Journal of Pipeline Systems Engineering and Practice

Abstract

As a cost-effective production, steel reinforced thermoplastic pipes (S-RTPs) are employed in many industrial fields, such as the marine and petrochemical industries. The design, manufacture, and mechanical properties of S-RTPs are more challenging than those of traditional pipes because of their complicated behaviors and damage mechanisms. A comprehensive review of the design, manufacture, and mechanical behavior of S-RTPs is presented to understand the predictive capabilities, efficiency, and drawbacks of S-RTPs. The limitations of the existing investigations of S-RTPs and the guidance on future research to improve the analysis and design of S-RTPs are also reviewed, discussed, and provided.

Practical Applications

This article is the first systematic review to consolidate the understanding of the mechanical performance and existing problems related to steel reinforced plastic pipes (S-RTPs). An S-RTP is a cost-effective production, and it has been employed in many industrial fields, such as the marine and petrochemical industries. The design, manufacture, and understanding of mechanical properties of S-RTPs are quite complicated, and they are more challenging than those of traditional pipes like carbon steel pipes, due to their complex behaviors and failure process. However, there has been no review to consolidate the understanding of the behavior and problems related to S-RTPs to the best of our knowledge. This paper reviews existing work in design, manufacture, mechanical properties, and some potential nondestructive testing techniques of S-RTPs to understand the predictive capabilities, efficiency, and drawbacks of S-RTPs. The limitations of the existing investigations of S-RTPs and the guidance on future research to improve the analysis and design of S-RTPs are also reviewed, discussed, and provided. People including scientists, engineers, workers, students, and potential purchasers can get the content of interest about different kinds of S-RTPs from this paper.

A pipeline is the most efficient way to transport water from one place to another. Due to aging, corrosion, and external factors, the pipeline is prone to damage, which causes leaks. Many machine learning (ML) and deep learning (DL) methods are available to address this issue. This paper does an experimental study on available methods in ML and DL for leak detection for the collected data using an acousto-optic sensor. The experimental setup comprises of an acousto-optic sensor made of an erbium-doped fiber laser (EDFL), galvanized iron pipeline, a tank, a pump, and a data acquisition unit. The dimensions of the galvanized pipeline looped with the water tank are a length of 40 m, an inner diameter of 89 mm, and an outer diameter of 90 mm. The diameter of the simulated leak aperture is 5 mm. The methods analyzed in this study are k-means, k-medoids, Naive Bayes (NB), support vector machines (SVM), k-nearest neighbor (KNN), decision tree (DT), categorical boosting (CatBoost), random forest (RF), XGBoost, AdaBoost, and one-dimensional convolutional neural network (1DCNN). ML algorithms need a feature extraction technique because the data collected from the experiment is too large and contains redundant information. Feature extraction reduces the data size by extracting essential information. This paper extracts ten features from raw data. Among the ML algorithms, AdaBoost gives the highest prediction accuracy of 98.02%. This paper also implements eight models of 1DCNN, and Model 1 shows the best prediction accuracy of 98.16%, which is the highest compared with all the other classifiers in ML and DL for one-dimensional time series acousto-optic sensor data.

Batch interface tracking and contamination control are the key technologies for the operation and management of multiproduct pipelines. Existing studies only focused on sequence arrangement or segment stoppage to control the generated contamination, and few considered this problem comprehensively and systematically. This study fully considers the contamination control process and develops a scheduling model for straight pipelines to minimize the cost of contamination loss caused by segment stoppage. The objective of this model is to minimize the restart cost of idle segments and the penalty cost due to improper interface placement during segment stoppage. Three types of constraints for contamination control are proposed, namely operation control, flow rate control, and stoppage control. Three real-world multiproduct pipelines are used as examples to validate the proposed method. Compared with the actual operation schemes, the method helps to reduce the number of stoppage operations by 20%–50%, and decrease the probability that the interface is improperly located by 40%–67%. Therefore, this work can reduce the cost of contamination treatment and bring greater benefits to pipeline operators.

In practical engineering, it is common to offset the drainage system axis to avoid conflicts with beams and columns. However, such an offset could significantly affect the system’s discharge capacity and drainage safety, leading to a prominent positive pressure above the offset floor. To address this issue, we propose optimizing the system by changing the elbow shapes at the offset position. In our study, we conducted comparative experiments on three elbow shapes, namely, a normal 90° elbow, double 45° elbow, and large curvature elbow, in two experimental systems. Pressure fluctuation and water seal losses were used as experimental parameters to determine the maximum discharge capacity in accordance with the Standard for Capacity Test of Vertical Pipe of the Domestic Residential Drainage System. Our experimental results show that replacing the 90° elbow with the double 45° elbow slightly reduces positive pressure. On the other hand, using the large curvature elbow effectively increases the drainage flow and reduces the positive pressure caused by the upper floor bias. Furthermore, we employed computational fluid dynamics (CFD) software to simulate the flow of water and air in the system, demonstrating the benefits of increasing the elbow at the connecting position.

Abstract

Accurate prediction of tensile strain capacity (TSC) is the key to the strain-based design and assessment (SBDA) of buried pipelines. Various TSC prediction models proposed by research institutions such as the Pipeline Research Council International (PRCI) and the Center for Reliable Energy Systems (CRES) are not applicable to pipelines with corrosion defects. In view of this, this study proposed a new TSC prediction model for corroded pipes based on extensive parametric finite-element analysis (FEA) validated by full-size pipe (FSP) tensile tests. In the proposed model, the corrosion defects were assumed to be a curved boxed shape. Based on the dependence of the TSC on the input parameters, the model considered the combined effect of material properties and internal pressure on TSC and employed a defect geometry factor function based on the corrosion defect depth, width, and length. The accuracy and reliability of the proposed method were validated by extensive parametric FEA and error analysis. The results showed that the method can be applied to the practical safety assessment of corroded pipelines in strain-based design guidelines.

Practical Applications

Assessing the safety of pipelines by axial strain can solve the problem that it is difficult to monitor the complex strain state of the pipe body under ground movements. Tensile strain capacity is an element to assess the tensile resistance of buried pipes to fracture. This study developed a tensile strain capacity prediction model for corroded pipes, which is faster to calculate compared to existing evaluation methods. When corrosion anomalies are discovered in areas with ground movement hazards, the proposed model can be used to quickly estimate the tensile strain capacity of the pipeline to make contingency measures. When the evaluated corroded pipe still has sufficient serviceability, the strain monitoring equipment can be used to monitor the remote strain of the pipeline in real time, which needs to be guaranteed to be less than the tensile strain capacity to fully utilize the pipeline performance. The work in this study can provide a reference for strain-based assessment and design of corroded pipes.

With the deterioration growth of water pipes, the frequency of identifying segments and unintended isolation in water distribution networks is expected to increase, because it can help identify the number of nodes that cannot be supplied and their locations when a failure occurs. Therefore, a methodology for identifying segments and unintended isolation was developed and verified using examples. In addition, a case study was conducted to confirm that the proposed methodology could be applied to actual fields. The proposed method only requires the node-based adjacency matrix $\mathbf{K}$ for identification, where the connectivity of the current network can be intuitively understood with the constructed matrices. Moreover, this methodology is easy to adopt using the single row-first search method. The proposed framework is compatible with the data storage format of hydraulic analysis software. Finally, the proposed methodology can be used in various research on configuring and analyzing water distribution networks.

The erosion of a gas well relief line elbow under supercritical conditions must be investigated, because under these conditions, it may be pierced within a few minutes, which may significantly affect control safety and cause casualties and environmental pollution. Herein, the Eulerian-Lagrangian method was used to establish a numerical model to examine a gas well relief line elbow erosion under supercritical conditions. The numerical model was verified by combining the unit erosion experiments and field failure cases. On this basis, we conducted a simulation analysis of elbow erosion under supercritical conditions according to actual blowout conditions. Nine influencing factors, including the angle of the elbow and discharge volume, were considered. The erosion mechanisms and laws of the elbow under gas–solid two-phase flow, liquid–solid two-phase flow, and gas–liquid–solid three-phase flow were also determined. The erosion laws of the elbow under gas–solid and gas–liquid–solid flow conditions were similar, with severe erosion occurring in the extrados of the elbow. The maximum erosion rate increases with increasing sand content, particle shape coefficient, and temperature, and it decreases with increasing outlet length. Moreover, the maximum erosion rate first increases and then decreases with increasing particle size before finally stabilizing with increasing discharge amount. The erosion rate of the gas–solid flow first increases and then decreases with increasing elbow angle. In contrast, severe erosion occurs on the side wall of the bend for liquid–solid flow conditions. The maximum erosion rate increases with increasing velocity and sand content and decreases with increasing elbow angle and particle shape coefficient. Moreover, it first decreases and then increases with increasing particle size, and it is barely affected by the outlet length and temperature. This study provides key theoretical support for elbow selection and structural optimization.

Vehicle load is the most frequent live load on cross-highway pipelines. Cross-highway pipelines will bear great stress under the rolling action of heavy vehicles, which leads to pipeline failure and damage. However, the current vehicle–road model and the pipeline–soil model are calculated independently of each other, which cannot accurately solve the stress response of buried pipelines in the process of a vehicle rolling in real time. This paper solves this problem. To study the real-time influence of vehicle load, vehicle speed, rolling position, and other dynamic factors on pipelines’ stress, the ultimate bearing capacity of small-diameter gas pipelines under the action of a heavy vehicle rolling was quantified. This paper established the vehicle–pavement–pipeline model by combining Adams version 2020 and Abaqus version 2020, and the pipeline’s dynamic stress response and ultimate bearing capacity under vehicle load were simulated numerically. According to the actual working conditions on-site, an experimental system for buried pipelines was established, and the test was verified with a 7-m-long pipeline. Simulated and experimental results were compared, and showed good agreement. The research results show that the stress on the pipeline is positively related to the vehicle load and diameter. However, with the increase in vehicle speed, the stress on the pipeline tends to decrease. When the burial depth of the pipeline reaches 1.5 m, the protective effect of soil on the pipeline is weakened. The study also found that the internal pressure has a certain protective effect which can reduce the impact of vehicle load on stress increment. This is a new discovery: the closer the rolling position is to the centerline of the pipeline, the greater is the peak stress. Finally, an evaluation method of the ultimate bearing capacity of buried pipelines and the overload warning index is proposed based on the von Mises stress, and vehicle overload warning curves were established. The vehicle overload warning curves can be used to determine whether the current vehicle load will be harmful to the pipeline, or to design the maximum vehicle load of a section in preliminary work. On-site application showed that the vehicle overload warning curves have high accuracy and can provide some guidance for the safe operation of the pipeline.

In the long-distance pigging operation of natural gas pipelines, severe wear of the sealing cups always leads to the failure and instability of the pig. Different from ordinary conditions, the wear process of the polyurethane rubber (PUR) used in the pigging operation occurs under the condition of interference fit, which generates a continuously varied wear rate. In this research, a custom-built wear experiment was designed and conducted by adopting a specially made indenter and specimen, and then the effect of contact pressure on the wear rate, i.e., the empirical expression of wear rate, was obtained based on the test results and Hertz contact theory. In addition, a numerical model that presents the PUR interference wear behaviors was established by combining the empirical expression of the wear rate and the analytical expression of the hyperelastic materials under the interference fitting. Through MATLAB programming with the iterative algorithm, the wear laws of two kinds of sealing cups were investigated, and the results indicate a coupling induction among the wear rate, wear loss, and contact pressure. The variation of wear loss with traveling distance was proven to agree with the exponential function, and thus a prediction model suitable for engineering practice was built to calculate the wear loss of the sealing cups.

Concrete pipes are used widely in sewage pipeline networks due to their superior stiffness, bearing capacity, and low price. However, as the service age increases, the microorganisms inside the pipeline react with the concrete pipe walls and induce concrete pipe wall corrosion. Microbiologically induced corrosion (MIC) is serious corrosion in concrete sewage pipe walls, resulting in a reduction of the wall’s thickness and causing the cover soil above the buried pipeline to collapse. The real-time corrosion in concrete sewage pipe walls was simulated in this study. A numerical simulation of the MIC in concrete sewage pipes was performed using the software COMSOL Multiphysics, in which the randomness of the MIC was considered by introducing the random distribution of concrete porosity and corrosive substance concentration; the influence of the turbulence and the transfer rate of ${\mathrm{H}}_2\mathrm{S}$ were considered by zoning the section of the pipe wall. Combined with the probability density evolution theory, a probability model is proposed to predict the maximum corrosion depth of the concrete sewage pipe wall. The results show that the maximum corrosion depth in the pipeline is more likely to occur in the vicinity of the sewage level and the pipe crown, and its dispersion increases with time and decreases as corrosive substance concentration increases. After verification, the model presented can be used to predict the time-dependent reliability and the service life of concrete sewage pipes.

Casing deformation often occurs in the process of shale gas volume fracturing, affecting the efficient development of shale gas. According to the actual deformation morphological characteristics of the casing and the mechanism of shear deformation, the three-dimensional physical model and finite element model of casing-cement sheath-formation under the interaction of multifault slip were established. The influence rules of sensitive factors such as fault slip distance, fault angle, elastic modulus of cement sheath, cement sheath defect, and casing eccentricity on casing deformation were studied. The prevention and control measures to reduce casing shear deformation were put forward. The results show that (1) the diameter shrinkage of the casing is more obvious when the relative slip distance between two faults exceeds 60 mm; (2) when the fault angle θ is from 60° to 75°, the shear deformation extent of the casing is large. Before fracturing construction, fault identification technologies should be used to detect the fault distribution in shale reservoirs and effectively prevent and control the risk points of casing deformation; (3) ultra-flexible cement sheath with elastic modulus less than 1 GPa can effectively reduce the shear deformation of the casing; (4) lack width of cement has a greater effect on reducing casing deformation than lack angle. Not sealing the well sections near the fault and leaving a certain buffer area for fault slip can delay or reduce the shear deformation failure of the casing; and (5) casing eccentricity makes casing more prone to failure. Therefore, it is necessary to ensure well-cementing quality.

Abstract

A single-phase incompressible bicomponent transient oil mixing model was established based on the mechanism of flow, heat, and mass transfer of refined oils to obtain the flow field and concentration distribution of the component at the mixed segment during the batch transportation process. A new model for the axial diffusion coefficient was proposed, which included the turbulent diffusion coefficient and the coefficients derived from the differences in fluid physical properties. The new diffusion coefficient calculation formula presented in this paper could make the error very small when the calculated concentration distribution value of the mixing section is compared with the experimental value. Based on the many factors that affect the flow features during the oil replacement process in the mixed-oil section, the main dimensionless parameters (Reynolds, Peclet, Prandtl, Richardson, Schmidt) are selected by using the dimensional analysis method. The mixed segment had a high $Pr$ and $Pe$. With an increasing concentration of lighter oil, the $Pr$ decreased while the $Pe$ and $Re$ increased. The $Ri$ was always less than 1, which implied that the buoyancy is not important, and the gravity effect can be ignored during the generation process of oil mixing according to the definition of $Ri$ that buoyancy is dominant when it is greater than 1. The change of $Sc$ versus the concentration of gasoline was small, and $Sc$ was always close to 100, which indicated that the momentum diffusion was greater than the mass diffusion in the process of oil mixing. Sensitivity analysis was carried out based on the tailing amount and tailing length of the mixed section. It was concluded that the influence of pipe diameter, flow rate, temperature, and diffusion coefficient on the tailing amount and tailing length had the same trend.

Practical Applications

A single-phase incompressible two-component transient oil-mixing model is established to obtain the flow field and component concentration distribution in the mixing section during batch transportation in this paper. The flow characteristics and tailing phenomena of the mixed-oil section are determined by studying the dimensionless numbers (Reynolds, Peclet, Prandtl, Richardson, Schmidt) in the oil change process of the mixed-oil section. The experimental results could predict the distribution of oil mixture concentration under specific working conditions. The influences of the difference of oil physical properties, turbulence effect as well as the adsorption effect of the pipe wall on the distribution of oil mixture concentration have also been discussed. The research results show that this technology provides the location and concentration distribution of the mixed oil segment when it is applied to practical cases, which gives technical support for detecting and cutting the mixed oil segment during the process of batch transportation and also guides the accurate cutting of the mixing segment.

To further accurately analyze the mechanical response of buried pipelines and pavement structures subjected to ground subsidence, this study established a dual-beam model consisting of an upper Euler–Bernoulli beam and a lower Euler–Bernoulli beam to simulate pavement structures and pipelines. The embankment fill layer and foundation soil were idealized as in the Winkler foundation model, which was divided into three zones. Coupling differential equations for the behavior of this dual-beam–foundation system were established according to the force equilibrium among the pavement structure, buried pipeline, embankment fill, and foundation soil with voids. The stress distribution and deformation of buried pipelines and pavement structures were solved according to the boundary and geometric conditions. The proposed mechanical model was evaluated on test data and corresponding finite element codes, concretely verifying the model’s effectiveness, and a parametric study was proposed to investigate the behavior of this system, such as the bending stiffness of pipelines, traffic load, the strength of fill layers, and foundation soil.

Leaks cause significant water loss in underground water distribution networks, which makes it critical that utilities quickly detect, localize, and repair them. Acoustic leak detection and localization methods using hydrophones and accelerometers are the most studied technology; however, most studies for localizing leaks have focused on simple straight pipe segments using the cross-correlation technique. Leak localization in a network of pipes is significantly more challenging, and this problem remains largely unexplored in the literature. The difficulty arises because the cross-correlation between two acoustic sensor measurements yields multiple time delays corresponding to multiple paths between the acoustic source and the sensors in a network. Hence, the problem of localizing the leak correctly requires taking such multiple paths into account. In this paper, we propose a new method for localizing leaks in a network of pipes. Our method operates on multiple time difference of arrival (TDOA) by calculating the cross-correlation of the signals from different pairs of hydrophone sensors. A conditional probability distribution function is calculated corresponding to each TDOA, and the leak location is found based on the principle of maximum likelihood estimation. We also formally propose a new term called interior points where we define the conditions in which leaks can be pinpointed or only localized to the closest pipe joint. Using simulation studies, the proposed method is shown to accurately pinpoint leaks for the cases when the simulated leak satisfies appropriate conditions. The method is also validated by conducting experiments on a laboratory test bed where a simulated leak is pinpointed to within 10 cm of the actual leak location.

Due to the mismatch between gas production and consumption in the gas delivery system (GDS), reduced customer demand can easily lead to overpressure and gas release. Companies mainly use high-pressure release (HPR) to lower the GDS pressure, which results in significant energy losses for compression and gas separation. This paper (1) proposes a low-pressure release method (LPR) for users’ low-demand scenarios to avoid regulating air separation units, and (2) provides a way for obtaining optimal parameters for LPR when used alone and together with HPR. LPR released low-pressure gas by lowing the compressor inlet guide vane opening (IGA) or the inlet valve opening to reduce the compressor load, outlet pressure, and GDS pressure, and reduce the energy consumption of the compressor. Case studies showed LPR reduced effectively the gas emissions and compressor operating energy. When LPR is used alone, the maximum reduction in gas emissions and compressor operating energy consumption compared with HPR was $\mathrm{6,049}{\mathrm{Nm}}^3$ and $900\mathrm{kW}·\mathrm{h}$, accounting for 62.96% and 2.18%, respectively, and the gas emission rate was reduced from 0.11% to 0.04%. When LPR was used together with HPR, 43.88% oxygen emissions and 0.28% total compressor energy consumption were reduced, and the total conditioning time can be effectively reduced from 42 to 10 min. The method increases the ability of the compressor to regulate the GDS.

Ground penetrating radar (GPR) is widely used in subsurface utility mapping. It is a nondestructive tool that has gained popularity in supporting underground drilling projects such as horizontal directional drilling (HDD). Even with the benefits including equipment portability, low cost, and high versatility in locating underground objects, GPR has a drawback of the time spent and expertise needed in data interpretation. Recent researchers have shown success in utilizing machine learning (ML) algorithms in GPR images for the automatic detection of underground objects. However, due to the lack of availability of labeled GPR datasets, most of these algorithms used synthetic data. This study presents the application of the state-of-the-art You Only Look Once (YOLO) v5 algorithm to detect underground objects using GPR images. A GPR dataset was prepared by collecting GPR images in a laboratory setup. For this purpose, a commercially available 2GHz high-frequency GPR antenna was used, and a dataset was collected with images of metal and PVC pipes, air and water voids, and boulders. The YOLOv5 algorithm was trained with a dataset that successfully detected and classified underground objects to their respective classes.

Corrosion perforation of wet gas pipelines occurs frequently, threatening the safe production on gas fields and bringing huge economic losses. Therefore, wet gas pipelines need to use a suitable corrosion prediction model to predict the corrosion rate along the pipeline and anticipate the corrosion risk. However, current corrosion prediction models have problems such as large errors and inaccurate predictions. This paper compares four common corrosion prediction models and combines these four models with OLGA multiphase flow simulation calculations to predict the corrosion condition of the same pipeline under three different conditions. Finally, it compares the model predictions with the actual corrosion internal detection data. The predicted value of the De Waard 95 model was closer to the corrosion internal detection data, but the error result still reached 82.03%. To further improve the accuracy of the corrosion prediction model for wet gas pipelines, the De Waard 95 model was optimized using the linear fitting method to obtain a new corrosion prediction model, W22. And the corrosion prediction experiment was conducted for the multiphase flow corrosion loop using the W22 model, and the error between the prediction results and the experimental results was only 6.8%. Compared with the De Waard 95 model, the average error of the W22 model was reduced by 81.32%. Therefore, the W22 model has high accuracy in predicting wet gas pipeline corrosion, which lays the foundation for improving the essential safety of wet gas pipeline operation.

Seismic-induced liquefaction is one of the main hazards for pipelines buried in saturated granular materials. When soil is partially or completely fluidized, a lifeline, although installed in superficial trenches in which the coarse backfill soil usually is compacted, may experience a sudden uplift and damages. To reduce pipeline uplift and thus limit the associated risks, the authors propose a sustainable and original mitigation strategy, suitable for both existing and new lifelines, based on both the use of a geomembrane and the compaction of the soil surrounding the pipeline. According to the design method proposed, the intervention geometry is selected on the basis of the pipeline maximum admissible displacement, whereas the minimum required relative density can be designed, based on the site-specific seismic demand, to avoid cyclically induced local accumulation in excess pore-water pressure. To prove the effectiveness of this strategy, a series of $1\text{-}g$ small-scale laboratory tests was performed on a pipe buried in a fluidized sand layer. A simplified displacement-based design approach, which was validated against the experimental data, is proposed.

Many underground metal drainage pipes start deteriorating by developing small perforation holes throughout their invert wall sections. Once the holes open up, drainage flow can enter into them and begin eroding the subsoil layer. This can lead to a partial loss of soil support under the pipe invert. In this technical note, the approach established by Spangler in his classic analysis of buried flexible metal pipes was employed to update the modified Iowa formula for buried pipes having a partial loss of soil support in the invert region due to perforation holes in the pipe wall. The new equations provide interesting insights into aspects of the pipe-soil interaction problem that is not often reported by other researchers. Furthermore, limited 2D simulations with a computer program CANDE-2007 appear to support the theoretical work carried out.