Parallel and distributed chimp-optimized LSTM for oil well-log reconstruction in China

Area of Science:

• Geoscience and Petroleum Engineering
• Artificial Intelligence and Machine Learning

Background:

• Well-log analysis is crucial for oil and gas extraction.
• Inconsistent or missing well-log data hampers geological interpretation and resource estimation.

Purpose of the Study:

• To develop a scalable and efficient deep learning model for reconstructing missing well-log data.
• To enhance the accuracy of geological analysis and hydrocarbon resource estimation.

Main Methods:

• A deep Long Short-Term Memory (LSTM) model was employed.
• The Parallel and Distributed Chimp Optimization Algorithm (PDCOA) was utilized for accelerated hyperparameter tuning.
• PDCOA enables parallel processing across multiple computers with regular communication for diversity and reliability.

Main Results:

• The proposed method demonstrates superior scalability, efficiency, and predictive accuracy compared to existing approaches.
• Successful reconstruction of missing well-log data was achieved.
• The model proves effective in addressing data inconsistencies in well-log analysis.

Conclusions:

• The PDCOA-enhanced LSTM model offers a robust solution for well-log data reconstruction.
• This advancement significantly aids geological interpretation and hydrocarbon resource assessment.
• The method presents a valuable tool for the oil and gas industry.