Water Leak Detection from Irrigation Pipelines in Botswana using Vibration Interpretation Technique

Background: The Agriculture sector in Botswana is the biggest water consumer and irrigation is a common practice. However, water leaks are common occurrence in irrigation systems. The current most prevalent methods for leak detection involves visual inspection of irrigation pipelines, which is not feasible due to water drainage in sandy soils and pipes being buried underground. Hence, the need to investigate the viability of vibration sensors for detection of water leaks. Proper water management requires constant monitoring of the water pipe network and fast detection of leaks. This paper discusses the use of an accelerometer to monitor water leaks and data analysis techniques that can be used for detecting leaks in crop irrigation systems. Objective: To detect water leaks in a farm piping system, hence help farmer reduce water loss. Method: We apply our algorithm on an experimental dataset to detect water leak. An unbiased statistic, Standardised Accuracy (SA), is used for evaluating our detection system, based upon; testing the result's likelihood relative to the baseline technique of random `predictions' (guessing), and calculation of effect sizes. Results: We demonstrate through our experiments based on experimental data sets leaks can be effectively detected. This gives an encouraging indication suggesting that it is feasible to use vibration sensors to predict water leaks in Agricultural setting. Conclusion: Using this approach it is possible for farmers to detect water leaks, even in sandy soils. Vibration interpretation techniques offer a viable option to water leak detection.