Evaluation of deep learning approaches for oil &amp; gas pipeline leak detection using wireless sensor networks

Spandonidis, Christos; Theodoropoulos, P.; Giannopoulos, Fotis; Galiatsatos, Nektarios; Petsa, Areti

doi:10.1016/j.engappai.2022.104890

Cited by 72 publications

(30 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Based on the Shortcut-ENN model, the pipeline's connection relationship and mechanism characteristics are retained. The reference (Spandonidis et al, 2022) constructed a 2D-Convolutional Neural Network model that undertakes supervised classification in spectrograms extracted by the accelerometers' signals on the pipeline wall. Then, the model was used to immediately detect leaks in metallic piping systems to transport liquid and gaseous petroleum products.…”

Section: Introductionmentioning

confidence: 99%

Deep-learning-based surrogate model for fast and accurate simulation in pipeline transport

Feng¹,

Yan²,

Yang³

et al. 2022

Front. Energy Res.

View full text Add to dashboard Cite

A new deep-learning-based surrogate model is developed and applied for predicting dynamic temperature, pressure, gas rate, oil rate, and water rate with different boundary conditions in pipeline flow. The surrogate model is based on the multilayer perceptron (MLP), batch normalization and Parametric Rectified Linear Unit techniques. In training, the loss function for data mismatch is considered to optimize the model parameters with means absolute error (MAE). In addition, we also use the dynamic weights, calculated by the input data value, to increase the contribution of smaller inputs and avoid errors caused by large values eating small values in total loss. Finally, the surrogate model is applied to simulate a complex pipeline flow in the eastern part of the South China Sea. We use flow and pressure boundary as the input data in the numerical experiment. A total of 215690 high-fidelity training simulations are performed in the offline stage with commercial software LeadFlow, in which 172552 simulation runs are used for training the surrogate model, which takes about 240 min on an RTX2060 graphics processing unit. Then the trained model is used to provide pipeline flow forecasts under various boundary conduction. As a result, it is consistent with those obtained from the high-fidelity simulations (e.g., the media of relative error for temperature is 0.56%, pressure is 0.79%, the gas rate is 1.02%, and oil rate is 1.85%, and water is 0.80%, respectively). The online computations from our surrogate model, about 0.008 s per run, achieve speedups of over 1,250 relative to the high-fidelity simulations, about 10 s per run. Overall, this model provides reliable and fast predictions of the dynamic flow along the pipeline.

show abstract

Section: Introductionmentioning

confidence: 99%

Deep-learning-based surrogate model for fast and accurate simulation in pipeline transport

Feng¹,

Yan²,

Yang³

et al. 2022

Front. Energy Res.

View full text Add to dashboard Cite

show abstract

“…Stator winding [42] Injection molding [43] Rolling element bearing [44] MLP [42] CNN [43], [44] Control systems [42]- [44] Productivity [42], [43] Stability [42], [44] Fault Detection Nuclear power plant [45], [46] Electric arc system [47] Wastewater treatment [48] Oil&Gas pipeline [49], [50] DNN [45] CNN [46], [47], [49] Deep Clustering [48] LSTM [49], [50] SVM [50] Control systems [45] Infrastructures [46]- [50] Stability [45]- [50] Fault Prediction…”

Section: Defect Detectionmentioning

confidence: 99%

“…Researchers in [46], [49], [50] propose various methods for pipeline leak detection. In [46], they implement a CNNbased approach using trajectory-based image features derived from time-series acoustic data.…”

Section: Defect Detectionmentioning

confidence: 99%

A Survey and Perspective on Industrial Cyber-Physical Systems (ICPS): From ICPS to AI-Augmented ICPS

Chae,

Lee,

Jang

et al. 2023

Trans. Ind. Cyb-Phy. Sys.

View full text Add to dashboard Cite

Digital Transformation integrates information technology across a broad spectrum of industrial sectors. Industrial Cyber-Physical Systems (ICPS) play a vital role in this transformation by harmonizing machinery, production, logistics, and societal needs through innovative information technologies. This paper investigates the adoption of industrial artificial intelligence (industrial AI) as a methodology for effective ICPS design, introducing AI-Augmented ICPS (AICPS). The study conducts a survey, focusing on the components and interactions of AICPS. We propose design considerations for the implementation of AICPS and investigate the application of cutting-edge industrial AI techniques in each interaction. From the standpoint of AI augmentation, this paper offers insights by identifying key perspectives, including uncertainty of information, safety of AI, explainability of AI, human-societal interactive ICPS, and standardization of industrial AI. This paper aims to enhance understanding of AICPS and lay the groundwork for integrating independent industrial AI techniques into ICPS.

show abstract

“…A network of sensors to monitor the changes in system conditions like pressure, temperature, or PH of the soil. (Spandonidis et al, 2022;Tariq et al, 2021)…”

Section: Sensing System Working Principle Examplesmentioning

confidence: 99%

State‐of‐the‐art review of leak diagnostic experiments: Toward a smart water network

Bakhtawar

Zayed

2023

WIREs Water

View full text Add to dashboard Cite

Effective water pipeline leak detection and localization can help water practitioners save water losses and reduce pipe rehabilitation budgets. Different sensing technologies can provide essential data aiding model development for predicting future leaks in the water distribution networks. Solutions for applying different sensing technologies to real water distribution networks are being developed for real‐time data collection, precise leak localization, reducing false alarms, and continuous pipeline monitoring. Experiments of different scales and specifications have been conducted to test the application and effectiveness of sensing technologies for leak diagnosis in water pipelines. Quality data collection is especially crucial for the successful implementation of these experiments. However, practitioners need help designing effective leak detection and localization experiments, wasting time, and resources. There needs to be a greater understanding of the design parameters for more accurate testing, such as sensor selection, sensor placement, ambient noise sources, and system operations. The current study, therefore, delineates design parameters for design leak diagnosis experiments for lab, testbed, and real networks. For application in real water networks, expert opinion is used to help identify the benefits and limitations and the implementation issues of available sensing technologies. An example experiment design using multi‐sensing technologies for smart leak diagnostics in Hong Kong is also presented. The study can guide water practitioners and researchers to design a multi‐sensing leak diagnostic experiments enabling a smart and sustainable utility management.This article is categorized under: Engineering Water > Methods Engineering Water > Sustainable Engineering of Water Science of Water > Water Extremes

show abstract

Evaluation of deep learning approaches for oil & gas pipeline leak detection using wireless sensor networks

Cited by 72 publications

References 34 publications

Deep-learning-based surrogate model for fast and accurate simulation in pipeline transport

Deep-learning-based surrogate model for fast and accurate simulation in pipeline transport

A Survey and Perspective on Industrial Cyber-Physical Systems (ICPS): From ICPS to AI-Augmented ICPS

State‐of‐the‐art review of leak diagnostic experiments: Toward a smart water network

Contact Info

Product

Resources

About