Leak Detection Algorithm for Pipelines in Noisy Environments

In this paper, two machine learning techniques are applied and compared in order to model leak detection in pipelines in noisy environments. A set of accelerometers, mounted on the surface of the pipes, was deployed for the data acquisition process. Measurements of noise during normal operating conditions were recorded as well as measurements of leaks, generated on various distances from the sensors. Using these measurement data, a training set was created from their time-domain and frequency-domain features. The leak detection process is then modeled as a binary classification problem (leak detection or not). For this problem, two machine learning classification techniques were evaluated, the support vector machines and the decision trees. The results for each learner are compared with the original data from the test dataset using representative performance indicators and, overall, high levels of accuracy are achieved.

Leakage of water in distributing urban waterworks is a matter of concerns for its impacts on environment and daily human life. Water is not unlimited resource especially running and fresh water. The research, included in vast framework, intends to render some outcomes regarding the application of leak detection in pipelines. The kind of pipeline considered in this research is a zigzag one with bends. Leak detection speed and its magnitude are most significant because they allow to save money and time. Because of the zigzag configuration and the small section, it is not easy to do that for this case. This research proposes a comparison between two important techniques used for computing and detecting leakage in pipelines. They are FDM (Filter Diagonalization Method) and impedance method

Digital twins are transforming the digitalization and automation of offshore gas pipeline systems by enabling realtime monitoring, predictive maintenance, and operational efficiency. This study introduces a novel adaptive digital twin framework designed for leak detection and localization in offshore gas pipelines. The framework integrates OLGA-generated synthetic data, validated experimental results, and advanced machine learning (ML) techniques, including transfer learning and ensemble models. The proposed framework achieves a classification accuracy of 98.2% for leak detection, with a mean absolute error (MAE) of 0.11 cm for leak size prediction and a mean absolute percentage error (MAPE) of 3.8% for leak localization. A core innovation of this framework is the calibration methodology, which recalibrates dimensionless nomographs and leak detection correlations for seamless adaptation to new pipeline geometries and operating conditions. Through systematic steps, the calibrated correlations predict leak size and location with high accuracy, leveraging pressure drop and mass flow difference data. Additionally, ML-driven models enable efficient generation of new nomographs for pipelines with varying configurations, enhancing scalability and reducing computational effort. The real-time implementation enables predictions with a latency of less than 2 seconds, significantly outperforming conventional methods in speed and accuracy. Also, the framework’s adaptability, supported by its digital twin visualization and real-time feedback mechanisms, significantly improves pipeline integrity management, operational safety, and environmental protection. The study demonstrates the framework’s robustness in handling complex flow dynamics and offers a scalable solution to enhance the digital transformation of offshore oil and gas operations.

Leak detection is an important issue in piping that deals with the management of water resources; nowadays large amounts of water in the network are dispersed as reported in current scientific literature. Among the methods for leak detection in water pipes, spectral analysis is very interesting. A classical spectral method is fast Fourier transform, but in this paper, we present an alternative method of spectral analysis, which has higher performance in terms of resolution and fast processing, namely decimated signal diagonalization (DSD). It is a nonlinear, parametric method for fitting time domain signals represented in terms of exponentially damped time signals. The aim is to reconstruct the unknown components as the harmonic variables, estimating the fundamental complex frequencies, and amplitudes. The DSD method partly uses the principles of the filter diagonalization method (FDM), which constructs matrices of a generalized eigenvalue problem directly from measured time signals of arbitrary length. However, the DSD because of its windowing technique produces a considerable reduction of size of the original data matrix, and consequently acquisition time can be shorter. We have tested the DSD method for leak detection problem in an experimental zigzag pipeline. We show as the DSD method produces good results in terms of resolution than FDM one.

The complex operational conditions of in-service pipelines and the extended service time make them susceptible to metal loss, which can lead to pipeline failure. However, quantifying the defect depth presents a challenge to current research. The depth quantification accuracy of pipeline defects can be improved, and the defect state of long-distance pipelines can be accurately controlled by building a model based on the internal magnetic flux signal. This study performed a pulling experiment for internal magnetic flux leakage detection in pipelines and extracted the triaxial leakage signals of 60 internal and external pipeline defects. Firstly, Pearson correlation coefficient is used to extract the seven main influencing factors affecting the depth of pipeline leakage defects, namely axial signal peak-valley value, axial signal waveform area, axial signal waveform energy, circumferential signal peak-valley value, radial signal peak-valley value, axial integral signal peak-valley value, and radial integral signal peak-valley value. Secondly, the initial weights and thresholds of the backpropagation neural network (BPNN) were optimized using the African Vulture Optimization Algorithm (AVOA). Furthermore, an AVOA-BPNN-based defect depth prediction model for pipe leakage detection was established to address the problem challenge of the BPNN algorithm easily falling into the local optimum. The magnetic triaxial leakage signals of the internal and external pipeline defects collected via 60 sets of drafting experiments were used as examples, while a particle swarm optimized BPNN (PSO-BPNN), genetic algorithm optimized BPNN (GA-BPNN), and an unoptimized BPNN were used as models for comparative analysis. The results showed that optimizing the BPNN using AVOA substantially improved the prediction accuracy of the BPNN model. The AVOA-BPNN model predicted 12 sets of pipe defect depths with an average relative error (MRE) of only 2.67% and a determination coefficient (R2) of 0.99800, which exceeded the values of the BPNN, PSO-BPNN, and GA-BPNN models. The AVOA-BPNN uses the internal leakage detection signal to predict the depth of pipe defects, supporting effective decision-making regarding pipe inspection and maintenance.

The commonly used cross-correlation technique for leak location requires that the acoustic speed is known and invariable. In practice, the gas leakage-induced acoustic waves propagate along multiple paths including in-pipe gas and pipe wall, and the acoustic waves in different transmission paths exhibit different acoustic speeds and different dispersive behaviors, which bring a great challenge for leak detection and location in the gas pipelines. In this study, based on the vibration theory of cylindrical elastic thin shell, the wavenumber formulae in different transmission paths are derived to predict the acoustic speeds and the acoustical coupling between the in-pipe gas and the pipe wall is analyzed to determine the dominant transmission path. In addition, the velocity dispersions in the dominant transmission path are suppressed by selection of a characteristic frequency band of the gas leakage-induced acoustic waves. The theoretical predictions are verified in the experiment and the results show that the theoretical acoustic speed is slightly larger than the measured acoustic speed. Thus, the theoretical acoustic speed formula is modified considering the effect of the structural loss factor and consequently the location error using the modified acoustic speed is reduced by two times compared to that using the theoretical acoustic speed.

Pipeline leak detection has attracted great research interests for years in the energy industry. Continuous pressure monitoring is one of the most straightforward approaches in leak detection which utilizes pressure point analysis (PPA) algorithms to exploit the transient pressure characteristics and identify leak events. However, a critical issue that jeopardizes the deployment of PPA based methods is the high false alarm rate. In this paper, a novel PPA based leak detection method is proposed which can accurately detect the leak events and dramatically decrease the number of false alarms compared to existing methods. Firstly, the proposed method takes advantage of the good approximation ability and fast learning speed of optimally-pruned extreme learning machine (OPELM) to produce a preliminary leak detection result. Then, the strong memorizing ability of bidirectional long-short term memory (BiLSTM) network is exploited to identify the true positive from the preliminary detection result, hence significantly decrease the number of false alarms. Furthermore, a feature extraction mechanism is also proposed to obtain both the dynamic and static characteristics from raw pressure wave. Experiments and verifications are performed on different real world data sets obtained from pipeline leak tests. It shows that the proposed method can achieve higher detection accuracy with significantly less false alarms. It enhances the practicality of pressure monitoring based leak detection schemes.

Leak detection of oil pipelines is an important issue for safe operation of pipelines, reducing oil loss and environmental pollution. A new approach based on principal component analysis (PCA) method to detect leaks of oil pipelines is proposed in this paper. In order to detect leaks, a classifier is designed to recognize negative pressure wave curve by training set. Results indicate that the method can detect many leak faults from a pressure curve, which might not be effectively detected by a traditional signal processing based method but can be recognized easily by human visual perception.

A Particle Filter Based Leak Detection Technique for Water Distribution Systems

In this study, we propose an innovative approach for real-time leakage detection in pipelines by integrating conservative Hamiltonian equations and experimental Internet of Things (IoT) technologies. The proposed method combines a hybrid model that utilizes sensors and IoT devices to acquire real-time data and solves the coupled system of Hamiltonian equations using the ODE45 numerical integration method. Spectral frequency analysis is an essential part of this method, as it reveals specific patterns in the pressure and flow signals. The findings highlighted 95% accuracy in leak detection, which was validated through a comparison of the theoretical and experimental data. The novelty of this approach lies in its ability to maintain constant total system energy, thereby enabling continuous monitoring for early leak detection. As an improvement, the proper handling of sensor signals is emphasized, underscoring its contribution to the efficient management of water resources in potable water distribution systems. Doi: 10.28991/CEJ-2024-010-04-01 Full Text: PDF

In order to further improve monitoring safety of oil and gas pipelines, a leakage detection technology for natural gas pipelines with distributed Fiber Bragg Grating (FBG) is put forward. Optical fiber, set up along pipelines acting as FBG sensors, is used to acquire pressure and vibration signals created by some events such as leakage, mechanical disturbance. Through grating matching and automatic distinguish technique leakage can be detected and located. Moreover, independent component analysis (ICA) technique is studied and used to separate useful signal from noise. Theory analysis and experiments showed this method has good capacity to detect and locate leakage effectively in pipelines.

Distributed Fiber optic sensing system is a unique tool for the evaluation of distributed temperature over several kilometers. It is a powerful diagnostic instrument for the identification and localization of potential problems, such as leakages in pipelines and dykes, hot-spots in high-voltage cables and other events that create temperature anomalies. Such distributed temperature sensing (DTS) systems have the advantage of being relatively easy to deploy over long pipeline sections and have been shown to detect leakages events with good accuracy and reliability. However, when distributed fiber optic sensing systems are deployed in security-critical applications, where availability and reliability are crucial, it is important to continuously verify and assess the correct functioning and reliability of the whole system, including the sensing cables, the measurement system, the data analysis software and the alert transmission. In the past, such testing have been performed periodically by the pipeline personnel, but testing frequencies are typically low, e.g. once per year. The DTS Automated Trip Testing System is a fully independent device that is able to produce a controlled and localized thermal anomaly (hot spot or cold spot) and verify its correct detection. This allows a continuous verification of the DTS system reliability and functionality and a periodic statistical evaluation of the confidence level (proven by experience SIL rating). This paper will present more specifically the development, the functioning and deployment, and its applications of an automated system and method for testing the efficacy and reliability of distributed temperature sensing (DTS) systems, in particular those DTS systems used for pipeline leakage detection.

In the continuation of authors' studies on leak diagnosis in pipelines, a new model-based Leak Detection and Isolation (LDI) algorithm is designed. This system only uses measures of flow and pressure coming from sensors placed at the ends of a pipeline. The present approach is based on a finite nonlinear pipeline model, and extended with variables related to the leak. On this basis, the purpose here is to investigate the use of a so-called algebraic observer to estimate the leak position and its magnitude. The corresponding observer design is thus presented, and its performances are illustrated both with simulation results, and experimental ones, with data taken from a real pipeline prototype.

Independent component analysis (ICA) is a feature extraction technique for blind source separation. Enhanced independent component analysis (EICA), which has enhanced generalization performance, operates in a reduced principal component analysis (PCA) space. SVM is a powerful supervised learning algorithm, which is rooted in statistical learning theory. SVM has demonstrated high generalization capabilities in many pattern recognition problems. In this paper, we integrate EICA with SVM and apply this new method to the leak detection problem in oil pipelines. In features extraction, EICA produces EICA features of the original pressure images. In classification, SVM classified the EICA features as leak or non-leak. The test results based on real data indicate that the method can detect many leak faults from a pressure curve, and reduce the ratio of false and missing alarm than conventional methods.

The article is concerned with detection of leakages in pipelines. The new method of improving diagnostic susceptibility of pipelines has been proposed. The essence of this new method is the acquisition of new diagnostic information (i.e. signals). In contrast to the standard signals (measurement of pressure, flow rate, temperature) used up to day, this information is the measurable signals of weak interactions between objects, resulting from the work of testing objects (correctors) joined to tested object (pipeline). The method has been verified experimentally. On the basis of the results of the experimental research it has been demonstrated that due to these new diagnostic signals it is possible to improve the diagnostic susceptibility of pipelines, and, thus increase the efficiency of leakages detection conducted by means of the methods used hitherto.