Short Term Operation of District Heating Systems: An Application of Mathematical Programming
This work treats the introduction of rigorous optimisation methods as a decision aid in the short term operation of district heating systems. The conditions for operation of district heating systems are investigated. Arguments for the necessity of a systems model that is capable of handling the discrete nature of the problem are presented.
A generic representation of district heating systems, representative for Swedish conditions is established. Subsequently a mixed integer programming modelling framework for the most complex case is developed.
The model covers unit commitment and operation of heat storage devices over a time period of one to several days, accounting for minimum load constraints, start-up costs, ramp-rate constraints, common steam headers, and other technical constraints, continuous and logical. The intricate structure of the Swedish system for taxes and emission fees is explicitly included in the model. The non-convexity resulting from the conditions for short term electricity exchange is modelled through the introduction of additional integer variables in order to guarantee global optima. It is also shown how dual information from the solution can be used to asses the economic advantage of storing heat in the network itself.
The model is implemented for use in the control room of a major Swedish district heating utility displaying all the possible complexity of the generic system.
The main experience from the development and implementation of the model is that the creation of a model is not a limiting factor, but rather the willingness and ability on behalf of the user to exploit the advantages of a systems model.