Abstract:Hydraulic fracturing pressure is one of the key criteria in defining the pumping schedule and in dimensioning surface plants for fracturing operations. However, it is difficult to predict at a practical field scale due to the complex determining factors, involving geological, hydrodynamic and rock mechanical parameters. This study proposes a synthetic data-driven workflow, integrating both hydrodynamic and rock mechanical models, to predict the pressure based on fracturing experience in neighboring wells. The data quality and performance of the geological, hydrodynamic and rock mechanical features are evaluated based on a backward elimination strategy and control variate method, using error evolutions as criteria. Relatively small errors (root mean square error 5.1 ~ 5.7 MPa and mean absolute error 5.6 ~ 8.3%) are returned for wells/cases within the same region as the training wells, demonstrating the superior performance of the workflow. The prediction errors increase significantly with increasing distance between training and testing wells – defining the range of applicability of the workflow. The rock mechanical feature (represented by the brittleness index) provides a larger contribution to the prediction than that of the hydrodynamic feature (represented by the proppant accumulation) in most of the testing cases. Young's Modulus exhibits a higher performance (induces the smallest errors in pressure prediction) to characterize the rock mechanical feature of the formation, compared with the brittle mineral ratio and Poisson's Ratio. The data quality of geological stresses and Poisson's Ratio may require improvements based on the irregular error evolutions. The remaining errors and proximity/regional limitations of the data-driven workflow are critically discussed. This new method provides a platform for the prediction of pressures at field-practical scales, which may be significant for both hydraulic fracturing and geological storage of CO 2 and H 2 in depleted reservoirs with sufficient historical hydraulic fracturing records.

Predicting the Minimum Horizontal Principal Stress Using Genetic Expression Programming and Borehole Breakout Data

Prediction of In-Situ Stresses by Using Machine Learning and Intelligent Optimization Algorithms

Prediction of the Maximum Horizontal Principal Stress from Breakout Data Using Generative Adversarial Networks and Backpropagation Neural Network

Prediction of Uniaxial Compressive Strength in Rocks Based on Extreme Learning Machine Improved with Metaheuristic Algorithm

Prediction of Peak Shear Strength of Rock Joints Based on Back-Propagation Neural Network

ROCKBURST PREDICTION USING PARTICLE SWARM OPTIMIZATION ALGORITHM AND GENERAL REGRESSION NEURAL NETWORK

Probabilistic Evaluation of Ground-Support Interaction for Deep Rock Excavation Using Artificial Neural Network and Uniform Design

Applied machine learning-based models for determining the magnitude of pore pressure and minimum horizontal stress

Estimation of minimum horizontal stress, geomechanical modeling and hybrid neural network based on conventional well logging data – a case study

Prediction of Fracturing Pressure and Parameter Evaluations at Field Practical Scales

Effets du gène « crête en pois » et d'autres gènes à effet visible sur le poids des coqs adultes

Enhanced prediction of bolt support drilling pressure using optimized Gaussian process regression

Prediction of hydraulic fracture initiation pressure in a borehole based on a neural network model considering plastic critical distance

Prediction of Uniaxial Compressive Strength Using Fully Bayesian Gaussian Process Regression (fB-GPR) with Model Class Selection

Developing a geomechanical model to predict breakdown pressure in a vertical borehole using failure analysis: a case study

Back-analysis for initial ground stress field at a diamond mine using machine learning approaches

A hybrid machine learning optimization algorithm for multivariable pore pressure prediction

Stability prediction of underground entry-type excavations based on particle swarm optimization and gradient boosting decision tree

A true triaxial strength criterion for rocks by gene expression programming

Ischemia-stimulated glucose uptake does not require catecholamines in rat heart.

A logistic regression classifier for long-term probabilistic prediction of rock burst hazard