Artificial Neural Network Modeling for the Prediction of Oil Production (original) (raw)
Related papers
Numerical simulations and decline curve analysis are classic tools used for predicting reservoir performance. Numerical simulations are a very complex tool, offering a nonunique solution with a high degree of uncertainty. Decline curve analysis does not take into account opening or closing intervals and variable injection rates. In this study, the authors designed a feedforward backpropagation neural network model as an alternative technique for predicting oil reservoir production performance. Real historical production data obtained from a Libyan oil field was used to train the network. This training network can serve as a practical reservoir production management tool.
Application of artificial neural network for predicting production flow rates of gaslift oil wells
Tạp chí Khoa học Kỹ thuật Mỏ- Địa chất, 2022
In petroleum industry, the prediction of oil production flow rate plays an important role in tracking the good performance as well as maintaining production flow rate. In addition, a flow rate modelling with high accuracy will be useful in optimizing production properties to achieve the expected flow rate, enhance oil recovery factor and ensure economic efficiency. However, the oil production flow rate is traditionally predicted by theoretical or empirical models. The theoretical model usually gives predicted results with a wide variation of error, this model also requires a lot of input data that might be time-consuming and costly. The empirical models are often limited by the volume of data set used to construct the model, therefore predicted values from the applications of these models in practical condition are not highly accurate. In this research, the authors propose the use of an artificial neural network (ANN) to establish a better relationship between production properties and oil production flow rate and predict oil production flow rate. Using production data of 5 wells which use continuous gas lift method in X oil field, Vietnam, an ANN system was developed by using back-propagation algorithm and tansig function to predict production flow rate from the above data set. This ANN system is called a back-propagation neural network (BPNN). In comparison with the oil production flow rate data collected from these studied continuous gas lift oil wells, the predicted results from the constructed ANN achieved a very high correlation coefficient (98%) and low root mean square error (33.41 bbl/d). Therefore, the developed ANN models can serve as a practical and robust tool for oilfield prediction of production flow rate.
2015
Porosity is one of the most significant parameters of hydrocarbon reservoirs describing the quality of reservoirs rocks. It is one of the most crucial characteristics that need to be predicted for evaluation of reservoirs. The conventional methods for porosity determination are core analysis and well test technique. These methods are however very expensive and time-consuming tasks. One of the comparatively inexpensive and readily available sources of inferring porosity is nuclear magnetic resonance (NMR) log. The aim of this paper is to present an application of two machine learning methodologies, which are christened general regression neural network (GRNN) and back-propagation neural network (BPNN), for prediction of NMR porosity using well log data and intelligent models. Available data of three was considered for training and testing the networks. Verification process was also performed by one remaining well. Obtained results have shown that the overall correlation coefficients ...
Neural Networks for Prediction of Oil Production
IFAC-PapersOnLine, 2018
Artificial neural networks are popular tool in oil and gas industry. The objective of this paper is to develop an artificial neural network that predicts oil production in State Oil Company of Azerbaijan Republic (SOCAR). Multilayer perceptron neural network is applied to predict oil production. As a result 99 percent accuracy in training procedure is achieved. The accuracy of the developed model was compared with empirical correlations. Great matching between neural network and actual data has been found.
Artificial Neural Network Surrogate Modeling of Oil Reservoir: A Case Study
Advances in Neural Networks – ISNN 2019, 2019
We develop a data-driven model, introducing recent advances in machine learning to reservoir simulation. We use a conventional reservoir modeling tool to generate training set and a special ensemble of artificial neural networks (ANNs) to build a predictive model. The ANN-based model allows to reproduce the time dependence of fluids and pressure distribution within the computational cells of the reservoir model. We compare the performance of the ANN-based model with conventional reservoir modeling and illustrate that ANN-based model (1) is able to capture all the output parameters of the conventional model with very high accuracy and (2) demonstrate much higher computational performance. We finally elaborate on further options for research and developments within the area of reservoir modeling.
Iraqi Journal of Chemical and Petroleum Engineering, 2009
A common field development task is the object of the present research by specifying the best location of new horizontal re-entry wells within AB unit of South Rumaila Oil Field. One of the key parameters in the success of a new well is the well location in the reservoir, especially when there are several wells are planned to be drilled from the existing wells. This paper demonstrates an application of neural network with reservoir simulation technique as decision tool. A fully trained predictive artificial feed forward neural network (FFNNW) with efficient selection of horizontal re-entry wells location in AB unit has been carried out with maintaining a reasonable accuracy. Sets of available input data were collected from the exploited grids and used in the training and testing of the used network. A comparison between the calculated and observed cumulative oil production has been carried out through the testing steps of the constructed ANN, an absolute average percentage error of t...
Application of Regression and Neural Network Models in Computing Forecasts for Crude Oil Productions
2015
In this study, a quadratic regression model and a two layered layer recurrent neural network (TLLRNN) method were used to model forecasting performance of the daily crude oil production data of the Nigerian National Petroleum Corporation (NNPC). The two methods were applied on the difference series and log difference series of the NNPC series. The results indicates that the two layered layer recurrent neural network model have better forecasting per-formance greater than the quadratic regression method based on the mean error square sense. The root mean square error (RMSE) and the mean absolute error (MAE) were applied to ascertain the assertion that the two layered layer recurrent neural network method have better forecasting performance greater than the quadratic regression method. The outcome of the analysis also indicates that modeling forecasting performance of the NNPC data with the log dif-ference series of the data gives greater forecasting performances than modeling with th...
Neural Network Based Performance Evaluation of a Waterflooded Oil Reservoir
In this paper, we considered the use of neural networks in the identification and prediction of a waterflooded reservoir consisting of eight injection wells and one production well with a 40% porosity. The data used for the non-linear identification was generated from a reservoir modelled in MATLAB Reservoir Simulation Toolbox (MRST). Likewise, in this study, the effect of number of hidden neurons on the accuracy, Mean Squared Error and oil production prediction of the reservoir was investigated. The study asserted the efficacy of the neural networks as regards to its predictive capacity. For the oil production rate, a mean squared error was recorded to be minimal for 2 hidden neurons as compared to the other three cases of neuron number. For water production rate, 8 hidden neurons were observed to be optimal compared to other cases. Oil and water production rate for a peak NPV value of 3 billion US dollars was recorded to be 2000m 3 /day and 4500m 3 /day respectively. The response was optimal for all cases except for the net present value, which requires a more substantial amount of data for the neural network model.
Analysis of Prediction of Pressure Data in Oil Wells Using Artificial Neural Networks
2010 IEEE Electronics, Robotics and Automotive Mechanics Conference, 2010
We present a methodology that integrates an artificial intelligent technology called Artificial Neural Networks (ANN´s) to develop and build a forecasting system that determines the behavior of the pressure of an oil reservoir, from its behavior, considered as reference in relation to four neighboring wells, which are producing at the same stratum. 356 data records were taken (a period of one year). During that period, it was observed that pressure curves show a decrease, which describes the behavior of the reservoir. It was also considered as an additional parameter the average pressure of the reservoir, whose information was obtained from the curves, describing the behavior of bottom pressure in the same stratum during the given period. Finally, we present the results of the predictions of pressure data, compared with the actual values of the reservoirs known, to discuss and assess the accuracy of the prediction of the proposed system.
Building the foundation for Prudhoe Bay oil production optimisation using neural networks
International Journal of Oil, Gas and Coal Technology, 2008
Field data from the Prudhoe Bay oil field in Alaska was used to develop a neural network model of the cross-country gas transit pipeline network between the production separation facilities and central gas compression plant. The trained model was extensively tested and verified using 30% of the data that was not used during the training process. The results show good accuracy in reproducing the actual rates and pressures at the separation facilities and at the gas compression plant. The correlation coefficient for rate and pressure were 0.997 and 0.998, respectively. This development builds the foundation for building a tool to maximise total field oil production by optimising the gas discharge rates and pressures at the separation facilities.