Pipeline Detection Research Articles

Time series prediction of hydrate dynamics on flow assurance using PCA and Recurrent neural networks with iterative transfer learning

Preventing gas hydrate formation is critical in offshore gas and oil production systems. Several models can predict hydrate formation, however, these empirical approaches have limitations due to dependency on geometries and fluid characteristics of the systems. The trends of hydrate formation or risk are considered statistical, which means there is no definite model to describe its behavior. Herein, we present a novel framework based on a combination of feature reduction methods and several deep learning models to predict the hydrate formation trend through the multivariate sensor data. Transition and segregation trends during hydrate formation were predicted in real-time using sequential time series data from the last 60 s. We employed various deep learning models (Dense, LSTM, GRU, ARLSTM), layers, and dropout to investigate and enhance the prediction ability of each model. Two groups of experimental data (200 rpm, 600 rpm) were used for training and testing the prediction to examine the universal applicability of the model. Transfer learning in training the model was employed to apply the discrete experimental set into time-series data and enhance the accuracy. The results with higher layer numbers and a dropout rate of 0.2 ∼ 0.6 showed the best performance. ARLSTM showed the smallest error among deep learning models and predicted the good trend of kinetic characteristics (transition and segregation part) during the hydrate formation. This approach based on deep learning can be adopted for risk and issue detection of pipelines in the gas production system.

Leakage Detection and Localization of Water Pipeline Using Multi-features and Adaptive Time Delay Estimation

The leakage of water in pipelines severely affects the environment and economy. However, there are limitations in the effectiveness of existing leak detection and localization techniques and methodologies. In this paper, we propose a novel leakage detection and localization method based on the multiple time-frequency features, a neural network, and an adaptive time delay estimation algorithm. First, we use spectral subtraction and wavelet denoising to reduce the effects of noise. In addition, to ensure and improve the accuracy of leakage detection in complex realistic environments, we propose the use of multi time-frequency features that can comprehensively represent the leak signal and make the neural network more robust to train a radial basis function (RBF)neural network to detect the leak signal. Further, we extract multiple features of the leakage signal and input into the RBF neural network to train. Moreover, to prevent the impulsive components of environmental noise and improve localization accuracy, we further propose the use of a fractional lower-order statistics (FLOS) based adaptive time delay estimation algorithm to estimate the time delay and locate the leakage. The simulation results show that the detection and localization performance of the proposed method is superior to those of existing schemes.

An analysis of detectable leakage rate for oil pipelines based on acoustic wave method

Oil pipeline leakage will not only cause economic losses, but also pollute the environment, so the leakage detection of pipelines is very important. The acoustic wave method is widely used in pipeline leak detection, and the leak acoustic signal collected by the acoustic wave sensor often contains a lot of noise, which makes it impossible to accurately determine the inflection point of the signal curve and reduces the accuracy of pipeline leak detection. This paper proposes a denoising algorithm based on mutual information optimization complete ensemble empirical mode decomposition with adaptive noise combined with cross-spectral analysis. Compared with other methods, this method can accurately select the effective intrinsic modal function for signal reconstruction, the denoising effect is more obvious, and the original information is preserved to a greater extent. Acoustic waves are attenuated during the propagation process, and will be affected by factors such as pipe connection ports and elbows, making it impossible to accurately determine the amplitude of acoustic waves around the pipeline. According to the propagation characteristics of acoustic waves and various factors that affect the propagation of acoustic waves, this paper establishes a model for calculating the amplitude of acoustic waves, which can accurately determine the amplitude of acoustic waves everywhere in the pipeline. Finally, according to the model, the relationship between pipeline characteristics and detectable leakage rate is analysed. Field experiments show that the proposed model is accurate and the denoising algorithm is efficient. The minimum detectable leakage rate of the oil pipeline can reach 0.43% when the acoustic wave method is used for leak detection.

Guided Wave Focusing Imaging Detection of Pipelines by Piezoelectric Sensor Array

The sensor array-based guided wave focusing detection method can effectively improve the detection sensitivity of defects and realize the visual imaging. In this research, the three-dimensional finite element simulation model of stainless steel pipeline for guided wave focusing detection was established, and the L(0, 2) mode of guided wave was excited by applying a load on the end surface of the pipeline. And in the experiment, the excitation and reception of L(0, 2) mode guided waves on the outer surface of the stainless steel pipeline were realized by the piezoelectric transducer array. A 16-channel guided wave focusing experimental system was integrated to conduct the defect detection experiments on a stainless steel pipe with diameter of 140 mm and wall thickness of 5 mm. The total matrix data acquisition was performed, and then the amplitude total focusing (TFM) imaging and sign coherence factor (SCF) imaging of the pipe were realized. In this way, the experimental results showed that the pipeline defect detection method and the system proposed in this research can achieve the longitudinal and circumferential positioning and imaging of defects, like holes and scratches.

Valve failure detection of the long-distance district heating pipeline by hydraulic oscillation recognition: A numerical approach

The long-distance district heating pipeline (LDHP) system can deliver waste heat from suburban areas to urban areas. Valve-induced hydraulic oscillation caused by valve failure will lead to large fluctuations of pressure, which may also result in pipeline rupture of the LDHP system. Fast and accurate detection of the faulty valve is a necessary measure for efficient repair and adjustment of the LDHP system. In this paper, the hydraulic transient model of the LDHP system based on the distributed parameter method is combined with the Particle Swarm Optimization algorithm to realize fast and accurate detection of the location and status of faulty valve, by recognizing the faulty valve-induced hydraulic oscillation in the LDHP. The effectiveness of the detection algorithm is verified by numerical simulation of a 20 km LDHP system, and the effects of instrumental error and sampling period on the robustness of the algorithm are investigated. Results show that when the sampling period is less than 10 s, the relative errors of the detected location and opening of the faulty valve are less than 8% and 0.1%, respectively. Therefore, the faulty valve-induced hydraulic oscillation in the LDHP can be effectively detected based on the method proposed in this paper.

Deeppipe: A deep-learning method for anomaly detection of multi-product pipelines

The multi-product pipeline is the main way of refined oil transportation to ensure the safety of the energy supply. Considering that abnormal conditions of the pipeline will cause huge economic losses, personal injury, and environmental pollution, it is important to conduct abnormal detection of multi-product pipelines in time. This work proposes a deep-learning method for multi-product pipeline anomaly detection from the perspective of the temporal and spatial characteristics of the pipeline system. First, analyzing the topological structure of the real pipeline and collecting relevant equipment parameters to establish an SPS simulation pipeline system. By simulating common normal conditions and abnormal conditions in the pipeline, the pressure matrixes of each station are constructed. After being processed by the adaptive padding network and the pairing network, it is input into a customized ranking model to realize the anomaly detection of the multi-product pipeline. Finally, a simulated pipeline and a real pipeline are taken as examples to verify the effectiveness of the proposed method. The results show that the proposed hybrid model has high accuracy, precision, recall, and F1 score of 99.5%, 99.5%, 99.4%, and 99.4%, respectively, which are better than the traditional ranking model. This method can be applied to guide the safe operation and management of on-site multi-product pipelines.

Decision with Uncertain Information: An Application for Leakage Detection in Water Pipelines

Decision with Uncertain Information: An Application for Leakage Detection in Water Pipelines

Simulation and Design of a Balanced-Field Electromagnetic Technique Sensor for Crack Detection in Long-Distance Oil and Gas Pipelines

Due to the extremely small size and arbitrary orientation of the cracks, a highly sensitive sensor based on the balanced-field electromagnetic technique was designed for in-line inspection of oil and gas pipeline cracks. A balanced-field electromagnetic technique sensor mutual inductance model was established and used to theoretically analyze the parameters affecting sensitivity. Finite element simulation was used to analyze the specific effects of the magnetically conductive medium, the number of coil turns, and the sensor lift-off height on the sensor output, respectively, and the sensor parameters of high sensitivity were determined. The detection effect of the sensor on the pipeline crack was tested by the single-sensor experiment and the pulling test. The results show that the designed balanced-field electromagnetic technique sensor is effective in detecting both circumferential and axial cracks of 0.5 to 6 mm in depth. As the crack depth increases, the sensitivity decreases and the detection voltage amplitude increases linearly. The sensitivity of the sensor is highest when detecting circumferential and axial cracks of 1 mm in depth at 1.76 and 0.87 mV/mm, respectively. In addition, the amplitude of the circumferential crack signal at the same depth is approximately twice that of the axial crack signal.

Using Twitter data to understand public perceptions of approved versus off-label use for COVID-19-related medications.

ObjectiveUnderstanding public discourse on emergency use of unproven therapeutics is essential to monitor safe use and combat misinformation. We developed a natural language processing (NLP)-based pipeline to understand public perceptions of and stances on COVID-19-related drugs on Twitter across time.MethodsThis retrospective study included 609,189 US-based tweets between January 29th, 2020 and November 30th, 2021 on four drugs that gained wide public attention during the COVID-19 pandemic: 1) Hydroxychloroquine and Ivermectin, drug therapies with anecdotal evidence; and 2) Molnupiravir and Remdesivir, FDA-approved treatment options for eligible patients. Time-trend analysis was used to understand the popularity and related events. Content and demographic analyses were conducted to explore potential rationales of people’s stances on each drug.ResultsTime-trend analysis revealed that Hydroxychloroquine and Ivermectin received much more discussion than Molnupiravir and Remdesivir, particularly during COVID-19 surges. Hydroxychloroquine and Ivermectin were highly politicized, related to conspiracy theories, hearsay, celebrity effects, etc. The distribution of stance between the two major US political parties was significantly different (p < 0.001); Republicans were much more likely to support Hydroxychloroquine (+55%) and Ivermectin (+30%) than Democrats. People with healthcare backgrounds tended to oppose Hydroxychloroquine (+7%) more than the general population; in contrast, the general population was more likely to support Ivermectin (+14%).ConclusionOur study found that social media users with have different perceptions and stances on off-label versus FDA-authorized drug use across different stages of COVID-19, indicating that health systems, regulatory agencies, and policymakers should design “targeted” strategies to monitor and reduce misinformation for promoting safe drug use. Our analysis pipeline and stance detection models are made public at https://github.com/ningkko/COVID-drug.

Passive detection in water pipelines using ambient noise II: Field experiments

A passive detection method has been proposed in a prior paper to extract key parameters and detect faults using the ambient noise present in water pipeline networks. This paper presents field experiments and data processing results to provide systematic experimental validation of this method. Field experiments were carried out in operational water pipeline networks at the University of Canterbury campus and the Waimakariri District, New Zealand, during which ambient noise was measured by pairs of pressure sensors installed at selected hydrants on pipelines of different materials, network topologies and simulated faults. Auto-correlation and cross-correlation analysis of noise at a single sensor and sensor pairs were carried out to estimate the wave speed and to locate faults in the networks. Data processing results indicate that water usage generating pressure transients are the dominant sources of ambient noise in operational water pipeline networks. This type of ambient noise can also be utilized by the passive detection method to achieve similar wave speed estimation accuracy and fault detection performance as the conventional active pressure wave detection methods.

Application of Alternating Current Stress Measurement Method in the Stress Detection of Long-Distance Oil Pipelines

With the development of pipeline networks, many safety accidents were caused by pipeline stress concentration; it is of great significance to accurately monitor the pipeline stress state for maintaining pipeline safety. In this paper, based on alternating current stress measurement (ACSM) methods, a 3D simulation model of a pipeline electromagnetic field was established by ANSYS software. The distribution law of the pipeline magnetic field and eddy current field were analyzed, and the influence of size and structure parameters of the coil inside the probe were studied. The internal stress detection system of the pipeline was designed, and the static tensile stress measurement experiment was carried out. Simulation and test results showed that the excitation coil with a larger diameter-to-height ratio had a higher measurement sensitivity. The sensitivity of the probe decreased monotonically with the increase of the difference between inner diameter and outer diameter of the detection coil. It increased monotonically with the increase of the equivalent radius of the detection coil. The best measurement results were obtained when the detection coil was located at the center of the two legs of the U-magnetic core. The results showed that the system could identify the pipeline stress concentration area effectively after detection engineering.

IndexPen

In this paper, we introduce IndexPen, a novel interaction technique for text input through two-finger in-air micro-gestures, enabling touch-free, effortless, tracking-based interaction, designed to mirror real-world writing. Our system is based on millimeter-wave radar sensing, and does not require instrumentation on the user. IndexPen can successfully identify 30 distinct gestures, representing the letters A-Z, as well as Space, Backspace, Enter, and a special Activation gesture to prevent unintentional input. Additionally, we include a noise class to differentiate gesture and non-gesture noise. We present our system design, including the radio frequency (RF) processing pipeline, classification model, and real-time detection algorithms. We further demonstrate our proof-of-concept system with data collected over ten days with five participants yielding 95.89% cross-validation accuracy on 31 classes (including noise). Moreover, we explore the learnability and adaptability of our system for real-world text input with 16 participants who are first-time users to IndexPen over five sessions. After each session, the pre-trained model from the previous five-user study is calibrated on the data collected so far for a new user through transfer learning. The F-1 score showed an average increase of 9.14% per session with the calibration, reaching an average of 88.3% on the last session across the 16 users. Meanwhile, we show that the users can type sentences with IndexPen at 86.2% accuracy, measured by string similarity. This work builds a foundation and vision for future interaction interfaces that could be enabled with this paradigm.

An FBG-based high-sensitivity structure and its application in non-intrusive detection of pipeline

Pipeline is one of the main transportation ways with increasingly wide applications in various fields. Generally, the pipelines are operated under high pressure, leakage accidents are extremely dangerous and it is necessary to ensure the safety of pipelines. In this study, a non-intrusive structure with high-sensitivity based on fiber Bragg grating (FBG) sensing is proposed for the detection of pipeline pressure. The core of the structure is a ditrigon strain beam that can transmit the strain on pipeline outside wall to the designed strain beam with an amplification function. Further, theoretical analyses for the structural sensitivity are deduced, and necessary structure optimization is implemented. To verify the practical function and effectiveness of the designed structure, laboratory experiments are conducted and the results indicate that the structure can amplify the strain on pipeline wall with an amplification coefficient of 5.44, and has the obvious advantages of convenient installation and simple-structure.

A semi-supervised leakage detection method driven by multivariate time series for natural gas gathering pipeline

Real-time and accurate leakage detection of natural gas gathering pipelines is critical to the safe and reliable operation of the gas and oil industry. Modern data-driven fault detection and diagnosis techniques have recently gained increasing attention because model-based techniques are practically prohibitive. However, most existing data-driven leakage detection approaches are obtained through supervised learning, which requires a substantial set of labelled data. Especially in real-world scenarios, leakage samples are rare. To reduce the dependence of the leakage detection method on leakage data and make full use of numerous normal datasets generated under normal working conditions, we propose a semi-supervised leakage detection method which consists of two components: an improved long short term memory autoencoder (LSTM-AE) network and one-class support vector machine (OCSVM). The LSTM-AE is first trained to learn the intrinsic features of a normal dataset of pipeline parameters which are multivariate time series. The OCSVM is then applied to calculate the score which is used to infer leak existence. The performance of the proposed method is evaluated on the real natural gas gathering pipelines, and the results confirm that the proposed method achieved 98% accuracy and 99% AUC (Area Under Curve) in a real-life dataset.

Developing and Implementing an AI-Based Leak Detection System in a Long-Distance Gas Pipeline

This research proposes an artificial intelligence (AI) detection model using convolutional neural networks (CNN) to automatically detect gas leaks in a long-distance pipeline. The change of gap pressure is collected when leakage occurs in the pipeline, and thereby the feature of gas leakage is extracted for building the CNN model. The gas leak patterns in the long-distance pipeline are analyzed. A pipeline detection model based on AI technology for automatically monitoring the leaks is proposed by extracting the feature of gas leakage. This model is tested by collecting gas pressure data from an existing natural gas pipeline system starting from Mailiao to Taoyuan in Taiwan. The testing result shows that the reduced model of leak detection can be used to detect the leaks from the upstream and downstream pipelines, and the AI-based pipeline leak detection system can obtain a satisfactory result.

Design of Pipe-inspection Robot Based on YOLOv3

In this paper, a pipe-inspection robot system which uses STM32F103 as control core is designed. The robot system adopts YOLOv3 network model to detect and locate defects of sewage pipeline. The robot uses closed-circuit television system for detection, which is composed of hardware and defect detection. The hardware is mainly composed of robot body, winding coil and upper computer. The motor drive module, body attitude sensor of robot and camera are placed in the pipeline robot. The router, lithium battery and length measurer constitute the winding coil, which is connected with the robot through cable for data transmission and power supply. the upper computer communicates with the cable coil through WIFI, and the staff can send various instructions through the upper computer to control robot operation, and view the status of the pipeline through video. The defect detection of pipeline is mainly realized by YOLOv3 network model, which is used to automatically detect the types of defects in sewage pipeline and mark the location of defects. Finally, the feasibility and effectiveness of defect detection based on YOLOv3 network model are verified by simulation experiments.

A Combined Semi-Supervised Deep Learning Method for Oil Leak Detection in Pipelines Using IIoT at the Edge

Pipelines are integral components for storing and transporting liquid and gaseous petroleum products. Despite being durable structures, ruptures can still occur, resulting not only in financial losses and energy waste but, most importantly, in immeasurable environmental disasters and possibly in human casualties. The objective of the ESTHISIS project is the development of a low-cost and efficient wireless sensor system for the instantaneous detection of leaks in metallic pipeline networks transporting liquid and gaseous petroleum products in a noisy industrial environment. The implemented methodology is based on processing the spectrum of vibration signals appearing in the pipeline walls due to a leakage effect and aims to minimize interference in the piping system. It is intended to use low frequencies to detect and characterize leakage to increase the range of sensors and thus reduce cost. In the current work, the smart sensor system developed for signal acquisition and data analysis is briefly described. For this matter, two leakage detection methodologies are implemented. A 2D-Convolutional Neural Network (CNN) model undertakes supervised classification in spectrograms extracted by the signals acquired by the accelerometers mounted on the pipeline wall. This approach allows us to supplant large-signal datasets with a more memory-efficient alternative to storing static images. Second, Long Short-Term Memory Autoencoders (LSTM AE) are employed, receiving signals from the accelerometers, and providing an unsupervised leakage detection solution.

Аналіз характеристик кумулянтної системи виявлення витоків рідини у трубопроводах

Acoustic noise signals that are generated by fluid leakage due to a leak in the pipeline have in general case non-Gaussian distribution, so for the further development of acoustic contact leak detection together with correlation-spectral methods it is advisable to use methods used for probabilistic analysis of non-Gaussian processes, in particular cumulant methods of noise diagnostics. The principle of cumulant method of liquid leakage detection in pipelines during operation is considered, which is based on the use of cumulants to distinguish acoustic noise signals registered at the control points of the pipeline section, namely on use of the distance between the cumulant vectors of signals and the mutual correlation function of signals. The block diagram of the cumulant system for liquid leakage detection is given. In order to use the cumulant leakage detection system for power plant diagnosis, a layout of the acoustic leak detection system has been developed, which is a hardware-software complex that performs the collection, registration and statistical processing of acoustic noise signals, as a result of which a decision is made on the presence or absence of leakage, the parameters of the hardware part of the system were given. To obtain estimates of the statistical characteristics of real leakage signals, an installation was used that allows to perform physical modeling of fluid leakage in a pipeline in standing water under sustainable conditions, with fixed leak dimensions and pressure level control. Acoustic noise generated by the flow of fluid moving in the heat pipeline is used as interference. Using the statistical characteristics of real signals, the characteristics of the cumulant leakage detection system were determined – leakage signal attenuation; the maximum distance between the transducers at which the leakage is detected; detector sensitivity; the influence of bandpass filtering on these characteristics is investigated. Keywords: acoustic contact leak detection, liquid leakage detector, acoustic noise signals, cumulant analysis

A method for leak detection in buried pipelines based on soil heat and moisture

Leakages in urban heating pipeline networks are mostly caused by underground corrosion, external environment interference and construction. The fluid leaking from pipes accumulates in surrounding soil for a long time, resulting in changes in soil machinery and hydraulic structure, and potential accidents threaten people's daily life and property safety. A preliminary judgment was made to determine whether the pipeline failed according to the infrared images of underground heating pipeline taken from the actual engineering. By comparing the soil moisture around the faulty pipe, we subsequently proposed a fault detection method for buried pipelines which coupled infrared thermography and the variation characteristics of surrounding soil moisture. In order to verify the effectiveness of this method, the numerical simulation was carried out on a leaking buried pipe, and the variation characteristics of soil heat and moisture in the direction of the failure point and its driving mechanism were analyzed. The results showed that the difference between the internal and external pressures at the failure point of the leaking buried pipeline enhanced the effect of soil pressure potential, resulting in significant heat and moisture transfer in soil, which contributed to detecting the buried pipeline failure and identifying the failure types of buried pipelines.

Application of GPR Underground Pipeline Detection Technology in Urban Complex Geological Environments

In the process of the continuous construction of underground pipelines, underground pipe network systems have become increasingly complex, which puts forward higher requirements for normal operation and maintenance. To address different kinds of complex conditions, this experiment in the present paper takes ground penetrating radar as the research basis and uses a self-correction and screening algorithm to innovatively detect underground pipelines. The results show that urban underground pipeline detection technology based on ground penetrating radar (GPR) can obtain a highly reliable number of pipelines and track predefined pipelines when detecting different numbers of verification pipelines. When detecting underground pipelines in different sections, the vertical and horizontal errors are no more than 0.199 m and 0.248 m, respectively, which means that the detection technology of urban underground pipelines based on GPR has high detection accuracy and can be performed on high-level detection tasks under various complex conditions. This research applies bottom detection radar to urban underground pipeline detection technology under complex conditions for the first time, innovatively uses the action mechanism of bottom detection radar, integrates its high precision and high efficiency into underground pipeline detection technology, and ensures the effectiveness of the detection work.