FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W4379347285> ?p ?o ?g. }

• W4379347285 abstract "Abstract The design and analysis of multiphase flows in wells, pipelines, and risers, as used in the oil and gas industry, are typically based on either empirical correlations (e.g., Beggs and Brill, Mukherjee and Brill, Duns and Ros) or on first-principles mechanistic models (e.g., the Shell Flow Correlations, the TUFFP Unified model, the Leda Flow Point model, or the OLGA flow correlations). Over the last three decades, Shell has developed its own set of two-phase and three-phase flow models that calculate the flow characteristics in pipelines, such as the flow pattern, the pressure drop, and the liquid hold-up. These multiphase flow (sub)models are collectively known as the Shell Flow Correlations (SFC). The objective of the present work is to explore the possibility of improving the prediction of multiphase pipeline transport through developing a data-driven method for the prediction of the flow pattern, pressure drop and liquid hold-up in a pipeline section. We expect to achieve higher accuracy than mechanistic tools by leveraging the knowledge from the experimental dataset and faster prediction of large amount of data. The method is based on using three surrogate models based on Machine Learning (ML) algorithms trained on a representative lab data set from the SFC experimental database. The first model predicts the flow pattern, the second one the liquid, and the third one the pressure gradient. This may extend the accuracy and the applicability range of the Shell Flow Explorer tool in predicting the multiphase flow characteristics in pipe sections. The approach taken to achieve the goal of the present work is that first synthetic datasets are generated from Shell Flow Explorer and OLGA tool. Different machine learning algorithms are trained on this synthetic dataset. The same trained Machine Learning (ML) algorithm can be used for transfer learning with experimental data sets and the models are evaluated with the predictions from Shell Flow explorer and OLGA tool. In the future, the same machine learning algorithm can be used to incrementally train and test on the new sets of experimental data. After iterations on several new experimental datasets, the machine learning algorithm performance can be assessed on different test dataset distribution independent of the train and validation instances. This document lists the details of the methodology of the data driven models, and the predictions from these data driven models are compared against the results from SFC and OLGA tools. The models developed could easily be integrated with real time data for the better prediction and anomaly detection. Such models can also be used to explore the design space and perform analyses. Additionally, they can be used for system simulation while an asset is operating and connected to IoT (Internet of Things) platform for enhanced monitoring, asset optimisation, diagnostics, and predictive maintenance. This opens a whole new era in value creation for industries to optimize operations and maintenance, as well as further accelerate the new- development process adding value along the value chain of the Shell business." @default.
• W4379347285 created "2023-06-05" @default.
• W4379347285 creator A5015156794 @default.
• W4379347285 creator A5088893559 @default.
• W4379347285 date "2023-06-05" @default.
• W4379347285 modified "2023-09-25" @default.
• W4379347285 title "Machine Learning Based Prediction of Pressure Drop, Liquid-Holdup and Flow Pattern in Multiphase Flows" @default.
• W4379347285 cites W1976883852 @default.
• W4379347285 cites W2005673201 @default.
• W4379347285 cites W2025196663 @default.
• W4379347285 cites W2899186308 @default.
• W4379347285 cites W2901795974 @default.
• W4379347285 cites W2945421839 @default.
• W4379347285 cites W4250984232 @default.
• W4379347285 doi "https://doi.org/10.2118/214348-ms" @default.
• W4379347285 hasPublicationYear "2023" @default.
• W4379347285 type Work @default.
• W4379347285 citedByCount "0" @default.
• W4379347285 crossrefType "proceedings-article" @default.
• W4379347285 hasAuthorship W4379347285A5015156794 @default.
• W4379347285 hasAuthorship W4379347285A5088893559 @default.
• W4379347285 hasConcept C114088122 @default.
• W4379347285 hasConcept C121332964 @default.
• W4379347285 hasConcept C127413603 @default.
• W4379347285 hasConcept C139514615 @default.
• W4379347285 hasConcept C144308804 @default.
• W4379347285 hasConcept C175309249 @default.
• W4379347285 hasConcept C199360897 @default.
• W4379347285 hasConcept C2779379648 @default.
• W4379347285 hasConcept C38349280 @default.
• W4379347285 hasConcept C41008148 @default.
• W4379347285 hasConcept C43521106 @default.
• W4379347285 hasConcept C57879066 @default.
• W4379347285 hasConcept C78519656 @default.
• W4379347285 hasConcept C98156149 @default.
• W4379347285 hasConceptScore W4379347285C114088122 @default.
• W4379347285 hasConceptScore W4379347285C121332964 @default.
• W4379347285 hasConceptScore W4379347285C127413603 @default.
• W4379347285 hasConceptScore W4379347285C139514615 @default.
• W4379347285 hasConceptScore W4379347285C144308804 @default.
• W4379347285 hasConceptScore W4379347285C175309249 @default.
• W4379347285 hasConceptScore W4379347285C199360897 @default.
• W4379347285 hasConceptScore W4379347285C2779379648 @default.
• W4379347285 hasConceptScore W4379347285C38349280 @default.
• W4379347285 hasConceptScore W4379347285C41008148 @default.
• W4379347285 hasConceptScore W4379347285C43521106 @default.
• W4379347285 hasConceptScore W4379347285C57879066 @default.
• W4379347285 hasConceptScore W4379347285C78519656 @default.
• W4379347285 hasConceptScore W4379347285C98156149 @default.
• W4379347285 hasLocation W43793472851 @default.
• W4379347285 hasOpenAccess W4379347285 @default.
• W4379347285 hasPrimaryLocation W43793472851 @default.
• W4379347285 hasRelatedWork W1978068056 @default.
• W4379347285 hasRelatedWork W1983330086 @default.
• W4379347285 hasRelatedWork W2007688570 @default.
• W4379347285 hasRelatedWork W2012680968 @default.
• W4379347285 hasRelatedWork W2056263910 @default.
• W4379347285 hasRelatedWork W2333561054 @default.
• W4379347285 hasRelatedWork W2361951699 @default.
• W4379347285 hasRelatedWork W2515874357 @default.
• W4379347285 hasRelatedWork W4213323399 @default.
• W4379347285 hasRelatedWork W4318571168 @default.
• W4379347285 isParatext "false" @default.
• W4379347285 isRetracted "false" @default.
• W4379347285 workType "article" @default.