Iterative learning-based many-objective history matching using deep neural network with stacked autoencoder

• Published in: Petroleum science Vol. 18; no. 5; pp. 1465 - 1482
• Main Authors: Kim, Jaejun, Park, Changhyup, Ahn, Seongin, Kang, Byeongcheol, Jung, Hyungsik, Jang, Ilsik
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.10.2021
• Subjects: Clustering; Deep neural network; History matching; Iterative learning; Many-objective; Stacked autoencoder; Stacked autoencoder; Deep neural network; History matching; Many-objective; Clustering; Iterative learning
• ISSN: 1995-8226; 1995-8226
• DOI: 10.1016/j.petsci.2021.08.001

This paper presents an innovative data-integration that uses an iterative-learning method, a deep neural network (DNN) coupled with a stacked autoencoder (SAE) to solve issues encountered with many-objective history matching. The proposed method consists of a DNN-based inverse model with SAE-encoded static data and iterative updates of supervised-learning data are based on distance-based clustering schemes. DNN functions as an inverse model and results in encoded flattened data, while SAE, as a pre-trained neural network, successfully reduces dimensionality and reliably reconstructs geo-models. The iterative-learning method can improve the training data for DNN by showing the error reduction achieved with each iteration step. The proposed workflow shows the small mean absolute percentage error below 4% for all objective functions, while a typical multi-objective evolutionary algorithm fails to significantly reduce the initial population uncertainty. Iterative learning-based many-objective history matching estimates the trends in water cuts that are not reliably included in dynamic-data matching. This confirms the proposed workflow constructs more plausible geo-models. The workflow would be a reliable alternative to overcome the less-convergent Pareto-based multi-objective evolutionary algorithm in the presence of geological uncertainty and varying objective functions.