Bottom hole pressure prediction based on hybrid neural networks and Bayesian optimization

Many scholars have focused on applying machine learning models in bottom hole pressure (BHP) prediction. However, the complex and uncertain conditions in deep wells make it difficult to capture spatial and temporal correlations of measurement while drilling (MWD) data with traditional intelligent models. In this work, we develop a novel hybrid neural network, which integrates the Convolution Neural Network (CNN) and the Gate Recurrent Unit (GRU) for predicting BHP fluctuations more accurately. The CNN structure is used to analyze spatial local dependency patterns and the GRU structure is used to discover depth variation trends of MWD data. To further improve the prediction accuracy, we explore two types of GRU-based structure: skip-GRU and attention-GRU, which can capture more long-term potential periodic correlation in drilling data. Then, the different model structures tuned by the Bayesian optimization (BO) algorithm are compared and analyzed. Results indicate that the hybrid models can extract spatial-temporal information of data effectively and predict more accurately than random forests, extreme gradient boosting, back propagation neural network, CNN and GRU. The CNN-attention-GRU model with BO algorithm shows great superiority in prediction accuracy and robustness due to the hybrid network structure and attention mechanism, having the lowest mean absolute percentage error of 0.025%. This study provides a reference for solving the problem of extracting spatial and temporal characteristics and guidance for managed pressure drilling in complex formations.

Application of artificial intelligence in predicting the dynamics of bottom hole pressure for under-balanced drilling:Extra tree compared with feed forward neural network model

This study used six fields data alongside correlation heat map to evaluate the field parameters that affect the accuracy of bottom hole pressure(BHP)estimation.The six oil field data were acquired using measurement while drilling device to collect surface measurements of the downhole pressure data while drilling.For the two case studies,measured field data of the wellbore filled with gasified mud system was utilized,and the wellbores were drilled using rotary jointed drill strings.Extremely Randomized Tree and feed forward neural network algorithms were used to develop models that can predict with high accuracy,BHP from measured field data.For modeling purpose,an extensive data from six fields was used,and the proposed model was further validated with two data from two new fields.The gathered data encompasses a variety of well data,general information/data,depths,hole size,and depths.The developed model was compared with data obtained from two new fields based on its capability,stability and accuracy.The result and model’s performance from the error analysis revealed that the two proposed Extra Tree and Feed Forward models replicate the bottom hole pressure data with R2 greater than 0.9.The high values of R^(2) for the two models suggest the relative reliability of the modelling techniques.The magnitudes of mean squared error and mean absolute percentage error for the predicted BHPs from both models range from 0.33 to 0.34 and 2.02%-2.14%,for the Extra tree model and 0.40-0.41 and 3.90%e3.99%for Feed Forward model respectively;the least errors were recorded for the Extra Tree model.Also,the mean absolute error of the Extra Tree model for both fields(9.13-10.39 psi)are lower than that of the Feed Forward model(10.98-11 psi),thus showing the higher precision of the Extra Tree model relative to the Feed Forward model.Literature has shown that underbalanced operation does not guarantee the improvement of horizontal well’s extension ability,because it mainly depends on the relationship between the bottomhole pressure and its corresponding critical point.Thus,the application of this study proposed models for predicting bottomhole pressure trends.

Bottom hole pressure estimation using hybridization neural networks and grey wolves optimization

An effective design and optimum production strategies of a well depend on the accurate prediction of its bottom hole pressure(BHP)which may be calculated or determined by several methods.However,it is not practical technically or economically to apply for a well test or to deploy a permanent pressure gauge in the bottom hole to predict the BHP.Consequently,several correlations and mechanistic models based on the knownsurfacemeasurementshave beendeveloped.Unfortunately,all these tools(correlations&mechanistic models)are limited to some conditions and intervals of application.Therefore,establish a global model that ensures a large coverage of conditions with a reduced cost and high accuracy becomes a necessity.In this study,we propose new models for estimating bottom hole pressure of vertical wells with multiphase flow.First,Artificial Neural Network(ANN)based on back propagation training(BP-ANN)with 12 neurons in its hidden layer is established using trial and error.The next methods correspond to optimized or evolved neural networks(optimize the weights and thresholds of the neural networks)with Grey Wolves Optimization(GWO),and then its accuracy to reach the global optima is compared with 2 other naturally inspired algorithms which are the most used in the optimization field:Genetic Algorithm(GA)and Particle Swarms Optimization(PSO).The models were developed and tested using 100 field data collected from Algerian fields and covering a wide range of variables.The obtained results demonstrate the superiority of the hybridization ANN-GWO compared with the 2 other hybridizations or with the BP learning alone.Furthermore,the evolved neural networks with these global optimization algorithms are strongly shown to be highly effective to improve the performance of the neural networks to estimate flowing BHP over existing approaches and correlations.

Low parameter model to monitor bottom hole pressure in vertical multiphase flow in oil production wells

The importance of the flow patterns through petroleum production wells proved for upstream experts to provide robust production schemes based on the knowledge about flow behavior.To provide accurate flow pattern distribution through production wells,accurate prediction/representation of bottom hole pressure(BHP)for determining pressure drop from bottom to surface play important and vital role.Nevertheless enormous efforts have been made to develop mechanistic approach,most of the mechanistic and conventional models or correlations unable to estimate or represent the BHP with high accuracy and low uncertainty.To defeat the mentioned hurdle and monitor BHP in vertical multiphase flow through petroleum production wells,inventive intelligent based solution like as least square support vector machine(LSSVM)method was utilized.The evolved first-break approach is examined by applying precise real field data illustrated in open previous surveys.Thanks to the statistical criteria gained from the outcomes obtained from LSSVM approach,the proposed least support vector machine(LSSVM)model has high integrity and performance.Moreover,very low relative deviation between the model estimations and the relevant actual BHP data is figured out to be less than 6%.The output gained from LSSVM model are closed the BHP while other mechanistic models fails to predict BHP through petroleum production wells.Provided solutions of this study explicated that implies of LSSVM in monitoring bottom-hole pressure can indicate more accurate monitoring of the referred target which can lead to robust design with high level of reliability for oil and gas production operation facilities.

Forecasting multiphase flowing bottom-hole pressure of vertical oil wells using three machine learning techniques

Flowing bottom-hole pressure(FBHP)is a key metric parameter in the evaluation of performances of oil and gas production wells.An accurate prediction of FBHP is highly required in the petroleum industry for many applications,such the hydrocarbon production optimization,oil lifting cost,and assessment of workover operations.Production and reservoir engineers rely on empirical correlations and mechanistic models exist in open resources to estimate the FBHP.Several empirical models have been developed based on simulation and laboratory results that involved many assumptions that reduce the model's accuracy when they are applied for the field applications.The technologies of machine learning(ML)are one discipline of Artificial Intelligence(AI)techniques provide promising tools that help solving human's complex problems.This study develops machine-learning based models to predict the multiphase FBHP using three machine learning techniques that are Random forest,K-Nearest Neighbors(KNN),and artificial neural network(ANN).Results showed that using an artificial neural network model give error of 2.5%to estimate the FBHP which is less than the random forest and K-nearest neighbor models with error of 3.6%and 4%respectively.The ML models were developed based on a surface production data,which makes the FBHP is predicted using actual field data.The accuracy of the proposed models from ML was evaluated by comparing the results with the actual dataset values to ensure the effectiveness of the work.The results of this study show the potential of artificial intelligence in predicting the most complex parameter in the multiphase petroleum production process.