Journals Proceedings

Abstract

This article proposes an optimal design methodology for designing a pump station and a geothermal water distribution network for a district heating system for various loads. A new approach to model based design and optimization using evolutionary algorithms, is presented. The approach makes use of functional and model based reasoning techniques, by applying elementary knowledge or information to optimize locally each element in a pump and a pipe system model. The global search methods for the pump arrangement, the topology and the pipe diameter design are adaptive genetic algorithms. The basic features of the presented approximations are illustrated by redesigning a district heating distribution system in use in Reykjavik under various loads, resulting in an improved design.