Using brillouin frequency shift to detect underground pipe leaks

Abstract

Underground pipe-networks and pipeline mains provide water to communities and industry, but some water is lost between the purification plant and the end user as Non-Revenue Water (NRW). A significant contributor to NRW is pipe leaks. Due to the fact that pipes are submerged underground, pipeline leaks are difficult to detect as there are potentially few to no visual indicators at ground level. Leaking pipes are the greatest contributing factor to the loss of purified water, accounting for more than 69% of NRW in South Africa, and accounting for between 25-50% of the total global water outlet from water purification plants, with the global average being 35%. Identifying leaking pipes to protect pipeline infrastructure and conserve water requires a sub-surface leak detection system.

An investigation was conducted on the performance of a potential leak detection system based on Distributed Fibre Optic Sensing using Fibre Optic Brillouin Frequency Shift Strain Interrogation. Brillouin Frequency Shift (BFS) is a phenomenon that occurs in optical fibres and is dependent on the fibres’ stress and temperature state.

Five different fibre optic cables were buried and monitored using the BFS phenomenon, measured along the length of the fibre optic cables at a spatial resolution of 50mm. Artificial leak experiments were conducted by introducing water into standpipes discharging at the depth of the fibre optic cables to invoke an artificial leak response. The introduction of water into the standpipes caused the moisture content of the soil surrounding the fibre optic cables to change. This change in soil moisture content caused strain and temperature changes to occur in the soil surrounding the fibre optic cables, which was in turn imposed onto the fibre optic cables. This resulted in changes in BFS in the fibre optic cables at the leak location. Monitoring fibre optic cables for changes in BFS therefore provides for a means of leak detection.

It was found from this study that it was possible to confidently detect moisture content changes resulting from the imposed artificial water leaks with each of the different fibre optic cable types tested, i.e. each cable worked as a functional leak detection sensor. The leak-induced-change in the optical fibres’ BFS was significant relative to that caused by natural temperature fluctuations and soil movements where no soil moisture content change took place. By analysing the BFS in the time domain it was also possible to determine the time of origin of the induced leaks in the experimental setup.