A warning thresholds scheme with dynamic oil parameters based on lasso regression and 6sigma

Yuyan Wu, Yueyang Chen, Yueting Shi, Chang Lu, Dongpeng Song

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

2 Citations (Scopus)

Abstract

Dynamic early warning makes great sense for oil management to keep safety and stability of oil production. In this paper, we derive the production regression model, predict production with 10 oil parameters based on Least Absolute Shrinkage and Selection Operator (Lasso) and Least Angle Regression (LARS) methods. The 10 most relevant oil parameters are decided by the warning parameters selection method from kinds of different parameters, which makes the prediction more reliable. The accuracy of regression model achieves 97%. Then we get the warning thresholds based on 6σ. Oil parameters for warning threshold partition experiment are from the database of Tianjin oilfield. The experiment results show that our method is capable of warning both mild and severe situation, and the accuracy is 95%, which runs ahead in oil industry and has great popularization value.

Original languageEnglish
Title of host publicationProceedings - 2016 IEEE International Conference on Digital Signal Processing, DSP 2016
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages190-193
Number of pages4
ISBN (Electronic)9781509041657
DOIs
Publication statusPublished - 2 Jul 2016
Event2016 IEEE International Conference on Digital Signal Processing, DSP 2016 - Beijing, China
Duration: 16 Oct 201618 Oct 2016

Publication series

NameInternational Conference on Digital Signal Processing, DSP
Volume0

Conference

Conference2016 IEEE International Conference on Digital Signal Processing, DSP 2016
Country/TerritoryChina
CityBeijing
Period16/10/1618/10/16

Keywords

  • early warning
  • lasso
  • oil production
  • regression

Fingerprint

Dive into the research topics of 'A warning thresholds scheme with dynamic oil parameters based on lasso regression and 6sigma'. Together they form a unique fingerprint.

Cite this