 Remote Geotechnical Monitoring of a Buried Oil PipelineAlla Yu. Vladova
Mathematics, 2022, vol. 10, issue 11, 1-14 Abstract: Extensive but remote oil and gas fields in Canada and Russia require extremely long pipelines. Global warming and local anthropogenic effects drive the deepening of seasonal thawing of cryolithozone soils and enhance pathological processes such as frost heave, thermokarst, and thermal erosion. These processes lead to a reduction in the subgrade capacity of the soils, causing changes in the spatial position of the pipelines, consequently increasing the number of accidents. Oil operators are compelled to monitor the daily temperatures of unevenly heated soils along pipeline routes. However, they are confronted with the problem of separating anthropogenic heat losses from seasonal temperature fluctuations. To highlight heat losses, we propose a short-term prediction approach to a transformed multidimensional dataset. First, we define the temperature intervals according to the classification of permafrost to generate additional features that sharpen seasonal and permafrost conditions, as well as the timing of temperature measurement. Furthermore, linear and nonlinear uncorrelated features are extracted and scaled. The second step consists of selecting a training sample, learning, and adjusting the additive regression model. Forecasts are then made from the test sample to assess the accuracy of the model. The forecasting procedure is provided by the three-component model named Prophet. Prophet fits linear and nonlinear functions to define the trend component and Fourier series to define the seasonal component; the third component, responsible for the abnormal days (when the heating regime is changed for some reason), could be defined by an analyst. Preliminary statistical analysis shows that the subsurface frozen soils containing the oil pipeline are mostly unstable, especially in the autumn season. Based upon the values of the error metrics, it is determined that the most accurate forecast is obtained on a three-month uniform time grid. Keywords: ESG; pipeline; remote monitoring; data analysis; machine learning; time series (search for similar items in EconPapers) Downloads: (external link) Related works: Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text Persistent link: https://EconPapers.repec.org/RePEc:gam:jmathe:v:10:y:2022:i:11:p:1813-:d:823786 Access Statistics for this article Mathematics is currently edited by Ms. Emma He More articles in Mathematics from MDPI |
Is your work missing from RePEc? Here is how to contribute.
Questions or problems? Check the EconPapers FAQ or send mail to .
EconPapers is hosted by the
School of Business at Örebro University.