Acoustic Measurements of Soil Pipeflow and Internal Erosion

Abstract

Internal erosion of soil pipes can lead to embankment failures, landslides, and gully erosion. Therefore, non‐intrusive methods are needed to detect and monitor soil pipeflow and the resulting internal erosion. This paper presents a laboratory study using both active and passive acoustic techniques to monitor and assess soil pipeflow and internal erosion. A 140 cm long by 100 cm wide soil bed, 25‐cm deep contained a single 6 mm diam. soil pipe at 15‐cm depth that extended from an upper water reservoir to the lower bed face. The soil pipe was maintained under a constant head of 2 cm and the flow rate and sediment concentration measured at 15 s intervals while measuring soil water pressures at several locations within the bed every 30 s. Acoustic measurements were conducted every 5 s, which consisted of two parts: actively monitoring the acoustic wave propagation at four locations along the soil pipe and passively recording water flow sounds at one location. For active measurements, the phase slope method was employed to measure the P‐wave velocity under noisy and dynamic conditions. The study showed that the variation of the P‐wave velocity reflected the ongoing internal erosion processes such as the onset of soil pipeflow, the buildup of positive water pressures within the soil pipe, the saturation of soil adjacent to the pipe, the variation of water pressures within and adjacent to the soil pipe as the soil drained following removal of the constant head, and relaxation of the soil. These observations can be analyzed and understood by using the concept of the effective stress and its relationship with the P‐wave velocity. For passive measurements, passive signals (including water flow sounds and ambient noises) were recorded by a sensor buried inside the soil and close to the soil pipe. Three signal processing algorithms were applied for the passive signal analysis, which revealed similar temporal characteristic of the water flow sounds. The passive study suggested that soil pipeflow can be identified and assessed from sound levels in terms of time‐domain root mean square (TD‐RMS) and frequency‐domain root mean square (FD‐RMS) and from the contrasts of the power spectrum image.