Existing classical techniques for monitoring pipelines are based on research works and techniques, such as robots, computational fluid dynamics, and hybrid systems. Robots are mainly used to monitor pipeline integrity. You Na et al. [18] proposed the use of biometric robots for detecting anomalies in pipelines. With an insect-like crawling feature, the proposed robot can effectively travel along the different paths of the complex pipeline network. Sujatha et al. [19] proposed a prototype robot for real-time, continuous, and autonomous monitoring of the pipeline. This was enabled by integrating the robots with a mobile application for tracking changes, which could be customised to the needs. A preliminary test conducted on the usage of the robots shows the robustness and practicability of usage. Kim et al. [10] developed a sensor-based system for pipeline monitoring and maintenance by combining topology-aware robots, mobile sensors, and a radio frequency identification (RFID) sensor-based localisation technique. Their integrated system is expected to prevent failures through the early detection and reporting of anomalies, after which, the robotic agents provide recovery by conducting repairs. Several experiments were conducted showing the cost-effectiveness and scalability of their system. However, research on robotics for overall pipeline monitoring is still in its early stages and has not been largely adopted due to its implementation complexity and the performance limitations for industrial purposes. Currently, pipeline inspection gauges are more commonly used to monitor structural integration with various limitations, such as one-directional movements.

Unlike robots, non-intrusive techniques exist, such as those based on computational fluid dynamics to detect and localise leakages in pipelines. Ostapkowicz, in his work [14], demonstrated the use of statistical leakage detection techniques based on pressure gradients and negative pressure waves. Experimental results were obtained by taking the pressure measurements and the fluid transmission speed. Both LDTs had good localisation accuracy and differed mainly in terms of energy consumption, i.e., the negative pressure wave had higher energy consumption because of its sampling rate compared to the gradient-based method, with lower energy consumption and accuracy. Similarly, Beush-ausen et al. [20] proposed transient leakage detection through an integrated analysis of the pressure and flow rate and the use of modified volume balance. With this technique, leakage detection was shown; however, some challenges persist. For instance, the localisation error was up to ten kilometres, some discrepancies in the flow metre were recorded, and there were limitations on the communication bandwidth and the effect of the type of fluid on this result. On the other hand, Santos et al. [21] proposed a detection method based on the usage of Doppler ultrasonic flow metres and transit-time ultrasonic flow metres. They carried out simulations to determine the impact of air bubbles on the accuracy of the detection and to understand its effectiveness. Karray et al. [22] also presented a leak detection technique based on the modified time of arrival difference, predictive Kalman filter, and the system-on-a-chip wireless sensor node. However, this work focused on reducing the energy consumption of LDMSs by incorporating data filtering, preprocessing, and compression in the algorithm. While existing statistical techniques present high accuracy in the DAL of leakages, challenges still exist, including energy consumption and single point of failure (SPOF) related to centralised implementation.

Other classical methods of DAL exist based on remote sensing, such as satellites, helicopters, vehicles, and unmanned aerial vehicles (UAVs) using radar, RGB images, or specialised cameras. S, Kostianoy et al. [23,24,25] presented the efficiency of using satellite systems for monitoring various pipeline transportation aspects. These include differentiating anthropogenic and natural effects of pipelines, ports, terminals, and the ecological impacts on the water bodies in proximity. The environmental impacts were evaluated using an integrated sea track web model to measure the spatial and temporal characteristics. Likewise, other works [26,27] demonstrated the use of a hyperspectral satellite for remote sensing in pipeline transportation by monitoring chemicals, gases, and other dynamic changes on the land, enabling the early detection of pipeline degradation. Although satellite-based monitoring provides multi-dimensional monitoring for pipelines and is suitable for difficult terrain, it is expensive to implement. Human-based monitoring systems, on the other hand, involve the use of observer vehicles, helicopters, or UAVs for the visual DAL of leakages (in addition to remote sensing, in some cases). Security personnel and community-based surveillance are also used to detect leakages, specifically those resulting from vandalism because of the size. However, studies [7,8,9] show that these techniques are expensive, ineffective, and impractical due to the very long detection times (days to weeks) and hazards to the environment.

3. WSN/IoT-Based Pipeline Monitoring