The monitoring approaches and non-destructive testing technologies for sewer pipelines

The drainage system is likened to the vein of the city, responsible for transporting domestic wastewater and rainwater to the sewage treatment plants or rivers, which together with the pipelines for water supply, heating and gas constitute the lifeline of a city (Keith 2008; Moss 2020).

To date, the total length of underground pipelines in China has been increasing with urbanization to nearly 3 million kilometers, some of which are decaying and aging over years, resulting in significant treatment and rehabilitation costs. A part of the reason for the deterioration of the pipelines is the aging or corrosion of the building materials, another is attributed to structural changes out of the design expectation due to traffic overload, construction disturbance, etc. (Hadjmeliani 2015; Rakitin & Xu 2015; Vanaei et al. 2017). As a result, infiltration, seepage, and structural damage are frequent. In addition, infiltration will carry away the surrounding soils causing cavities, which further worsen structural damage, and structural damage will, in turn, aggravate infiltration or leakage (Kuliczkowska 2016a, 2016b). These defects, if not were monitored, detected, and treated in time, would further evolve into urban environmental hazards, water losses, and surrounding pavement damage, causing a surcharge (Wu & Sage 2008; Kuliczkowska 2016a, 2016b, Barton et al. 2020). Therefore, periodic inspection and assessment for the municipal pipeline condition and their repair and replacement accordingly are necessary for every country. Furthermore, deterioration models can be developed using continuous and complete historical detection results (Gedam et al. 2016). The remaining life can also be predicted if these continuous inspections are of sufficient years (Malek Mohammadi et al. 2019). Addressing these, real-time monitoring, or non-destructive testing (NDT) techniques are often used.

The subject of this paper is to describe the technology available for monitoring and detecting defects in sewage pipes. Although there already are some great related articles on this topic (such as Makar 1999; Duran et al. 2002; Liu & Kleiner 2013; Moradi et al. 2019; Makris et al. 2020; Bahnsen et al. 2021), the contribution of this paper is to provide an introduction of potential advanced techniques, and a summary of the capabilities and limitations of common NDT techniques for sewers.

The paper is organized as follows: Section 2 describes the approaches used for real-time monitoring of sewage pipes, Section 3 describes the various NDT methods that have been used or could be used to inspect sewage pipes, Section 4 summarizes the capabilities and limitations of each NDT method and discusses future efforts that should be aimed at regarding underground pipeline inspection.

The hydraulic performance of sewer pipes in service can gradually deteriorate or even cause blockages due to the deposition of insoluble fat, oil, and grease (FOG), further resulting in overflows (He et al. 2011).

Sewer monitoring began in the 1970s when water level, flow rate, and rainfall were monitored to predict and prevent overflows in combined sewers (Anderson & Callery 1974). After 2000, monitoring methods and sensors for pollutant concentrations and toxic gas concentrations in sewage pipes have developed considerably (Lynggaard-Jensen 1999; Gruber et al. 2005; Gruber et al. 2006; Tatiparthi et al. 2021), in addition, the development of Internet of Things (IoT) technologies has enabled real-time online monitoring (Stoianov et al. 2008; Liu et al. 2014; Raza & Salam 2020).

Continuous monitoring data of water level, flow rate, pollutant, and toxic gas concentration at manholes can not only provide a prediction of overflow but can also help identify the location of defects in pipelines such as inflow and infiltration, collapse, deposition, and blockage (Bareš et al. 2012; Panasiuk et al. 2015).

One way to achieve structural health monitoring (SHM) of pipelines is geological monitoring along the pipeline using remote sensing, video monitoring, fiber optic sensing technology, and earth pressure sensors. With these sensors, the ground settlement, the soil deformation, and the malicious human action along the pipeline are monitored real-timely (Li & Gao 2017). This paper focuses on an alternative monitoring method, where the target is the pipeline structure.

Vibration-based monitoring（VBM） technology is an SHM method that can determine and identify the type and location of defects, which is implemented via collecting the characteristic signals of the dynamic response of pipelines under the excitation of environmental loads such as traffic loads (Doebling et al. 1998; Toh & Park 2020). VBM was originally proposed for monitoring large-span bridge structures, which requires multiple sensors arranged equidistantly on the structure and thus involves extensive sensor maintenance. This may be one of the reasons why the technology has still not been used for SHM of sewages, and another reason may be the difficulty in identifying and interpreting the characteristic signals due to the complex environmental seismic sources surrounding sewers (Brownjohn et al. 2011). This technology can be also used as a damage identification or inspection method.

Acoustic emission (AE) monitoring technology is similar to vibration-based monitoring methods, whereby defects are identified through interpreting the collected acoustic signals generated by the elastic waves that are emerging when pipeline deformation, or cracks happening (Gholizadeh et al. 2015; Martini et al. 2017; Goldaran & Turer 2020). Commonly used sensors for AE technology are hydrophone arrays, piezoelectric sensors, fiber optic sensors, etc. (Quy & Kim 2021). The disadvantage is similar to VBM techniques, where a large amount of environmental noise needs to be filtered out before saving, identifying, and interpreting anomalous signals.

Distributed fiber optic sensing technology features anti-electromagnetic interference, high sensitivity, lightweight size, and wide monitoring range advantages (Buchmayer et al. 2021). It has a bright future in the operation monitoring of infrastructure including power lines (Chai et al. 2019), bridges (Minardo et al. 2012), dams (Platt et al. 2011), tunnels (Lienhart et al. 2019), and high-speed railway (Filograno et al. 2013).

Unlike traditional sensors that rely on discrete sensors measuring at pre-determined points, distributed sensing does not rely upon manufactured sensors but uses optical fiber. The optical fiber is made of pure glass (silica) as thin as human hair. It consists of two parts: the inner core and the outer cladding. Cladding is a glass layer made up of lower refractive index glass to maintain the guidance of light within the core. Both parts are encapsulated by a single or multiple layers of primary polymer coatings for protection and easiness of handling (Richards et al. 2012). Thus, it is also a cost-effective method that can be easily deployed even in the harshest and most unusual environments.

The optical fiber is the sensing element without any additional transducers in the optical path. The interrogator operates according to a radar-style process: it sends a series of pulses into the fiber and records the return of the naturally occurring scattered signal against time. The parameters such as power, frequency, and phase of these waves can be affected by temperature, stress, and vibration to produce an abnormal signal; therefore, the signal characteristics distributed along the path of optical propagation can be used to identify and locate abnormal structural conditions (Lauro et al. 2014).

The SHM for sewer pipes can be achieved by laying fiber along the new pipeline, but there is not yet a well developed approach for applying this technology to installed pipelines.

Durability monitoring is achieved by monitoring sewer environmental parameters to predict the residual life expectancy of concrete pipes. This is performed based on corrosion models whereby the annual thickness loss of concrete protection cover can be calculated as a function of temperature, relative humidity, and hydrogen sulfide concentration (Wells & Melchers 2014; Wells & Melchers 2015). Surface humidity is derived by measuring conductivity and the correlation between conductivity and humidity (Thiyagarajan et al. 2020), or by measuring the anomalous signal of the fiber when it is stressed by the hygroscopic swelling coat outside (Rente et al. 2021).The effectiveness and limitations of these monitoring methods are compared and discussed in Table 1.

Closed circuit television (CCTV) technology was first applied to pipeline corrosion detection in the mid-1960s (Page 1965), and has been widely used for pipeline inspection worldwide since the 1980s.

As shown in Figure 1, CCTV system is mainly composed of a control platform, crawler, camera, cable tray, and recorder. Vision information recording the defect conditions including rust layer, scaling, corrosion, perforation, cracks within the pipeline can be accessed by controlling the crawler carrying a high-frequency camera into the sewer interior. These images or videos are translated by experienced engineers or well-trained image recognition software into information usable for pipeline condition ratings, according to standards issued by the authorities (Guo et al. 2009; Sarshar et al. 2009; Li et al. 2019; Moradi et al. 2019).

CCTV is unable to detect areas covered with sewage and sludge, so the pipes need to be plugged, extracted, and flushed before the actual testing begins, to ensure that the depth of water and sludge within the pipes is in an acceptable range.

Sewer Scanner and Evaluation Technology (SSET) was originally developed jointly by the Toa Grout, Core Corporation, and the Tokyo Metropolitan Government Sewerage Service in Japan in 1994. It was introduced to the North American market in 1997 as part of a Civil Engineering Research Foundation (CERF) evaluation program (Purdue ECT Team 2007).

Based on a CCTV inspection system, SSET uses a fisheye lens with an optical scanner and gyroscope to automatically pre-process video data as the robot moves from one end of the pipe to the other, displaying the entire internal surface of the pipe as a flat image (Haurum & Moeslund 2020).

Quickview (QV) is a visual inspection technology similar to CCTV technology, mainly composed of an intelligent display controller, transmission cable, camera, telescopic pole, U-probe. It is highly integrated, battery-powered, compact, and lightweight, suitable for field or mobile operating conditions. The disadvantage is that it is time-consuming, and the detection distance is limited (Wang et al. 2021).

Another visual inspection technology is the manhole zoom camera. The internal conditions of the pipeline in a range of about 20–30 m can be photographed using the camera installed at the manhole by adjusting the sight distance (Plihal et al. 2016).

The mentioned above traditional camera-based inspection system requires the operator to control the camera or fix it in a manhole to inspect the inside of pipes. For a long time in the past, the identification and evaluation of defects often require the examination of huge collections of videos, which is not only time-consuming and labor-intensive but the reliability of detection results hinges on the operator's work experience (Moradi et al. 2019). Now with the help of artificial intelligence and image recognition, the efficiency and accuracy of these vision detection techniques have been greatly improved (Haurum & Moeslund 2020).

The waveform and phase of a sound wave change accordingly when encountering the discontinuities producing reflection, refraction, diffraction, and attenuation. Thus, the discontinuity profile information can be derived by analyzing and processing these characteristic signals (Brook 2012). Using a probe comprised of a piezoelectric element capable of deforming and generating high-frequency acoustic waves that travel at a specific velocity dependent on the material. Usually, as shown in Figure 2, the horizontal coordinate represents the time (distance), and the vertical coordinate represents the energy intensity, the defect depth can be determined by the horizontal coordinate distance between wave peaks, and the defect size can be determined by the wave height.

Based on ultrasonic inspection technology, advanced technologies such as phased-array ultrasonic inspection technology and guided-wave ultrasonic inspection technology (long-range ultrasonic inspection) were developed (Alleyne et al. 2001; Hagglund et al. 2012).

As shown in Figure 3, the phased-array ultrasound probes consist of multiple fixed elements that can be individually phased or time-delay controlled to create a beam at a specific angle, thus enabling signal amplification (focusing) or moving focus in the target area. Phased-array ultrasonic inspection can achieve accurate detection of cracks, rust, abrasion. and other structural defects of the pipeline within the target range, but it requires installing the probe on the surface of the pipes, which is not very suitable for buried pipelines (Kim et al. 2013).

Long-range ultrasonic testing (LRUT) also known as guided-wave ultrasonic testing, is an ultrasound method using Lamb waves, a low-frequency wave that propagates along the plates or shells (Barkanov et al. 2018). As shown in Figure 4, Lamb wave is a kind of long-wavelength ultrasonic wave, which will form a waveform passing parallel to the direction of the thin plate after being reflected and interfered with by the upper and lower walls several times in the thin plate. The guided-wave ultrasonic testing allows for full-section and long-distance (up to 200 m) inspection of homogeneous materials such as metal and plastic pipes, requiring only a small excavation to install the ultrasonic transducer (Mirchev et al. 2018). However, the guided-wave transmission distance is greatly reduced by the discontinuity of the concrete material, resulting in a testing distance of only 2–8 m at a time for concrete pipes (Tatarinov et al. 2018).

The basic principle of sonar detection is the same as ultrasonic testing, the distinction being that the former uses the propagation of sound waves in the water to detect the submerged target. It has been widely used for inner surface defect and sedimentation detection in operating pipes.

The laser profiling technique uses laser profing to project a circumferential ring onto the inner wall of the pipeline, which is further recorded by a CCTV system. The three-dimensional (3D) images of the inner surface of the pipeline can be further available by stitching the contour rings of each frame drawn from the video. Accordingly, the pipeline deformation, the wall thickness variation, and the height of sludge can be visualized. Compared to CCTV inspection systems, this technology replaces lighting equipment with laser emitters, while improving the high efficiency and accuracy of inspection, and offers the possibility of quantitative analysis of defects. The disadvantage is that stitching many frames together in the right location is not quite achievable (Duran et al. 2007; Clemens et al. 2015). In practice, visual inspection techniques are the most viable for assessing underground pipes.

Surface 3D information can be obtained by laser triangulation, where a laser is projected onto a surface and then observed with a camera at a specific offset angle to achieve a 3D measurement. Aligning both the laser and camera to the same target area, the angular offset and distance between the laser and camera are configured, thus the depth difference can be calculated.

Both laser mapping and laser 3D scanning are a technique based on laser triangulation. With laser 3D scanning or laser mapping, the dense point information including 3D coordinates, texture, and reflectivity of the measured pipeline surface is measured by the laser rangefinder. The created point clouds with the surface contour information are used to be stitched together to build a 3D pipeline model (Son et al. 2015).

Light detection and ranging (LiDAR) also uses lasers to scan 3D space, but it employs time of flight, wavelength, and projection angle to determine the distance and location of objects in laser click rather than relying on geometric calculations as laser 3D scanning does. The multiplication of millions of points results in a point cloud that can be stitched together into a geometric grid. Performing sewer pipeline and manhole inspections, LiDAR can create a digital twin model of the pipelines. Deformation, surface contours, joint gaps, location offsets, size and location of fracture defects, deposition thickness, etc., can all be determined from this model (Ékes 2017; Ékes 2021).

Stereo vision inspection is another emerging technology that can quantitatively evaluate defects. Usually, depth cameras such as binocular cameras, structured light cameras, or binocular structured light cameras are used as inspection elements to obtain point cloud data for each pixel in the field of view. Like LiDAR and 3D laser scanning, a 3D digital twin model for the pipe surface profile can be built by stitching these point clouds together (Huynh et al. 2016; Zanuttigh et al. 2016; Zhang et al. 2017).

Current leak detection technology, also known as focus electrode leak locator, is a non-destructive testing technology to determine pipeline leakage defects according to the change in impedance of the pipeline leakage parts. Non-metallic pipes or metal pipes with insulated lining feature high impedance at the inner wall, while the impedance of water and soil is low. Therefore, the leakage enables the generation of a current flow in the closed circuit between the probe inside the pipe and the ground electrode. As the probe of the electrical leak detector goes through the broken area where leakage occurs, the current feedback is recorded and analyzed to determine the size and location of defects. This technique does not apply to metallic pipes and requires the pipe to be tested in full flow condition during detection (Soares et al. 2011).

With the mechanical assessment of conduits (MAC), a radial force is applied to the inner wall of the pipe, and the strain within a certain range in the direction of the pipeline axis is recorded, whereby the structural integrity of the pipe and the stability of the surrounding soil can be analyzed. Young's modulus can be inferred from the loading curve when the shape of the pipe section and the thickness of the pipe wall are known, and a quality rating factor can be determined by comparing Young's modulus with the expected modulus. This method is only applicable to the pipe whose diameters are large enough to accommodate the entry of people and equipment (Zhao et al. 2001).

As shown in Figure 5, the magnetic flux leakage sensing technology identifies defects by creating a magnetic field in the pipe wall and then measuring the magnetic flux leakage caused by cracks, corrosion pits, and wall thickness thinning. Usually, the magnetic field is generated by permanent poles at the ends of the detection device, and such fields are disturbed when they pass through defects in the pipe wall, which are measured by Hall-effect devices. In addition, analysis of magnetic flux test results also allows the determination of wall thickness (Tiratsoo 1992; Shi et al. 2015).

This technology is now widely used in petroleum pipelines, where a device called Smart Pipeline Inspection Gauge (pig) is developed. The Smart pig starts at the upstream station and records the defect data along the pipe wall as it moves with fluids during pipeline operation. The pig is centralized in the pipe center by odometer wheels during the entire travel, and the distance traveled is recorded by them as well. Once received at the downstream end, the data is downloaded and analyzed. The location of the pipe corresponding to the defect data is related to the mileage information (Nagaraj 2013).

This technology can only be used to detect metal pipes that can be magnetized. An improved sensor called remote field transformer coupling sensor can be used to detect corrosion defects of reinforcement bars or steel cylinders in prestressed concrete pipes (Mergelas & Kong 2001).

Ground-penetrating radar (GPR) uses high-frequency and usually polarized electromagnetic waves to penetrate the ground, whereby the defects of buried pipes and surrounding soils can be detected by analyzing the echo signal (Lai et al. 2018). The GPR moves along the pipelines and emits electromagnetic waves downwards, the material, size, radius, and contours of the target can be measured according to the collected echo signal discrepancy which is caused by the dielectric permittivity contrast (Chandra et al. 2019). Structural profiles of the pipe and surrounding soil can be generated by connecting measurements from different locations. The significant parameters for GPR are the depth penetration (detection distance) and the resolution. However, this is a contradictory pair of indicators, which means that to increase the resolution, the penetration capacity is inevitably reduced. Therefore, this technology is more applicable to the shallow buried underground pipelines whose dielectric constants are differing distinctly from air and soil (Daniels 2004). In addition, the need to block traffic limits the practical application of GPR in municipal pipeline inspection.

The pipe penetrating radar (PPR) is a kind of GPR used inside the pipelines. With the PPR method, internal structural defects, the wall thickness of pipes, and cavity defects in its surrounding soil can be quantitatively detected (Ekes & Neducza 2012).

Infrared thermography (IRT) is a method based on the measurement of heat gradients caused by the presence of defects in the structure surface, which enables large-area and rapid detection of defects in a non-contact approach. Any object with a surface above absolute zero will emit electromagnetic waves, and the wave parameters including wavelength and intensity will be varied by microstructural defects. As a result, surface defects can then be inspected by analyzing these parameters which are related to the temperature (Speakman & Ward 1998). The heat absorption and radiation vary on the defect area, on the basis of which the active IRT was developed by employing the excitation source such as halogen lamp, infrared lamp, hot blower, ultrasounds, microwaves, eddy currents, lasers (Keo et al. 2015; Szymanik et al. 2016; Ciampa et al. 2018; Jang & An 2018; Ishikawa et al. 2021).

Due to the weak dielectric nature of the concrete material, the ability to absorb and radiate electromagnetic waves has been quite low, so it is still facing challenges using infrared detection technology for concrete pipe inspection. Alternatively, the thermal differences in the soil created by leaking pipes buried underground may propagate to the surface, which is feasible to detect using IRT (Kavi 2018).

Since single data from inspection technology does not fully indicate the condition of the pipeline, in the 1980s, Germany started to explore the application of multiple sensors integrated with a vehicle that travels inside the pipeline to inspect. This led to a prototype of a pipeline inspection robot.

The German pipeline inspection robot KARO carries optical, acoustic, chemical, and electromagnetic detection sensors that allow simultaneous acquisition of pipeline defect information such as pipe diameter, wall deformation, cracks, surrounding soil voids, and water environment information as it moves (Kuntze et al. 1995; Eiswirth et al. 1999). Meanwhile, the Australian Water Authority and Commonwealth Scientific and Industrial Research Organization (CSIRO) collaborated to develop an inspection system that can automatically evaluate defects, pipeline inspection rapid assessment techniques (PIRAT). PIRAT is configured with two semi-independent subsystems: an inspection system consisting of a CCTV system and either a laser scanner (for waterless pipes) or a sonar scanner (for full water pipes) to collect pipe geometry data; and the other system to interpret geometry data to identify, classify and evaluate defect levels (Campbell et al. 1995). In recent years, a variety of advancing robots has been developed (Tur & Garthwaite 2010; Daniyan et al. 2022; Sawarkar 2021; Jain et al. 2022). The applications, advantages and limitations of these methods are compared and discussed in Table 2 

As the service age of sewage pipelines increases, the transport capacity of the pipe will be decayed due to deposition and grease scaling, the concrete protective cover of bars will be diminished due to corrosion, the joint offset and gap might occur disturbed by the deformation of the ground or construction, groundwater infiltration or over-excavation of the surrounding tunneling may cause loss of soil and further form cavities behind the pipe wall. These defects are hidden risks in our cities. To address this problem, various NDT techniques have been developed for identification and assessment.

This paper presents the introduction of the various monitoring methods and NDT techniques available for sewage pipes. The applicability, data available, and limitations of these techniques are summarized in Tables 1 and 2. Some of these techniques, such as CCTV, QV, SSET, zoom camera, are for visual observation of operational defects of pipes including deposition, roots, infiltration, obstacles, or structure defects including crack, fracture, broken, hole, joint offset, joint separate, weld failure. All these techniques are currently widely used. Some others allow quantitative evaluation of the loss of pipe walls, such as PPR and acoustic methods. The soil condition around the pipe wall can be evaluated by GPR, PPR, and sometimes MAC. Of these technologies, GPR is mature but not very popular due to the requirement to intercept traffic, while PPR and MAC are still in their commercialization way and only have a few use cases in Europe and the US. Descriptions for technologies that have potential in the development of digital twin models of pipelines including Laser 3D scanning, LiDAR, and stereo vision are also provided. With the development of technology and commercial competition, it is foreseeable that pipe inspection robots equipped with multiple technologies will be more affordable in the future.

The visual inspection techniques are mature and achieve automatic identification and rating of cracks with the help of artificial intelligence technology. The disadvantage is that they have difficulties in quantitative evaluation and can only detect surfaces. Acoustic and laser-based techniques have shown merit in quantitatively detecting defects such as cracks, delamination or thinning, however, these techniques have limitations in practice, as they take a long time when inspecting long pipelines; the implementation of automation may remedy this deficiency. Current leak detection is only suitable for non-metallic pipes, magnetic leaks are only suitable for magnetizable metallic pipes, their accuracy of defect location and quantitative assessment is poor.

The introduction of stereo vision technology into mature vision inspection technology is promising. The forward direction should be to develop a robot with simultaneous localization and mapping (SLAM) capabilities that can build a digital model rapidly while advancing through the pipeline, that is the named digital twin. This will enable automated identification, quantitative evaluation of pipeline defects, and their real-time presentation by employing a single inspection system.

In addition to the ongoing development of NDT technology, future efforts should focus on wireless communication technology in the underground space, to eliminate the need for robots to drag long cables and enable the deployment of real-time monitoring sensors inside pipes. Also, robust sensors that can bear the harsh environment in sewage pipes are desired.

This research was undertaken, thanks in part to funding from the Beijing major science and technology projects (No. Z191100008019002).

All relevant data are included in the paper or its Supplementary Information.