The back propagation based on the modified group method of data-handling network for oilfield production forecasting

Abstract

In this paper, a novel hybrid forecasting model combining modified group method of data handling (GMDH) and back propagation (BP) is introduced for time series oilfield production forecasting. The proposed model takes advantages of both the modified GMDH networks in effective parameter selection and the BP network in excellent nonlinear mapping and provides a robust simulation ability for oilfield production with higher precision. Various production parameters of an actual oilfield were utilized to analyze and test the annual output predicted by proposed model (modified GMDH-BP). The performance of the proposed model was compared with the multiple linear regression (MLR), GMDH, modified GMDH, BP, and the hybrid model combining group method of data handling and back propagation (GMDH-BP) using time series annual production data. The relative error, correlation coefficient (R), root mean square error, mean absolute percentage of error, and scatter index were utilized to investigate the performance of the presented models. The evaluation results indicate that the hybrid model provides more accurate production forecasts compared to other models and exhibits a robust simulation ability for capturing the nonlinear relation of complex production time series prediction of oilfield.

Prediction of Permeability Using Group Method of Data Handling (GMDH) Neural Network from Well Log Data

“…In an iterative manner, the procedure will be repeated until entire pairs are estimated to produce the novel pairs of regression which would be kept into another novel matrix termed as Y matrix in Equation (10). These newly regression pairs generated could be taken as the newly enhanced variables which have the best predictability compared to that of data set X (Equation 3)…”

“…The extensively applied intelligent method in permeability prediction with improved accuracy over that of statistical methods is the artificial neural network (ANN) [3,9]. This is as a result of its ability to discover patterns in non-linear and complex data systems [10][11][12][13]. However, standard ANN is known to exhibit several drawbacks such as overfitting, converging at local minima, and low computational speed.…”

Research on the Application of Integrated Learning Models in Oilfield Production Forecasting

“…Liu et al introduced a LSTM learning method for set and empirical pattern decomposition. Guo et al proposed a hybrid prediction model that combines an improved grouped data processing method (GMDH) with a backpropagation algorithm. A hybrid approach based on linear, statistical, and machine learning models has been successfully applied in many other fields , but has not received sufficient attention in well-production forecasting.…”