Optimization of Enhanced Pipeline Leak Detection Using Advanced Dense Hebbian Neural Network Architectures for Improved Accuracy and Efficiency

The key to leak detection and location in water supply pipelines is signal denoising and feature extraction. First, in this paper, an improved spline-local mean decomposition (ISLMD) is proposed to eliminate noise interference. Based on the ISLMD decomposition of a signal, the cross-correlation function between the reference signal and the product functions component can be obtained. And then the PF component containing the leak information can be extracted reasonably. Compared with improved local mean decomposition, the ISLMD has higher accuracy in leak location. Second, an image recognition method using a convolutional neural network for leak detection is proposed, which can better address the problem that the features of different leak apertures or locations are highly similar to each other. The images from the conversion of the reconstructed signals are used as the input of the AlexNet model, which is capable of adaptive extraction of leak signal features. The trained AlexNet model can effectively detect different leak apertures. Finally, the signal time-delay between the upstream and downstream pressure transmitters caused by the leak and propagation of negative pressure wave is determined according to generalized cross-correlation analysis, and thereby, the leak location is obtained. The experimental results show that the proposed method is effective for leak detection and location.

Single-hidden-layer feed-forward quantum neural network based on Grover learning

IntroductionThe application of artificial intelligence, particularly neural network models, has expanded throughout various fields of academic and industrial studies, recently. This research focuses to investigate aeroelastic phenomena and predict flutter speeds without expensive computational and experimental methods utilizing these algorithms, specifically the artificial neural network (ANN).MethodsFundamentally, the approach involves an ANN algorithm to navigate the complexities of aeroelastic systems, achieved by layering multiple ANNs. The network's neurons effectively interpret diverse flight data by this structure. The study also incorporates state-space representation and Theodorsen's unsteady aerodynamic theory to create a comprehensive dataset on aeroelastic flutter speed across different wing parameters. In order to forecast the flutter speeds, a feed-forward neural network (FFNN) model, which is an approach of ANN method, containing sigmoid hidden neurons and a linear output neuron has been proposed for the conditions above.ResultsThe proposed model accomplished a regression value of 0.92 for flutter speeds according to the experimental findings. As presented, the suggested FFNN model is successful and can be employed to predict the flutter speed estimation. The findings of the FFNN structure are validated with the previous work for aeroelastic flutter speed analysis in the literature.DiscussionFinally, the findings indicate that the FFNN model formulated in this study is highly accurate in predicting flutter speeds with the accordance of numerical results of aeroelastic structure modeled with unsteady aerodynamic theory. In addition, the correctness of the predicted flutter speeds demonstrates that analyzing without expensive tests and high computational costs is in the near future.

Prompt and reliable detection of pipeline leaks is vital for human safety, the economy, the environment, and corporate reputation. However, a simple mechanistic model for accurately predicting leak characteristics in different flow regimes is lacking. To fill this gap, a novel methodology was used to develop a multiphase flow leak detection model using only inlet and outlet parameters. The gas-liquid two-phase leak mass flow rate, location, and size are computed through iterative processes in the upstream and downstream sections of the leak. Data sets were generated for a wide range of geometric (3 – 5 inch pipe diameter, 2000 – 10000 feet pipe length, 500 – 1500 feet leak location, 0.2 – 3 inch leak opening diameter), hydrodynamic (Newtonian, air, CO2, N2) and operating conditions (0.3 – 0.628 outlet liquid fraction). These data sets were utilized to develop contour plots and a data-driven model using statistical analysis based solely on the inlet and outlet parameters. The results indicate that a change in total mass flow rate and pressure in the inlet and the outlet section of the leak can be a good indicator for determining the location and size of the leak. The effect of different pressure constraints, pipe length, pipe diameter, two-phase fluid rheology, leak diameter, leak location, outlet liquid volume fraction, and flowing liquid hold-up on leak size, pressure, and flow rate is analyzed. Decreasing the liquid fraction in the outlet section of the leak leads to a slight increase (6% average) in the inlet mass flow rate and a significant decrease (50% average) in the outlet mass flow rate for fixed pressure constraints, resulting in an increased leak flow rate, pressure, and density. Similarly, longer pipe lengths, bigger pipe diameters, heavier gas phase, and lower liquid fraction at the outlet have higher leak pressure for the same leak locations due to higher leak flow rate. Furthermore, contour plots revealed that identifying a leak near the pipe inlet is easier, although determining its size remains challenging. On the other hand, detecting a leak near the pipe outlet is more difficult, but assessing its size is comparatively easier. The developed data-driven showed good agreement with different literature data sets with a MAPE of less than 20%. The mechanistic model's key advantage lies in its reliance on fundamental equations and physics, making it applicable to various operating conditions for field applications. Moreover, the data-driven model is straightforward and accurate, eliminating the need for complex simulations. This study has the potential to assist industries in determining leak location, size, and pressure using only the inlet and outlet parameters, without requiring multiple sensors along pipelines.

Pipelines are a safe energy-efficient mode of transporting fluids and gases. However, risks of pipe leaks still exist. Various pipe leak detection methods are in use; some traditional methods rely on acoustic receivers that must be located on the pipe, thereby restricting their implementation. Microseismic monitoring resolves this restriction because the ground rather than the pipe is monitored. However, microseismic monitoring has, thus far, only been applied using 2D methods for pipeline leak detection, and its accuracy is limited. To address this problem, a new 3D leak detection and location method for buried pipelines is proposed. First, using a 3D staggered-grid finite-difference method for acoustic equations, we simulated a microseismic wave field generated by a leaking pipeline, and recorded it with a ground surface geophone array. Then, a generalized cross-correlation technique was used to extract the time difference. Finally, the simulated annealing method was used to obtain the leak location in 3D space. Simulation and field data results show that the proposed method can locate the leakage point and has good anti-noise capability. Accurate 3D leak location requires at least two survey lines. Non-parallel survey lines can locate the approximate plane position of the leakage. Parallel line and orthogonal line pairs can be used to obtain the leak position accurately; the parallel line pair is the most accurate. The proposed method was concluded to be effective and accurate at detecting and locating leakages in buried pipeline networks.

Pipeline leak detection has always been a relatively difficult technical problem; especially in urban pipeline leak detection, there are still many problems to be solved. A multipoint leak detection and location method for urban gas pipelines based on variational mode decomposition and relative entropy was proposed. Firstly, the experiment pipeline system was built, and the original signal was collected by acoustic emission technology; then, a variational model method was used to decompose the signal to obtain multiple intrinsic mode function (IMF) components with different characteristic scales. According to the characteristics of relative entropy, each component was analyzed, the appropriate IMF component was selected, and the selected component was reconstructed to obtain the observation signal. The multipoint leakage location model of the urban gas pipeline was established. The number of source signals was estimated by singular value decomposition, and the leakage signals were separated; finally, the accurate location of leakage point was achieved by cross-correlation positioning. The results showed that the average relative error of the pipeline leak location results is 2.97%, and the leak location accuracy is significantly improved, achieving the purpose of precise location.

Fatigue crack growth and deformation in weldments of aluminium alloys Hermann, R. Mater. Sci. Eng. A (15 Oct. 1994) 186 (1–2), 45–53

Statistical Approach to Neural Network Model Building for Gentamicin Peak Predictions

There is an increasing need for timely pipeline leak detection and localization methods, pipeline leak could lead to not only the loss of the goods but also considerable environmental and economic problems. With the rapid development of hardware and software, the pipeline leak detection and localization algorithms have been widely researched and applied in many Fields. However, traditional methods are usually limited by extracting features manually, which is inefficient and time-consuming. Convolutional neuron network is an effective method to extract features automatically. In this paper, a novel ensemble transfer learning one-dimension convolutional neural network (TL1DCNN) for the pipeline leak detection and localization is proposed. The TL1DCNN plays the role of base learner. The results of a set of obtained base learners are integrated to achieve the task of pipeline leak detection and localization. Firstly, one-dimension convolutional neural network (1DCNN) models with different parameters are pretrained with source domain data. A small learning rate is set to retrain the above 1DCNN models for target task with target domain data in order to obtain TL1DCNN base learners. Then, the four ensemble strategies with different number base learners whose ensemble weights are optimized by particle swarm optimization algorithm are obtained by minimizing the sum of similarity. The dataset simulated by pipeline network model is used to evaluate the effectiveness of the proposed approach. The indicators such as classification accuracy, precision, recall, F_score and confusion matrix are used to compare the proposed approach with traditional methods and other deep learning methods. The experimental results show that the performance of the proposed approach is superior to other compared methods.

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.

The purpose of this study is to develop a feed-forward neural network (FFNN) model with back-propagation learning algorithm to predict the dissolved oxygen from water temperature and 5 days-biological oxygen demand in the Tigris River, Baghdad-Iraq. The Artificial Neural Networks model was implemented utilizing measured data that were gathered from laboratories of water treatment plant, Baghdad-Iraq, during the year 2008. The correlation analysis between dissolved oxygen and dependent parameters were utilized in selecting the major inputs from water quality parameters for commencing of ANN models. The performance of ANN models were tested utilizing the coefficient of correlation (R), the efficiency coefficient of Nash-Sutcliffe (NS), mean square error (MSE) and mean absolute errors (MAE). The outputs revealed that the feed-forward neural networks using back-propagation learning algorithm which was prepared by temperature and biological oxygen demand offered a relatively high correlation coefficient of 0.885, and efficiency coefficient of 0.782, meanwhile a reasonably low mean square errors of 1.133, and mean absolute errors of 0.369 values for whole array period. The results of the present study demonstrate that the artificial neural networks using FFNN model is capable to forecast the dissolved oxygen values with acceptable accuracy. This is suggesting that the artificial neural network is a useful tool for Tigris River management in Baghdad-Iraq.

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.

This paper presents a Wavelet-based Artificial Neural Network (WANN) approach to model rainfall-runoff process of the Delaney Creek and Payne Creek watersheds with distinct hydro-geomorphological characteristics, located in Florida. Wavelet is utilized to handle the multi-frequency characteristics of the process in daily and monthly time scales. Thus, rainfall and runoff time series were decomposed into several sub-series by various mother wavelets. Due to multiple components obtained through wavelet decomposition, input sets to the Feed Forward Neural Network (FFNN) were enhanced. The application of two information content based criteria (i.e., entropy, H, and mutual information, MI) to select more reliable input sets (among all potential input sets) and to have better insight into the physics of process is considered as the basic innovation of the study which led to a more accurate and compact model. The increase in the number of input of the FFNN might lead to a complex structure and low performance. The results demonstrated that MIÂ Â as a supervised feature extraction criterion could lead to more reliable outcomes due to its non-linear nature. Furthermore, results indicated the superiority of proposed entropy-based WANN model (EWANN) in comparison to simple FFNN. Moreover, multi-step-ahead FFNN, conventional WANN and classic Auto Regressive Integrated Moving Average with eXogenous inputs (ARIMAX) models could not reveal appropriate forecasting results with regard to EWANN model. The superiority of the EWANN over the WANN and FFNN models is not only in terms of efficiency criteria, but also due to its appropriate ability to provide information about the physics of the process. The consequences of EWANN for rainfall-runoff modeling of two watersheds revealed that the proposed EWANN could simulate the process of a small and flat sub-basin slightly reliable than a sloppy and wild watershed. The poor outcome of monthly modeling in regard to daily modeling might be due to involvement of more uncertainty in the monthly data.

This paper presents a methodology for leak detection and location in a great longitude pipeline, which belongs to E.P. PETROECUADOR, through the combination of line balance and speed propagation of the negative pressure wave. Two different approaches, slope projection and a first order dead time (FODT) model, are used for computing the pressure change time, which in combination with speed propagation of the sound is used for estimating leak location. A better performance in leak location accuracy than a commercial installed system is obtained, whose results are compared.

Pipeline leak detection using an acoustic emission (AE) method requires highly sensitive transducers responding to less attenuative and dispersive wave motion in order to place the discrete transducer spacing in an acceptable approach. In this paper, a new piezoelectric transducer geometry made of PZT-5A is introduced to increase the transducer sensitivity to the tangential direction. The finite element analysis of the transducer geometry is modeled in the frequency domain to identify the resonant frequency, targeting 60 kHz, and the loss factor. The numerical results are compared with the electromechanical characterization tests. The transducer response to wave motion generated in different directions is studied using a multiphysics model that couples mechanical and electrical responses of structural and piezoelectric properties. The directional dependence and the sensitivity of the transducer response are identified using the laser-induced load function. The transducer response is compared with a conventional thickness mode AE transducer under simulations and leak localization in a laboratory scale steel pipe.