The Golem: A General Data-Driven Model for Oil & Gas Forecasting Based on Recurrent Neural Networks

Abstract

Oil & gas forecasting is one of the most critical issues in reservoir management. Physics-based simulations are the most common models used for production forecasts in oilfields. Previous works based on Machine Learning (ML) developed models focused on the oil rate as unique target variable, a forecasting by one-day output , and just one class of reservoir (synthetic or actual). This work introduces a general data-driven model based on Recurrent Neural Networks to forecast an adaptive sequence of timestamps for the complete production rates (oil, gas, and water), and we also included the wellbore pressure as target variable, for both classes of reservoirs as actual as synthetic. The first dataset was obtained from the synthetic benchmark UNISIM-II-H, which simulates a carbonate reservoir in the Brazilian pre-salt; the second dataset is from an actual reservoir, the Volve oilfield, a decommissioned reservoir in the Norwegian North Sea. The forecasting is calculated using an input sequence of daily values from the historical record of the production rates and the pressure; the output is also a set of the values to the next sequence of days for one selected production variable (oil, gas, water, or pressure). The size of both input and output sets is adaptive and its adjustment depends on the dataset size and the production time. We built the model and compared it between the Long Short-Term Memory (LSTM) and the Gated Recurrent Unit (GRU) implementations. We tuned the architectural parameters of the model, the input size of historical records, and the output forecasting days. We performed the training/testing procedures with several sizes for the training dataset from the target wellbore and tested with the remaining data to evaluate the model stability. We adopted the Symmetric Mean Absolute Percentage Error (SMAPE) and the coefficient of determination r-square (R 2 ) metrics to compare our forecasting values to the production rates and the pressure, most of the results for both synthetic and actual oilfields exhibited that the model can follow an accurate trend of the production rates and the pressure; and the output values are approximated. Forecasting values from the designed model exhibited closer values when compared to the expected data from the wellbores in most of the experiments, some cases exhibited an SMAPE lower than 2 and R 2 up to 0.99. The model can learn the behavior of each production variable according to the training set and the forecasting output can be adapted for a set of several timestamps.

Predictive Analytics Performance on Oil and Gas: A Significant Review

“…The time domain of the reservoir focuses on the Volve and UNISIM-IIH oilfields, utilized Long Short-Term Memory (LSTM) and GRU models for the classification of 3,257 samples based on oil, gas, water, or pressure levels [63]. Regarding O&G forecasting, the GRU model emerged as the frontrunner, with an amazing R 2 of 99%.…”

“…The proposed model LSTM-AE-OCSVM gets a greater accuracy of 98%, and the researcher proposed using anomalous data in future studies. Meanwhile, Martinez & Rocha [63] focused on reservoirs and used 3,257 samples from the Volve and UNISIM-IIH oilfields to examine LSTM and GRU models. With an impressive R 2 of 99%, the GRU model demonstrated its superiority in O&G forecasting when classifying oil, gas, water, or pressure.…”

A multivariate natural gas load forecasting method based on residual recurrent neural network

“…As shown in Figure 3a, although the gas change pattern can be largely grasped only based on history data, the essential law of gas forecasting has been explored after multi-factor introduction, based on which more accurate results can be obtained. Moreover, the fitting results in terms of Mean Square Error (MSE), Mean Absolute Error (MAE), R-Square (R2) and Symmetric Mean Absolute Percentage Error (SMAPE) [7,9] have been summarized in Table 1. It can be seen that the forecasting performances based our multifactors datasets are mostly better than that of history-only datasets.…”

“…Moreover, the fitting results in terms of Mean Square Error (MSE), Mean Absolute Error (MAE), R‐Square (R2) and Symmetric Mean Absolute Percentage Error (SMAPE) [7, 9] have been summarized in Table 1. It can be seen that the forecasting performances based our multi‐factors datasets are mostly better than that of history‐only datasets.…”

“…Reference [7] utilized a fully connected deep neural network (DNN) to track and extract the dynamic features of natural gas load patterns in correlated multivariate time series, thereby improving the prediction accuracy. Reference [9] introduces a general data‐driven model based on Recurrent Neural Networks to forecast an adaptive sequence of timestamps for the gas load, which was tested under a wide variety of conditioning and is shown to produce excellent results.…”