Dynamic prediction method for surface pump pressure based on spatio-temporal data features

Patent metadata

Field	Value
Publication number	US-12560081-B2
Application number	US-202519183020-A
Country	US
Kind code	B2
Filing date	Apr 18, 2025
Priority date	Jul 19, 2024
Publication date	Feb 24, 2026
Grant date	Feb 24, 2026

Abstract

A dynamic prediction method for surface pump pressure based on spatio-temporal data features includes: S 1 . obtaining original fracturing treatment data and formation evaluation data of a plurality of wells; S 2 . extracting fracturing treatment data from a prepad fluid initiation stage to a shut-down stage in the fracturing treatment data; S 3 . constructing time series data with the fracturing treatment data, and construct active independent variable data by combining the fracturing treatment data and formation evaluation data; S 4 . establishing a hybrid neural network model and a LightGBM gradient boosting model; S 5 . training the hybrid neural network model and the LightGBM gradient boosting model with the time series data and the active independent variable; S 6 . obtaining fracturing treatment data of a new fracturing well, and predicting with the trained models; and S 7 . initializing weights of the prediction models, updating and iterating the weights, and obtaining a final prediction value.

First claim

What is claimed is:

1 . A hydraulic fracturing method, comprising:
checking a fracturing truck to ensure clearance of a ground pipeline; starting a high-pressure pump of the fracturing truck and testing pressure bearing performance of equipment above a wellhead valve and the ground pipeline; starting the fracturing truck and extruding fracturing fluid into a formation based on a pump rate; injecting sand carrying fluid including the fracturing fluid and a proppant; pushing the sand carrying fluid in a tubing or a casing into the formation; and dynamically predicting a surface pump pressure based on spatio-temporal data features and adjusting abnormal bottomhole pressure in time, wherein dynamically predicting a dynamic prediction method for surface pump pressure based on spatio-temporal data features, the dynamic prediction method is applied in predicting and adjusting an abnormal bottomhole pressure in time comprises the following steps:
S 1 : obtaining original fracturing treatment data of a complete fracturing treatment cycle of a plurality of wells and formation evaluation data of a corresponding fracturing formation, wherein the fracturing treatment data in the step S 1 comprises: the pump rate, proppant concentration and proppant ratio; the formation evaluation data comprises maximum horizontal principal stress, minimum horizontal principal stress, formation pressure, pore pressure, and vertical stress, wherein the pump rate is obtained by a fracturing pump; the formation pressure and the pore pressure are obtained by a Drill Stem Testing, (DST) drill pipe testing tool;
S 2 : dividing the complete fracturing treatment cycle into different intervals according to the original fracturing treatment data, and extracting fracturing treatment data from a prepad fluid initiation stage to a shut-down stage;
S 3 : preprocessing the fracturing treatment data obtained in the step S 2 , constructing time series data with a sliding window algorithm, preprocessing and then combining the fracturing treatment data and formation evaluation data to construct active independent variable data, wherein the active independent variable data is data that is actively adjusted to influence bottomhole pressure;
S 4 : establishing a Convolutional Neural Network-Long Short-Term Memory, (CNN-LSTM)-Attention hybrid neural network model for receiving the time series data and establishing a Light Gradient Boosting Machine, (LightGBM) gradient boosting model for receiving the active independent variable data;
S 5 : training the CNN-LSTM-Attention hybrid neural network model with the time series data obtained in the step S 3 , training the LightGBM gradient boosting model with the active independent variable data, and performing hyperparameter optimization on the training process;
S 6 : obtaining fracturing treatment data and formation evaluation data of a new fracturing well, and performing future multi-step prediction on the bottomhole pressure with the models trained in the step S 5 at a sand carrying fluid initiation stage;
S 7 : initializing weights of prediction results of the CNN-LSTM-Attention hybrid neural network model and the LightGBM gradient boosting model, separately calculating errors of the two models according to prediction results of each round and a real pressure, and optimizing the weights of the prediction results of the two models to obtain a prediction value of the surface pump pressure;
determining that the predicted pressure is abnormal; and
during a hydraulic fracturing construction, adjusting construction parameters in advance upon determination that the predicted pressure is abnormal, to avoid sand plugging.

2 . The hydraulic fracturing method according to claim 1 , wherein the preprocessing obtained fracturing treatment data in the step S 3 to obtain preprocessed data comprises the following steps:
S 31 : deleting null values and interpolating missing value data;
S 32 : standardizing the fracturing treatment data; and
S 33 : dividing the standardized fracturing treatment data in a certain proportion to obtain a training data set, a verification data set and a test data set.

3 . The hydraulic fracturing method according to claim 1 , wherein the constructing time series data with a sliding window algorithm in the step S 3 comprises the following steps:
constructing a time window by data of a part of time points in the past, wherein a number of time points contained in each subsequence is a size of the time window; and moving the time window forward one or more time units in an entire data set to form a plurality of time windows.

4 . The hydraulic fracturing method according to claim 1 , wherein the active independent variable data in the step S 3 comprises pump rate, proppant concentration, and proppant ratio.

5 . The hydraulic fracturing method according to claim 1 , wherein the weights of the prediction results of the two models are optimized with a backpropagation method in the step S 7 .

6 . The hydraulic fracturing method according to claim 1 , further comprising:
S 8 : modifying the active independent variable and checking predicted surface pump pressure.

7 . The hydraulic fracturing method according to claim 6 , wherein the step S 8 further comprises:
S 81 : obtaining pump rate, proppant concentration, proppant ratio and formation evaluation data of n pieces of data before a time point predicted backward, and taking mean values of the pump rate, the proppant concentration and the proppant ratio;
S 82 : setting active variable pump rate, proppant concentration and proppant ratio of m time points predicted backward as the mean values obtained in the step S 81 , remaining the formation evaluation data unchanged, and inputting into a LightGBM gradient boosting model for prediction; and
S 83 : when predicted pressure trends under different fracturing treatment parameters need to be observed, adjusting the obtained mean values and inputting the average values into the LightGBM gradient boosting model for prediction.

8 . A computer device, comprising: a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor, when executing the computer program, implements the steps of the hydraulic fracturing method according to claim 1 .

9 . A non-transitory computer-readable storage medium, having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the hydraulic fracturing method according to claim 1 .

Assignees

• Univ Southwest Petroleum

Inventors

• Zeng Fanhui
• Zhang Yu
• Du Xiaoyang
• Hu Dagan
• Guo Jianchun

Classifications

• E21B2200/22

  Fuzzy logic, artificial intelligence, neural networks or the like · CPC title

• E21B2200/20

  Computer models or simulations, e.g. for reservoirs under production, drill bits · CPC title

• E21B43/267

  reinforcing fractures by propping · CPC title

• E21B43/2607

  Surface equipment specially adapted for fracturing operations · CPC title

• Y02A10/40

  Controlling or monitoring, e.g. of flood or hurricane; Forecasting, e.g. risk assessment or mapping · CPC title