Towards On-line Fault Detection and Diagnostics in District Heating Systems

Abstract

This paper gives a brief introduction for detecting faults in pressure sensors and diagnostics in a district heating network. Proper pressure throughout the network is essential in maintaining the requirements for the end-user. A district heating network library has been constructed in OpenModelica for the purpose of developing a district heating network representation of Skultuna, Sweden. The use of object-oriented program will give the ease of expanding the network to encompass the entire network distribution from Mälarenergi AB, Västerås, Sweden. The physical model can then be used in conjunction with sensor data to calculate residual values. These residuals are then used as input into a Bayesian Network to determine the possibility of three different operating outcomes. This approach will allow for operators to evaluate a systems performance, help in decision support mechanisms, and can provide assistance in scheduling maintenance.