Fracturing optimization identification method and system based on microseismic signal analysis

Abstract

The invention relates to the technical field of seismic signal analysis, in particular to a fracturing optimization identification method and system based on micro-seismic signal analysis. The method comprises the following steps: acquiring historical micro-seismic event data, including historical micro-seismic waveform data and historical pressure curve data of a fracturing pump, and integrating the data to obtain an original data set; performing data preprocessing on the original data set to obtain a de-noised micro-seismic waveform and a smooth pressure curve, performing simultaneous analysis, identifying bridge plug setting signal characteristics, and determining a corresponding micro-seismic event type; and based on the bridge plug setting signal characteristics in the microseism event, identifying and acquiring the microseism event with the classification label by adopting an improved random forest model. According to the method, a hierarchical heterogeneous tree group is introduced by improving a random forest model, so that the comprehensiveness and reliability of classification are enhanced; by eliminating redundant features, generating a second-order feature interaction item and performing weighted calculation on an output probability, the recognition precision of the model on a complex micro-seismic event is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of seismic signal analysis, and in particular to a fracturing optimization identification method and system based on microseismic signal analysis. Background Art

[0002] Microseismic signal analysis is a technology that inverts the characteristics of fracture development by monitoring the tiny seismic signals caused by fracturing during oil and gas production; fracturing operation is a key project that creates fractures in the reservoir by pumping high-pressure fluid to increase oil and gas recovery rates.

[0003] Existing methods for fracturing optimization and identification based on microseismic signal analysis have the following deficiencies: bridge plug setting signal identification relies on a single waveform or pressure curve without performing joint analysis, and feature matching relies on empirical thresholds, resulting in missed detection or misjudgment of key events; redundant features are not effectively eliminated during feature processing, and there is a lack of second-order feature interaction analysis, which affects the quality of model input; classification models mostly use traditional random forests, and a single tree structure makes it difficult to take into account both local details and global associations. The classification accuracy and reliability are insufficient, making it difficult to meet the requirements of fracturing optimization for accurate identification of microseismic events. Therefore, there is an urgent need for a fracturing optimization and identification method based on microseismic signal analysis that can solve the above problems. Summary of the Invention

[0004] The purpose of the present invention is to address the problems existing in the background technology and propose a fracturing optimization identification method and system based on microseismic signal analysis.

[0005] The technical solution of the present invention is a method for optimizing and identifying fracturing based on microseismic signal analysis, comprising the following steps: S1. Obtain historical microseismic event data, including historical microseismic waveform data and historical fracturing pump pressure curve data, and integrate them to obtain an original data set; S2. Preprocess the original data set to obtain denoised microseismic waveforms and smoothed pressure curves, and perform simultaneous analysis. Based on the pressure curve, locate the fracturing interval as a candidate window for bridge plug setting. Extract waveform feature parameters within the window, match features using a historical template library, identify the bridge plug setting signal characteristics, and determine the corresponding microseismic event type. S3. Based on the characteristics of the bridge plug setting signal in the microseismic event, an improved random forest model is used to identify and obtain microseismic events with classification labels; S3 includes: S31. Performing parallel classification using a hierarchical heterogeneous tree population, including constructing a fine recognition tree and a global association tree based on a pre-assigned number of microseismic events as the number of training samples, and obtaining a classification probability output by each tree; S32. The output probability is weighted and calculated according to the ratio of the number of pre-allocated training samples to obtain the classification label of the microseismic event.

[0006] Preferably, the method of simultaneous analysis of S2 includes: S21. Identify the interval between the end of the fracturing and the start of the next fracturing according to the smoothed pressure curve, and mark the interval as a candidate window for bridge plug setting; S22. Obtain a de-noised microseismic waveform of a candidate window for bridge plug setting, calculate waveform characteristic parameters, and obtain candidate signal characteristics; wherein the waveform characteristic parameters include rise time, peak amplitude, duration, main frequency, bandwidth, polarization direction angle, and energy decay rate; S23. Obtain historical setting signals of the same type of bridge plug, establish a bridge plug setting signal library, and summarize the template features; S24. Calculate the feature matching degree between the candidate signal feature and the template feature, obtain the qualified candidate signal feature based on the feature matching degree, and mark the bridge plug setting signal; S25. Determine a microseismic event based on the bridge plug setting signal and use it as the result of the joint analysis.

[0007] Preferably, S31 includes: S311, classifying the features according to the feature type, and obtaining a list of several similar features based on the feature type; S312. In the list of similar features, all feature combinations are obtained by traversing, several feature pairs are constructed, the Spearman correlation coefficients between the feature pairs are calculated, and feature pairs with a Spearman correlation coefficient greater than a preset redundancy threshold are marked as redundant feature pairs; S313, obtaining the influence weight of the feature, sorting the two features in the redundant feature pair according to the influence weight, marking the lowest-ranked feature as the cropping feature, removing the cropping feature from the list of similar features, and updating the list of similar features; S314, determining the core features of the microseismic event based on the features in the updated similar feature columns, traversing to obtain all core feature combinations, and generating second-order core feature interaction terms; S315. Take both the interaction item and the core feature as the target item, calculate the information gain between the target item and the event type, and select the target item with high discrimination.

[0008] Preferably, S31 further includes: S316. Calculate the information gain between the target item and the microseismic event type using the following formula: ; ; ; Where H(C) is the basic information entropy; Ci is the microseismic event type; i is the event type number; i=(1,2,3); among them, C1 is the artificial fracture type; C2 is the natural fracture type; C3 is the interference type; is the prior probability of the i-th type of microseismic event; N0 is the number of historical microseismic event samples; k is the number of possible value combinations of the two core features of the target item X in the historical microseismic event samples; j is the number of possible value combinations; Nj is the number of samples when the target item X takes the j-th value among the k possible values; is the conditional entropy of the target item X under event C; The information gain between the target item and the microseismic event type is calculated as the posterior probability of the i-th type by the following formula; IG(X) is the information gain value of the target item X.

[0009] Preferably, S31 further includes: S317 , screening the information gain values ​​IG(X) of the target item X, retaining the target items whose information gain values ​​IG(X) are greater than a preset gain threshold and selecting them as high-discrimination target items; S318, through the formula Calculate the dynamic weight W(X) of the retained target item X; Where, and are the minimum and maximum information gain, respectively; G(X) is the adaptation coefficient between the target item X and the underlying lithology, which is determined based on industry experimental data; α is the balance coefficient.

[0010] Preferably, S31 further includes: S319, obtaining target item values ​​of microseismic events and corresponding microseismic event classification labels and classification accuracy, and taking target item values ​​with classification accuracy greater than a preset accuracy threshold as tree nodes; S3110. Sort the target items in descending order according to their information gain values ​​to obtain an information gain value ranking, determine the target items of each layer according to the information gain value ranking, and perform deep splitting as a fine recognition tree structure.

[0011] Preferably, S31 further includes: S3111. Traverse all target item combinations and construct several target item pairs. For each target item pair, calculate the global relevance score of the two target items in the target item pair using the following formula: ; ; Where MI is the mutual information value of the target item pair; Zh is the value of the target item pair, h is the value number of the target item pair, and h is a positive integer; Score is the global correlation score; Acc is the classification accuracy of the target item pair in historical samples; S3112. Sort the target item pairs in descending order according to their global relevance to obtain a global relevance descending order. Determine the target item pairs at each level based on the global relevance descending order and perform breadth splitting as a global relevance tree structure.

[0012] Preferably, S31 further includes: S3113. Perform node path simulation on each tree. After reaching the leaf node, count the proportion of each type of microseismic events in the historical training samples of the node and use it as the output classification probability.

[0013] The present invention also discloses a fracturing optimization identification system based on microseismic signal analysis, which applies the above-mentioned fracturing optimization identification method based on microseismic signal analysis, specifically comprising: The data acquisition module is used to obtain historical microseismic event data, including historical microseismic waveform data and historical fracturing pump pressure curve data, and integrate them to obtain the original data set; The data processing and analysis module is used to pre-process the original data set to obtain the denoised microseismic waveform and smoothed pressure curve, perform joint analysis, identify the characteristics of the bridge plug setting signal, and determine the corresponding microseismic event type; The signal recognition and event output module is used to identify and obtain microseismic events with classification labels based on the bridge plug setting signal characteristics in microseismic events using an improved random forest model.

[0014] Compared with the prior art, the above technical solution of the present invention has the following beneficial technical effects: The present invention uses a joint analysis of pressure curves and microseismic waveforms, combined with feature matching criteria, to accurately identify bridge plug setting signals, thereby improving the pertinence of key event feature extraction; by introducing a hierarchical heterogeneous tree group through an improved random forest model, parallel classification is achieved by constructing a fine recognition tree based on deep splitting of information gain and a global correlation tree based on global correlation breadth splitting of mutual information and accuracy, thereby enhancing the comprehensiveness and reliability of classification; by eliminating redundant features through the Spearman correlation coefficient, generating second-order feature interaction terms and weightedly calculating the output probability, the model's recognition accuracy for complex microseismic events is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 This is a flow chart of the first embodiment of the present invention. DETAILED DESCRIPTION

[0016] Example 1, as Figure 1As shown, the present invention proposes a fracturing optimization identification method based on microseismic signal analysis, which includes the following steps: S1. Obtain historical microseismic waveform data, historical fracturing pump pressure curve data, and classification labels and classification accuracy of microseismic types, and integrate them to obtain an original data set; wherein the microseismic waveform can be obtained by high-frequency sampling and recording of a microseismic acquisition instrument; S2. Preprocess the original data set to obtain denoised microseismic waveforms and smoothed pressure curves, perform joint analysis, identify the characteristics of the bridge plug setting signal, and determine the corresponding microseismic event type; The steps for data preprocessing of the original dataset include: Taking the start time of the fracturing pump as the standard time, all data timestamps are uniformly converted into the time difference relative to the standard time; The data with different sampling rates in the original dataset are resampled to the same time grid using cubic spline interpolation; The correlation between the pressure curve and the microseismic signal is used to eliminate interference signals without pressure response. The wavelet threshold denoising algorithm is used to denoise the microseismic waveform to obtain a processed data set, including the denoised microseismic waveform and the smoothed pressure curve. The methods for simultaneous analysis of S2 include: S21. Identify the interval between the end of the fracturing and the start of the next fracturing according to the smoothed pressure curve, and mark the interval as a candidate window for bridge plug setting; S22. Obtain a de-noised microseismic waveform of a candidate window for bridge plug setting, calculate waveform characteristic parameters, and obtain candidate signal characteristics; wherein the waveform characteristic parameters include rise time, peak amplitude, duration, main frequency, bandwidth, polarization direction angle, and energy decay rate; S23. Obtain historical setting signals of the same type of bridge plug, establish a bridge plug setting signal library, and summarize the template features; S24. Calculate the feature matching degree between the candidate signal feature and the template feature, obtain the qualified candidate signal feature based on the feature matching degree, and mark the bridge plug setting signal; wherein the feature matching degree is obtained by calculating the cosine similarity method, and when the cosine similarity is ≥0.85, it is determined that the feature matching degree meets the standard; Eliminate interference signals from the bridge plug setting signal, including interference signals from fracturing equipment operation and random activation signals from natural fractures. For example, interference signals from fracturing equipment operation, such as pump truck start and stop signals, last for 5-10 seconds and have no sudden pressure rise characteristics. Random activation signals from natural fractures have no fixed time window, and the pressure curve does not rise synchronously, with dispersed energy. S25, determining a microseismic event based on the bridge plug setting signal and using it as a joint analysis result; S3. Based on the characteristics of the bridge plug setting signal in the microseismic event, an improved random forest model is used to identify and obtain microseismic events with classification labels. Specifically, the following steps are included: S31. A hierarchical heterogeneous tree group is used to perform parallel classification, including constructing a fine recognition tree and a global association tree based on the pre-assigned number of microseismic events as the number of training samples, and obtaining the classification probability output by each tree. Specifically, the following steps are included: S311, classifying the features according to the feature type, and obtaining a list of several similar features based on the feature type; S312. In the list of similar features, all feature combinations are obtained by traversing, a number of feature pairs are constructed, the Spearman correlation coefficients between the feature pairs are calculated, and feature pairs whose Spearman correlation coefficients are greater than a preset redundancy threshold are marked as redundant feature pairs; wherein the preset redundancy threshold is determined based on historical microseismic big data; S313. Obtaining the influence weight of the feature, sorting the two features in the redundant feature pair by the influence weight, marking the lowest-ranked feature as a clipping feature, removing the clipping feature from a list of similar features, and updating the list of similar features; wherein the influence weight of the feature can be obtained by fitting the recognition accuracy of the feature and the microseismic event in historical microseismic events, or can also be obtained based on the importance score statistics of experts; S314, determining the core features of the microseismic event based on the features in the updated similar feature columns, traversing to obtain all core feature combinations, and generating second-order core feature interaction terms; S315. Take both the interaction item and the core feature as the target item, calculate the information gain between the target item and the event type, and select the target item with high discrimination; S316. Calculate the information gain between the target item and the microseismic event type using the following formula: ; ; ; Where H(C) is the basic information entropy; Ci is the microseismic event type; i is the event type number; i=(1,2,3); among them, C1 is the artificial fracture type; C2 is the natural fracture type; C3 is the interference type; is the prior probability of the i-th type of microseismic event; N0 is the number of historical microseismic event samples; k is the number of possible value combinations of the two core features of the target item X in the historical microseismic event samples; j is the number of possible value combinations; Nj is the number of samples when the target item X takes the j-th value among the k possible values; is the conditional entropy of the target item X under event C; The information gain between the target item and the microseismic event type is calculated as the posterior probability of the i-th type by the following formula; IG(X) is the information gain value of the target item X; the larger the information gain value IG(X), the stronger the effect of the target item X on reducing the uncertainty of event classification; S317. Filter the information gain values ​​IG(X) of the target item X, and retain the target items whose information gain values ​​IG(X) are greater than a preset gain threshold and select them as high-discrimination target items; wherein the preset gain threshold is obtained based on historical screening big data training; S318, through the formula Calculate the dynamic weight W(X) of the high-discrimination target item X; Where, and are the minimum and maximum information gain values, respectively; G(X) is the adaptation coefficient between the target item X and the underlying lithology, determined based on industry experimental data; α is the balance coefficient, used to prioritize the dominant role of the information gain value; S319, obtaining target item values ​​of microseismic events and corresponding microseismic event classification labels and classification accuracy, and taking target item values ​​with classification accuracy greater than a preset accuracy threshold as tree nodes; wherein the preset accuracy threshold may be 0.85; S3110, sorting the target items in descending order according to their information gain values ​​to obtain an information gain value ranking, determining the target items of each layer according to the information gain value ranking and performing deep splitting as a refined recognition tree structure; For example, the first layer splitting is performed by using whether the main frequency is greater than 300 Hz as a node to preliminarily separate high-main-frequency events such as artificial cracks from low-main-frequency events such as natural cracks or interference; Second layer splitting: Layer 2: Split based on whether the polarization angle is <30° to further focus on the directional characteristics of the artificial seam; The third level splitting: splitting based on whether the amplitude decay rate is greater than 0.7, distinguishing between the rapid amplitude decay of natural cracks and the irregular amplitude decay of interference signals; S3111. Traverse all target item combinations and construct several target item pairs. For each target item pair, calculate the global relevance score of the two target items in the target item pair using the following formula: ; ; Where MI is the mutual information value of the target item pair; Zh is the value of the target item pair, h is the value number of the target item pair, and h is a positive integer; Score is the global correlation score; Acc is the classification accuracy of the target item pair in historical samples; S3112. Sort the target item pairs in descending order of global relevance to obtain a global relevance descending order, determine the target item pairs of each layer based on the global relevance descending order, and perform breadth splitting as a global relevance tree structure; S3113. Perform node path simulation on each tree. After reaching the leaf node, count the proportion of each type of microseismic events in the historical training samples of the node and use it as the output classification probability.

[0017] S32. The output probability is weighted and calculated according to the ratio of the number of pre-allocated training samples to obtain the classification label of the microseismic event; illustratively, the ratio of the number of training samples used to construct the fine recognition tree and the global association tree is 6 to 4, and the classification probabilities finally output by the two trees are weighted by 0.6 and 0.4 respectively.

[0018] By jointly analyzing the pressure curve and microseismic waveform and combining it with the feature matching criterion of cosine similarity ≥ 0.85, the bridge plug setting signal is accurately identified, thereby improving the pertinence of key event feature extraction. By improving the random forest model and introducing hierarchical heterogeneous tree groups, parallel classification is achieved by constructing a fine recognition tree based on deep splitting of information gain and a global correlation tree based on global correlation breadth splitting of mutual information and accuracy, thereby enhancing the comprehensiveness and reliability of the classification. The Spearman correlation coefficient is used to eliminate redundant features, generate second-order feature interaction terms, and calculate the output probability in a weighted manner, thereby improving the model's recognition accuracy for complex microseismic events.

[0019] In a second embodiment, the present invention proposes a fracturing optimization identification system based on microseismic signal analysis, which is applied to the fracturing optimization identification method based on microseismic signal analysis proposed in the first embodiment, and specifically includes: The data acquisition module is used to obtain historical microseismic event data, including historical microseismic waveform data and historical fracturing pump pressure curve data, and integrate them to obtain the original data set; The data processing and analysis module is used to pre-process the original data set to obtain the denoised microseismic waveform and smoothed pressure curve, perform joint analysis, identify the characteristics of the bridge plug setting signal, and determine the corresponding microseismic event type; The signal recognition and event output module is used to identify and obtain microseismic events with classification labels based on the bridge plug setting signal characteristics in microseismic events using an improved random forest model.

[0020] The embodiments of the present invention are described in detail above with reference to the accompanying drawings, but the present invention is not limited thereto. Various changes can be made within the scope of knowledge possessed by those skilled in the art without departing from the spirit of the present invention.

Claims

1. A fracturing optimization identification method based on microseismic signal analysis, characterized in that: The following steps are involved: S1. Obtain historical microseismic event data, including historical microseismic waveform data and historical fracturing pump pressure curve data, and integrate them to obtain an original data set; S2. Preprocess the original data set to obtain denoised microseismic waveforms and smoothed pressure curves, and perform simultaneous analysis. Based on the pressure curve, locate the fracturing interval as a candidate window for bridge plug setting. Extract waveform feature parameters within the window, match features using a historical template library, identify the bridge plug setting signal characteristics, and determine the corresponding microseismic event type. S3. Based on the characteristics of the bridge plug setting signal in the microseismic event, an improved random forest model is used to identify and obtain microseismic events with classification labels; S3 includes: S31. Performing parallel classification using a hierarchical heterogeneous tree population, including constructing a fine recognition tree and a global association tree based on a pre-assigned number of microseismic events as the number of training samples, and obtaining a classification probability output by each tree; S32. The output probability is weighted and calculated according to the ratio of the number of pre-allocated training samples to obtain the classification label of the microseismic event.

2. The fracturing optimization identification method based on microseismic signal analysis according to claim 1, characterized in that: The methods for simultaneous analysis of S2 include: S21. Identify the interval between the end of the fracturing and the start of the next fracturing according to the smoothed pressure curve, and mark the interval as a candidate window for bridge plug setting; S22. Obtain a de-noised microseismic waveform of a candidate window for bridge plug setting, calculate waveform characteristic parameters, and obtain candidate signal characteristics; wherein the waveform characteristic parameters include rise time, peak amplitude, duration, main frequency, bandwidth, polarization direction angle, and energy decay rate; S23. Obtain historical setting signals of the same type of bridge plug, establish a bridge plug setting signal library, and summarize the template features; S24. Calculate the feature matching degree between the candidate signal feature and the template feature, obtain the qualified candidate signal feature based on the feature matching degree, and mark the bridge plug setting signal; S25. Determine a microseismic event based on the bridge plug setting signal and use it as the result of the joint analysis.

3. The method for fracturing optimization and identification based on microseismic signal analysis according to claim 2, characterized in that: S31 includes: S311, classifying the features according to the feature type, and obtaining a list of several similar features based on the feature type; S312. In the list of similar features, all feature combinations are obtained by traversing, several feature pairs are constructed, the Spearman correlation coefficients between the feature pairs are calculated, and feature pairs with a Spearman correlation coefficient greater than a preset redundancy threshold are marked as redundant feature pairs; S313, obtaining the influence weight of the feature, sorting the two features in the redundant feature pair according to the influence weight, marking the lowest-ranked feature as the cropping feature, removing the cropping feature from the list of similar features, and updating the list of similar features; S314, determining the core features of the microseismic event based on the features in the updated similar feature columns, traversing to obtain all core feature combinations, and generating second-order core feature interaction terms; S315. Take both the interaction item and the core feature as the target item, calculate the information gain between the target item and the event type, and select the target item with high discrimination.

4. The method for fracturing optimization and identification based on microseismic signal analysis according to claim 3, characterized in that: The S31 also includes: S316. Calculate the information gain between the target item and the microseismic event type using the following formula: ； ； ； Where H(C) is the basic information entropy; Ci is the microseismic event type; i is the event type number; i=(1,2,3); among them, C1 is the artificial fracture type; C2 is the natural fracture type; C3 is the interference type; is the prior probability of the i-th type of microseismic event; N0 is the number of historical microseismic event samples; k is the number of possible value combinations of the two core features of the target item X in the historical microseismic event samples; j is the number of possible value combinations; Nj is the number of samples when the target item X takes the j-th value among the k possible values; is the conditional entropy of the target item X under event C; The information gain between the target item and the microseismic event type is calculated as the posterior probability of the i-th type by the following formula; IG(X) is the information gain value of the target item X.

5. The method for fracturing optimization and identification based on microseismic signal analysis according to claim 4, characterized in that: The S31 also includes: S317 , screening the information gain values ​​IG(X) of the target item X, retaining the target items whose information gain values ​​IG(X) are greater than a preset gain threshold and selecting them as high-discrimination target items; S318, through the formula Calculate the dynamic weight W(X) of the retained target item X; Where, and are the minimum and maximum information gain, respectively; G(X) is the adaptation coefficient between the target item X and the underlying lithology, which is determined based on industry experimental data; α is the balance coefficient.

6. The method for fracturing optimization and identification based on microseismic signal analysis according to claim 5, characterized in that: The S31 also includes: S319, obtaining target item values ​​of microseismic events and corresponding microseismic event classification labels and classification accuracy, and taking target item values ​​with classification accuracy greater than a preset accuracy threshold as tree nodes; S3110. Sort the target items in descending order according to their information gain values ​​to obtain an information gain value ranking, determine the target items of each layer according to the information gain value ranking, and perform deep splitting as a fine recognition tree structure.

7. The method for fracturing optimization and identification based on microseismic signal analysis according to claim 6, characterized in that: The S31 also includes: S3111. Traverse all target item combinations and construct several target item pairs. For each target item pair, calculate the global relevance score of the two target items in the target item pair using the following formula: ； ； Where MI is the mutual information value of the target item pair; Zh is the value of the target item pair, h is the value number of the target item pair, and h is a positive integer; Score is the global correlation score; Acc is the classification accuracy of the target item pair in historical samples; S3112. Sort the target item pairs in descending order according to their global relevance to obtain a global relevance descending order. Determine the target item pairs at each level based on the global relevance descending order and perform breadth splitting as a global relevance tree structure.

8. The method for fracturing optimization and identification based on microseismic signal analysis according to claim 7, characterized in that: The S31 also includes: S3113. Perform node path simulation on each tree. After reaching the leaf node, count the proportion of each type of microseismic events in the historical training samples of the node and use it as the output classification probability.

9. A fracturing optimization identification system based on microseismic signal analysis, applied to the fracturing optimization identification method based on microseismic signal analysis according to any one of claims 1 to 8, characterized in that: Specifically include: The data acquisition module is used to obtain historical microseismic event data, including historical microseismic waveform data and historical fracturing pump pressure curve data, and integrate them to obtain the original data set; The data processing and analysis module is used to pre-process the original data set to obtain the denoised microseismic waveform and smoothed pressure curve, perform joint analysis, identify the characteristics of the bridge plug setting signal, and determine the corresponding microseismic event type; The signal recognition and event output module is used to identify and obtain microseismic events with classification labels based on the bridge plug setting signal characteristics in microseismic events using an improved random forest model.

Images