Cristin-resultat-ID: 2054475
Sist endret: 16. mars 2023, 13:56
Resultat
Vitenskapelig Kapittel/Artikkel/Konferanseartikkel
2022

Sequential Multi-Realization Probabilistic Interpretation of Well Logs and Geological Prediction by a Deep-Learning Method

Bidragsytere:
  • Sergey Alyaev
  • Adrian Ambrus
  • Nazanin Jahani og
  • Ahmed Elsheikh

Bok

Transactions of the SPWLA 2022
ISBN:
  • 978-0123009562

Utgiver

SPWLA

Om resultatet

Vitenskapelig Kapittel/Artikkel/Konferanseartikkel
Publiseringsår: 2022
ISBN:
  • 978-0123009562

Klassifisering

Fagfelt (NPI)

Fagfelt: Matematikk
- Fagområde: Realfag og teknologi

Beskrivelse Beskrivelse

Tittel

Sequential Multi-Realization Probabilistic Interpretation of Well Logs and Geological Prediction by a Deep-Learning Method

Sammendrag

The majority of geosteering operations rely on traditional shallow sensing logging tools as sources of information. Many such operations rely on stratigraphic-based steer-ing when the logs from the drilled well are matched to logs from an offset well by modifying the lateral shape of stratigraphy. The match of the logs indicates a plau-sible interpretation, but due to the scarcity of log data in many situations, this interpretation is not unique. In man-ual workflows maintaining several likely interpretations is not realistic and in automated workflows, multiple in-terpretations are seldom used. We describe a deep neural network (DNN) that outputs a selected number of stratigraphic interpretations using a single evaluation of the input log data in two millisec-onds. The input data defined prior to training consists of one or several log pairs consisting of one current lat-eral and one offset-well log. For each of the interpreta-tions, the DNN also estimates the respective probability and can be configured to produce likely ahead-of-data predictions of the geology, which are based on the data mismatches and the likelihood of geological configura-tions with respect to the training dataset. The described probabilistic interpretation and prediction is enabled by the supervised training of a mixture density DNN (MDN) with a stable multiple-trajectory-prediction loss function. In this paper, we apply the MDN for the sequential in-terpretation of well logs. We use the interpretations and the probabilities from the previous interpretation step as starting points for the probabilistic interpretations and predictions for the current step. We avoid the curse of dimensionality by discarding the unlikely starting points. The batchable MDN evaluation enables tracking of hun-dreds of solutions while still maintaining sub-second per-formance, compared to minute(s) reported in other recent papers. The performance of the method is verified on syn-thetic test data as well as the realistic well data from the Geosteering World Cup 2020 (based on the Middle Woodford formation, located in the South Central Okla-homa Oil Province in the United States) and stratigraphic configurations provided by geologists. In all cases, the method manages to capture likely interpretations. At the same time, the accuracy of predictions deteriorates for the configurations which were not typical for the train-ing dataset.

Bidragsytere

Aktiv cristin-person

Sergey Alyaev

  • Tilknyttet:
    Forfatter
    ved NORCE Energi og teknologi ved NORCE Norwegian Research Centre AS

Adrian Ambrus

  • Tilknyttet:
    Forfatter
    ved NORCE Energi og teknologi ved NORCE Norwegian Research Centre AS

Nazanin Jahani

  • Tilknyttet:
    Forfatter
    ved NORCE Energi og teknologi ved NORCE Norwegian Research Centre AS

Ahmed Elsheikh

  • Tilknyttet:
    Forfatter
    ved University of Liverpool
1 - 4 av 4

Resultatet er en del av Resultatet er en del av

Transactions of the SPWLA 2022.

Torres-Verdin, Carlos. 2022, Vitenskapelig antologi/Konferanseserie
1 - 1 av 1