Optimal Sensor Placement and Influence of Noise on Pressure Wave Evaluation for Leakage Localization in a District Heating Network

Abstract

Pipe bursts and leakages in district heating networks are a problem both for the economic operation and the supply reliability. In order to minimize costs and secondary damages, an efficient and accurate leakage localization is necessary. If this is possible, only the affected network part needs to be shut down and operation in the remaining heating network can be maintained. One method will be the pressure wave detection. In case of a large spontaneous leakage, the pressure wave propagates throughout the entire network. Each sensor within the network is reached at a different time point due to its position. This allows a leakage localization. In this paper, a framework published previously was extended to deal with noise. Leakage attribution to exclusion areas is calculated based on pressure drop time points extracted from measurement data. A performance criterion is used to evaluate how well leakage localization is working. The framework can be used for optimal sensor placement and for the investigation of noise. Optimal sensor placement for up to five sensors in a network with 90 km trench length is calculated with a high performance criterion. Further, the influence of noise of the pressure drop time points is investigated. To this end, the extent of noise on the pressure drop time points determined from real measurement data of ten real events was examined. Acceptable noise levels, which allow sufficient leakage localization quality, are evaluated for different numbers of optimal placed sensors. The results are presented and further improvements are discussed.