oa Protecting Water Resources via Smart-Sensing Infrastructure: From Silicon to Systems

Accessible fresh water in rivers and lakes represent only 0.03 % of Earth's total water reservoir [1]. The protection and efficient utilization of this precious and limited resource is one of the most pressing issues in the 21st century. Today, it is estimated that 22 % percent of urban water resources is lost in the infrastructure due to leakage [2]. This article proposes a next-generation smart-sensing platform aimed at providing low-cost water-pipe inspection and leakage detection. Pipeline inspection techniques used in the traditional oil and gas industries are often not directly applicable to water systems [3]. Industry-standard pigging platforms require special launch and recovery facilities and they can not tolerate complex surface conditions caused by bio-fouling and corrosion. The current generation of water-pipe surveying instruments rely on ultrasound distance measurement which is prone to interference from road traffic, construction, and air pockets. The wavelength of ultrasound also limits its sensing accuracy [4]. Furthermore, all existing solutions requires highly trained on-site operators and thereby incurring significant deployment cost. The proposed smart-sensing platform in Figure 1 is designed to be fully autonomous, low-maintenance, and non-invasive to the existing infrastructures. The in-pipe roving sensor detects water leakage and wall-thinning and communicates this information in real-time via an acoustic-radio hybrid cellular sensor network. Portable self-powered base-stations are installed along the water-pipe at intervals of 10-100 meters and communicates with the roving sensor using a wide-band 1 MHz ultrasonic channel which, unlike electro-magnetic radio waves, can penetrate the metallic pipe wall without invasive retrofitting. The attenuation problem at this frequency range [5] is solved by the short distance between neighbouring base-stations. Channel-State-Information (CSI) is used to optimize transmitter power allocation; delays are tolerated in exchange for longer battery life. The data packets received from the roving sensor are relayed between the base-stations via the electro-magnetic radio frequency (RF) medium to the central server. The roving sensor in Figure 2 uses a compressive image sensor assisted by an acoustic transceiver to visually detect leakage sites. The image sensor's on-chip image compression (Figure 3) is facilitated by a novel Analog-to-Information architecture which allows the image to be sampled at sub-Nyquist rates with significant power savings in both image capture and processing [6]. This optic-acoustic hybrid fault detection scheme allows the image sensor to be utilized more efficiently by only waking up the image sensor when fault detection likelihood is high. The bases-station integrates a number of sensors (Figure 4) to provide complimentary pipe-line status information. Temperature [7] and humidity sensors are installed both near the pipe and at ground level. Leakage sites can be detected by observing a drop in temperature and rise in humidity in the soil near the pipe when compared to the ground level references. Zinc-oxide nano-wire gas sensors [8] is added to the sensing repertoire. All sensors on the base-station share a common ADC and analog circuit to minimized power consumption and cost. The base-station is self-powered by a solar-cell panel and a small backup battery.