finding a cost optimal scheduling of the heat and power production plants, which satisfies both the network load demand and operational constraints. The dynamic optimization problem that involves both continuous and discrete variables is typically solved using linearized models with the energy flows as continuous decision variables. This simplified approach based on non-physical plant models is limited as it cannot well describe critical parameters such as the energy stored in the network, the electrical efficiency of the co-generation plants or temperature and pressure constraints at the customer substations.
In this talk, it will be shown how physical plant models can be integrated into the standard approach to substantially improve the production plans. The idea is to first solve for the status (on/off) of the production units using simplified models and thereafter focus on the economic dispatch using nonlinear models and optimization techniques. The feasibility and potential of the proposed method have been shown using experimental data from a Vattenfall owned plant in Nyköping, Sweden.