Prediction and application of porosity based on support vector regression model optimized by adaptive dragonfly algorithm

ABSTRACT

Porosity is an important parameter of reservoir physical properties and plays an important role in reservoir evaluation. Porosity can be measured by direct and indirect methods. However, most methods are very time-consuming and expensive. Therefore, it is of great significance to establish a fast, economical and effective method for accurate prediction of porosity. In this paper, an alternative method of porosity prediction, based on the integration between novel adaptive population heuristic intelligent optimization algorithm (ADA) and support vector regression (SVR) is presented. In this study, the simulation results of the new model were compared with the DA-SVR, BP, and ELM methods. The experimental results show that the ADA-SVR prediction model can achieve higher prediction accuracy and the prediction accuracy is 96.3%. Therefore, the proposed ADA-SVR model is feasible and effective for predicting porosity and can be used as an effective tool for predicting other reservoir parameters.

KEYWORDS:

• Dragonfly algorithm
• adaptive strategy
• support vector regression
• porosity prediction
• geophysical well logs

Acknowledgments

This work is supported by PetroChina Innovation Foundation (2016D-5007-0305)，Class A Strategic Pilot Science and Technology Project of Chinese Academy of Sciences - Project of Distribution Law and Exploration Evaluation of Deep Oil and Gas （M1701001A）.

Additional information

Funding

This work was supported by the Class A Strategic Pilot Science and Technology Project of Chinese Academy of Sciences - Project of Distribution Law and Exploration Evaluation of Deep Oil and Gas [M1701001A];PetroChina Innovation Foundation [2016D-5007-0305];