Fractured rock mass recognition method, device and equipment based on drilling data and medium

Abstract

The invention provides a fractured rock mass recognition method and device based on drilling data, equipment and a medium, and the method comprises the steps: obtaining the drilling data of a rock mass, and constructing a relational graph based on the drilling data of the rock mass; a pre-configured neighborhood radius and a minimum point number are obtained, the neighborhood radius is determined by a slope inflection point in a k distance map generated based on the historical drilling data, and the minimum point number is determined by an identification evaluation index obtained by carrying out region limitation based on the determined neighborhood radius and carrying out fractured rock mass identification on the historical drilling data; and on the basis of the neighborhood radius and the minimum point number, data clustering processing is carried out on the drilling data of the rock mass in a preset relational graph, an abnormal data cluster is determined, and the abnormal data cluster represents regional information of the fractured rock mass in the preset relational graph. According to the method, clustering of any shape can be processed, the recognition capability of fractured rock masses with diversified forms is enhanced, and more accurate basic data is provided for engineering application.

Description

Technical Field

[0001] The present invention relates to the technical field of underground exploration, and particularly to a method, device, equipment and medium for identifying fractured rock masses based on drilling data. Background Art

[0002] At present, the identification techniques for fractured rock masses mainly include on-site testing, theoretical derivation, model identification with supervised learning, and model identification with unsupervised learning. On-site testing requires a series of complex processes such as core sampling (the process of taking rock cores) and laboratory tests. Moreover, extreme care is needed during the core sampling process, as there is a high probability of damaging the rock cores, and the real-time performance is poor, making it difficult to guide the actual drilling process. Theoretical derivation means inferring the possible locations of fractured rock masses by setting relevant thresholds based on experience.

[0003] The supervised learning method requires pre-setting labels, dividing the borehole data into two categories: fractured rock masses and intact rock masses, and then identifying them by training a regression model. This method depends on the accuracy and integrity of the labels and is easily affected by noise and data distribution.

[0004] The unsupervised learning method requires pre-setting the number of clusters and is easily affected by noise and initial values (such as easily misidentifying fractured rock masses as noise), resulting in unstable clustering results.

[0005] It can be seen that the above methods all lack the ability to accurately identify fractured rock masses and cannot provide more accurate basic data for engineering applications. Summary of the Invention

[0006] Aiming at the problems existing in the prior art, the present invention provides a method, device, equipment and medium for identifying fractured rock masses based on drilling data.

[0007] The present invention provides a method for identifying fractured rock masses based on drilling data, including: Obtaining the drilling data of the rock mass and constructing a relationship graph based on the drilling data of the rock mass; Obtaining a pre-configured neighborhood radius and minimum number of points, where the neighborhood radius is determined by the slope inflection point in the k-distance graph generated based on historical drilling data, and the minimum number of points is determined by the identification evaluation index obtained by performing fractured rock mass identification on the historical drilling data with regional limitation based on the determined neighborhood radius; Based on the neighborhood radius and the minimum number of points, performing data clustering processing on the drilling data of the rock mass in a preset relationship graph to determine abnormal data clusters, and the abnormal data clusters represent the regional information of the fractured rock mass in the preset relationship graph.

[0008] A method for identifying fractured rock masses based on drilling data provided by the present invention. The drilling data of the rock mass includes drilling depth, drilling pressure, rotational speed, torque, and drilling speed. Accordingly, a drilling speed normalization parameter is determined based on the drilling pressure, rotational speed, torque, and drilling speed, and a relationship graph is constructed based on the drilling depth and the drilling speed normalization parameter.

[0009] A method for identifying fractured rock masses based on drilling data provided by the present invention. The step of determining the drilling speed normalization parameter based on the drilling pressure, rotational speed, torque, and drilling speed includes: According to the drilling pressure, rotational speed, torque, and drilling speed, the following calculation formula is used to determine the drilling speed normalization parameter; Wherein, is the drilling speed normalization parameter, is the drilling speed, is the drilling pressure, is the rotational speed, is the torque, , , and are preset coefficients.

[0010] A method for identifying fractured rock masses based on drilling data provided by the present invention. The method further includes: Judging the regional continuity of the abnormal data cluster. When it is determined that the abnormal data cluster has continuity characteristics, identifying the central position of the abnormal data cluster and determining the continuous distance of the abnormal data cluster; when it is determined that the abnormal data cluster does not have continuity characteristics, re-dividing the abnormal data cluster to determine multiple abnormal data sub-clusters, identifying the central position of the abnormal data sub-cluster and determining the continuous distance of the abnormal data sub-cluster; wherein, the continuous distance represents the continuous depth of the fractured rock mass.

[0011] A method for identifying fractured rock masses based on drilling data provided by the present invention. The method further includes a step of obtaining the neighborhood radius, including: Calculating the distance from each data point in the historical drilling data to its nearest k neighbors, and constructing a k-distance graph; Determining the inflection point region in the k-distance graph; Determining the inflection point with the largest slope in the inflection point region, and determining multiple points near the inflection point based on the inflection point with the largest slope, and obtaining the slopes corresponding to the points near the inflection point; Based on the determined slopes, performing data clustering processing on the historical drilling data set to obtain a clustering result, and determining the best slope as the neighborhood radius based on the clustering.

[0012] A method for identifying fractured rock masses based on drilling data provided by the present invention, the method further includes a step of obtaining the minimum number of points, including: Under the regional limitation of the determined neighborhood radius, a traversal method is adopted with the initial number of points as the starting value, and the number of points is gradually increased according to a preset increment; different data clusters are formed based on different numbers of points; Based on different data clusters, evaluation indexes corresponding to different numbers of points are determined; Based on the evaluation indexes of different numbers of points, the continuous times of consistent evaluation indexes are counted. When the continuous times reach the preset times, based on the visualization results of the data clusters corresponding to the continuously consistent evaluation indexes, one of the numbers of points is determined as the minimum number of points.

[0013] A method for identifying fractured rock masses based on drilling data provided by the present invention, before distributing the drilling data of the rock mass in a preset relationship graph, the method further includes: Based on the obtained drilling data, effective drilling process data is identified, outliers are removed, and drilling data is spliced.

[0014] The present invention also provides a device for identifying fractured rock masses based on drilling data, including: A collection module for obtaining the drilling data of the rock mass and constructing a relationship graph based on the drilling data of the rock mass; An acquisition module for obtaining a pre-configured neighborhood radius and the minimum number of points, wherein the neighborhood radius is determined by the slope inflection point in the k-distance graph generated based on historical drilling data, and the minimum number of points is determined by the recognition evaluation indexes obtained by performing fractured rock mass identification on the historical drilling data under the regional limitation based on the determined neighborhood radius; A recognition module for performing data clustering processing on the drilling data of the rock mass in a preset relationship graph based on the neighborhood radius and the minimum number of points, and determining abnormal data clusters, where the abnormal data clusters represent the regional information of the fractured rock mass in the preset relationship graph.

[0015] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, and when the processor executes the program, it implements any one of the above-mentioned methods for identifying fractured rock masses based on drilling data.

[0016] The present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements any one of the above-mentioned methods for identifying fractured rock masses based on drilling data.

[0017] The present invention also provides a computer program product, including a computer program which, when executed by a processor, implements any one of the above-mentioned fracture rock mass identification methods based on drilling data.

[0018] A fracture rock mass identification method, device, equipment and medium based on drilling data provided by the present invention, by reasonably optimizing the neighborhood radius and the minimum number of points, and then based on the neighborhood radius and the minimum number of points, performing data clustering processing on the drilling data of the rock mass in the relationship graph to determine abnormal data clusters, can handle clusters of any shape, enhance the identification ability of fracture rock masses with diverse morphologies, provide more accurate basic data for engineering applications, not only improve the accuracy of identification, but also effectively reduce the dependence on manual intervention and reduce the risk of human errors. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0020] Figure 1 is a schematic flowchart of the fracture rock mass identification method based on drilling data provided by the present invention.

[0021] Figure 2 is a relationship graph between the drilling depth and the normalized parameters of the drilling speed provided by the present invention.

[0022] Figure 3 is an identification result graph of the fracture rock mass identification method based on drilling data provided by the present invention.

[0023] Figure 4 is a K-distance graph for determining the neighborhood radius provided by the present invention.

[0024] Figure 5a is the identification result based on the neighborhood radius provided by the present invention Figure 1 .

[0025] Figure 5b is the identification result based on the neighborhood radius provided by the present invention Figure 2 .

[0026] Figure 5c is the identification result based on the neighborhood radius provided by the present invention Figure 3 .

[0027] Figure 5d is the identification result based on the neighborhood radius provided by the present invention Figure 4 .

[0028] Figure 6 It is a diagram showing the identification results of SS and CH values of fractured rock masses under different minimum point numbers provided by the present invention.

[0029] Figure 7a It is the evaluation result under different minimum point numbers in the same neighborhood provided by the present invention Figure 1 .

[0030] Figure 7b It is the evaluation result under different minimum point numbers in the same neighborhood provided by the present invention Figure 2 .

[0031] Figure 7c It is the evaluation result under different minimum point numbers in the same neighborhood provided by the present invention Figure 3 .

[0032] Figure 8 It is a schematic structural diagram of a fractured rock mass identification device based on drilling data provided by the present invention.

[0033] Figure 9 It is a schematic structural diagram of an electronic device provided by the present invention. Detailed implementation manners

[0034] To make the objectives, technical solutions and advantages of the present invention clearer, the technical solutions in the present invention will be clearly and completely described below with reference to the accompanying drawings in the present invention. Obviously, the described embodiments are some but not all of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present invention without making creative efforts fall within the scope of protection of the present invention.

[0035] The following will describe Figures 1 - 9 the fractured rock mass identification method, device, equipment and medium based on drilling data of the present invention.

[0036] Figure 1 Fig. shows a schematic flow chart of a fractured rock mass identification method based on drilling data provided by the present invention. Refer to Figure 1 , and the method includes the following steps: Step 11: Obtain the drilling data of the rock mass, and construct a relationship diagram based on the drilling data of the rock mass.

[0037] Step 12: Obtain a pre-configured neighborhood radius and minimum point number. Among them, the neighborhood radius is determined by the slope inflection point in the k-distance diagram generated based on historical drilling data, and the minimum point number is determined by the identification evaluation index obtained by performing fractured rock mass identification on the historical drilling data with regional limitation based on the determined neighborhood radius.

[0038] Step 13: Based on the neighborhood radius and the minimum number of points, perform data clustering on the drilling data of the rock mass in the relationship graph to determine abnormal data clusters, where the abnormal data clusters represent the regional information of the fractured rock mass in the preset relationship graph.

[0039] Regarding Steps 11 to 13, it should be noted that a large number of sensors such as displacement, hydraulic pressure, and rotational speed are installed on the geological drill to collect the drilling data in real time during the drilling process of the rock mass by the drill. Due to different types of sensors, there are many types of data types of the obtained drilling data. For example, drilling depth, drilling pressure, rotational speed, torque, and drilling speed are not listed one by one here.

[0040] In the present invention, for the intuitive display of the drilling data and for the intuitive observation of the fractured area of the rock mass, the drilling data is presented in the form of data points in a drawing. Specifically, a relationship graph can be constructed based on the drilling data of the rock mass. This relationship graph can represent the rock mass conditions encountered by the drill during the drilling process of the rock mass, such as at what drilling depth the fractured rock mass is found. Therefore, the relationship graph of the present invention can establish the corresponding relationship between the drilling data and the fractured rock mass conditions.

[0041] In the present invention, since the drilling data is reflected in the relationship graph in the form of data points. Therefore, clustering processing needs to be performed on the data points in the relationship graph to identify the area that can represent the fractured rock mass.

[0042] At present, the identification technologies for fractured rock masses mainly include on-site testing, theoretical derivation, model identification of supervised learning, and model identification of unsupervised learning. On-site testing requires a series of complex processes such as core sampling (the process of taking rock cores) and laboratory tests. Moreover, extreme care is needed during the core sampling process, and it is very likely to damage the rock core, and the real-time performance is not good, making it difficult to guide the actual drilling process. Theoretical derivation is to infer the possible location of the fractured rock mass by setting relevant thresholds through experience.

[0043] The supervised learning method needs to preset labels, divide the borehole data into two categories: fractured rock mass and intact rock mass, and then identify through training a regression model. This method depends on the accuracy and integrity of the labels and is easily affected by noise and data distribution.

[0044] The unsupervised learning method needs to preset the number of clusters and is easily affected by noise and initial values (such as easily misidentifying the fractured rock mass as noise), resulting in unstable clustering results.

[0045] It can be seen from this that the above methods all lack the ability to accurately identify the fractured rock mass and cannot provide more accurate basic data for engineering applications.

[0046] Therefore, in the identification of fractured rock masses, a clustering algorithm (such as the DBSCAN algorithm) is used to calculate the identification results independently, so that the identification results of fractured rock masses can be more accurate, effectively reducing human errors and the amount of calculation. During the clustering process, by fitting the actual adverse conditions brought by the fractured rock masses, a more reasonable neighborhood radius and minimum number of points are optimized to improve the accuracy of the clustering results.

[0047] Therefore, the present invention does not need to preset the number of clusters, has strong adaptability to data, and is suitable for rock mass data with complex shapes and distribution characteristics. The clustering algorithm of the present invention forms clusters based on the density of data points, can divide high-density regions into clusters, and mark low-density regions as noise. This characteristic enables the clustering algorithm to effectively distinguish dense fractured rock masses from relatively sparse intact rock masses.

[0048] There are two key parameters in the clustering algorithm, namely eps (neighborhood radius) and minPts (minimum number of points required to form a cluster). The algorithm automatically detects fractured rock masses by finding the data center points with higher density and forming different clusters. Therefore, the optimized neighborhood radius of the present invention is determined by the slope inflection point in the k-distance graph generated based on historical drilling data, and the minimum number of points is determined by the regional limitation based on the determined neighborhood radius and the identification evaluation index obtained by identifying fractured rock masses from historical drilling data.

[0049] Based on the neighborhood radius and the minimum number of points, the present invention performs data clustering processing on the drilling data of rock masses in a preset relationship graph to determine abnormal data clusters, and the abnormal data clusters represent the regional information of fractured rock masses in the preset relationship graph.

[0050] The specific implementation steps of the clustering process of the present invention are as follows: 1. Input data: Input data set: Input or import the rock mass data to be identified. Input structure parameters: Input the confirmed optimal eps value (i.e., neighborhood radius) to define the influence range of each point; input the confirmed optimal minPts value to define the minimum number of points required for a cluster.

[0051] 2. Algorithm execution 1) Initialization: Mark all points as unprocessed and create an empty list to store clusters.

[0052] 2) Traverse data points: Randomly select an unprocessed point P . Mark the point P as processed. With the point P as the center and eps as the radius, find all points within the P neighborhood of the point ε , which is called Nε ( P ).

[0053] 3) Judgment key point: If Nε ( P ) the number of points in is greater than or equal to minPts, then the point P is a core point. Create a new cluster C , and add the point P and Nε ( P ) all points in to the cluster C .

[0054] 4) Expand the cluster: For each unprocessed point C in the cluster Q , perform the following operations: a) Mark the point Q as processed.

[0055] b) Find all points within the Q neighborhood of the point, called ε ( Nε ). Q )

[0056] c) If Nε ( Q ) the number of points in is greater than or equal to minPts, then add the points in Nε ( Q ) to the cluster C .

[0057] d) Repeat the above process until the cluster C no longer grows.

[0058] 5) Process noise points: If Nε ( P ) the number of points in is less than minPts, then the point P is marked as a noise point (outlier).

[0059] 6) Repeat steps 2 - 5: Until all points are processed.

[0060] The method for identifying fractured rock masses based on drilling data provided by the present invention, by reasonably optimizing the neighborhood radius and the minimum number of points, and then based on the neighborhood radius and the minimum number of points, performing data clustering processing on the drilling data of rock masses in the relationship graph to determine abnormal data clusters, can handle clusters of any shape, enhance the ability to identify fractured rock masses with diverse morphologies, provide more accurate basic data for engineering applications, not only improve the accuracy of identification, but also effectively reduce the dependence on manual intervention and reduce the risk of human errors.

[0061] In a further method of the above method, the process of constructing the relationship diagram is mainly explained. The drilling data of the rock mass includes drilling depth, drilling pressure, rotational speed, torque, and drilling speed. Correspondingly, the drilling speed normalization parameter is determined according to the drilling pressure, rotational speed, torque, and drilling speed, and the relationship diagram is constructed based on the drilling depth and the drilling speed normalization parameter.

[0062] In the present invention, during the drilling process of the drill rig on the rock mass, a large amount of drilling data is generated, and there are many types of data. Therefore, the required drilling data can be obtained based on specific usage requirements. For example, the required drilling data can include drilling depth, drilling pressure, rotational speed, torque, and drilling speed. Since the position of the fissure area in the rock mass needs to be determined, it is necessary to reasonably analyze the drill rig data and the relationship between the drilling depth and other types of drilling data. When the drilling data of the rock mass includes drilling depth, drilling pressure, rotational speed, torque, and drilling speed, correspondingly, the drilling speed normalization parameter is determined according to the drilling pressure, rotational speed, torque, and drilling speed, and the relationship diagram is constructed based on the drilling depth and the drilling speed normalization parameter. For details, see Figure 2 。

[0063] During the drilling process, the drilling speed is used as the standard for evaluating the rock mass parameters. However, when the drilling pressure and the drilling rotational speed change, the discreteness of the drilling speed is very large. It is inappropriate to use only the drilling speed as the sole standard for evaluating the rock mass parameters. Based on this, in order to eliminate the influence of the drilling pressure F, rotational speed N, and torque M on the drilling speed V, a drilling speed normalization parameter is proposed. This index is unique in homogeneous materials and is as follows: According to the drilling pressure, rotational speed, torque, and drilling speed, the following calculation formula is used to determine the drilling speed normalization parameter; Among them, is the drilling speed normalization parameter, is the drilling speed, is the drilling pressure, is the rotational speed, is the torque, 、 、 and are preset coefficients, which reflect the influence weights of various factors on the drilling speed. 、 、 and are coefficients determined through preliminary calibration tests.

[0064] The process of the preliminary calibration test is as follows: 1. Select a homogeneous rock mass sample: First, a rock mass sample with homogeneous characteristics needs to be selected to ensure the reliability of the test results.

[0065] 2. Determine the drilling parameters: Set different drilling parameters, including the drilling pressure F, rotational speed N, and torque M. The changes in these parameters will be used to analyze their effects on the drilling speed V.

[0066] 3. Conduct the drilling test: Use different combinations of drilling parameters to conduct the drilling test and record the drilling speed V in each test.

[0067] 4. Data collection and analysis: Collect all test data, including the drilling pressure F, rotational speed N, torque M, and drilling speed V. Use statistical methods and data fitting techniques (such as multiple linear regression, non - linear regression, etc.) to analyze the data and determine the relationship between each parameter and the drilling speed. Through the functional relationship obtained by fitting, determine the constant coefficients in the formula. For example, the data of the drilling pressure F and the drilling speed V can be fitted using a functional relationship. At this time, the value of the constant coefficient b is - 0.5, which is the negative value of the coefficient in the functional relationship. By analogy, the values of the constant coefficients b, c, d, etc. can be obtained. Then multiply the coefficients similar to 2.53 in all functional relationships to obtain the value of the constant coefficient a. Through the above steps, the drilling characteristics of the homogeneous rock mass under different drilling parameters can be obtained, and the corresponding parameter model can be established to obtain the V normalized parameter expression formula of the drilling speed.

[0068] After obtaining the drilling data during the drilling process, through normalization, the effects of the drilling pressure, rotational speed, and torque on the drilling speed are eliminated, making the evaluation results more comparable. In homogeneous materials, this index is unique, making the evaluation of rock mass parameters more accurate.

[0069] In a further method of the above method, abnormal data clusters are those clusters that account for a very small proportion in the dataset. The cluster centers of these clusters deviate significantly from the distribution of the overall data, indicating that they may represent the characteristics of fractured rock masses. This deviation is because the physical properties of fractured rock masses are significantly different from those of homogeneous rock masses, resulting in a sudden change in the normalized drilling speed.

[0070] For these abnormal data clusters, it is also necessary to further analyze the characteristics of the data points within the clusters. Fractured rock masses usually show a significant increase in the drilling speed because the resistance encountered by the drill bit when passing through the fractures decreases; at the same time, the pressure and torque may show abnormal decreases because the rock in the fractures is more fragmented and easier to drill through.

[0071] Further analyze the characteristics of data points within the clusters. For example, the central position of the fractured rock mass can be determined. To determine the central position of the fractured rock mass, the center point of each abnormal cluster can be calculated. This center point is determined by taking the average of all core points within the cluster, which roughly corresponds to the central position of the fractured rock mass.

[0072] However, during the process of determining the cluster center, fractured rock masses at different depths may occur. Although they are discontinuous, they are assigned to the same cluster because of similar data forms. If not processed, it will directly affect the determination of the central positions of fractured rock masses at different depths. Therefore, after analyzing all fractured rock masses, it is necessary to check whether the fractured rock masses within the same cluster are continuously distributed in depth. So, judge the regional continuity of the abnormal data clusters. When it is determined that the abnormal data clusters have continuity characteristics, identify the central position of the abnormal data clusters and determine the continuous distance of the abnormal data clusters. See Figure 3 , in Figure 3 , the normal data clusters (Cluster 1) and abnormal data clusters (Cluster 2 and Cluster 3) can be identified. For example, for the abnormal data clusters, the minimum depth and the maximum depth can be used to calculate the continuous distance of the abnormal data cluster (Cluster 2).

[0073] When it is determined that the abnormal data clusters do not have continuity characteristics, re-divide the abnormal data clusters to determine multiple abnormal data sub-clusters, identify the central positions of the abnormal data sub-clusters and determine the continuous distances of the abnormal data sub-clusters; among them, the continuous distance characterizes the continuous depth of the fractured rock mass. Continue to see Figure 3 , if the abnormal data cluster (Cluster 2) has discontinuity characteristics, the abnormal data cluster (Cluster 2) can be re-divided to divide out multiple abnormal data sub-clusters.

[0074] To judge whether there are discontinuous regions, spatial continuity checks and visual image checks can be carried out. The spatial continuity check is carried out by calculating the normalized parameter values of the drilling speeds of all fractured rock masses V . If within the same cluster, the adjacent depth values of the ordinates corresponding to the normalized parameter values of the drilling speed V are greater than 3, then there is very likely a discontinuous region. At the same time, through the observation of the visual image, if it is found that the distribution of the fractured rock mass is too scattered, it can be determined that the region where these two adjacent points are located is the discontinuous region within the same cluster.

[0075] Next, the fractured rock masses in the discontinuous area are regrouped. The fractured rock masses of each adjacent part form a group, and the boundary points determined through spatial continuity inspection and visual image inspection are the boundaries of each group of fractured rock masses. When determining the central position of the fractured rock masses, the center point of the cluster cannot be calculated for this area. Instead, the center points of each group of fractured rock masses need to be calculated, and then the average value of the core points of all groups is taken to determine the central position of the fractured rock masses.

[0076] Next, the boundaries of each fractured rock mass cluster in the vertical direction are determined. This is achieved by identifying the minimum and maximum depth values of the normalized parameter data points of the drilling speed V within the cluster. The minimum depth value represents the starting position of the fractured rock mass, while the maximum depth value represents the termination position of the fractured rock mass.

[0077] The continuous distance of the fractured rock mass is estimated by calculating the difference between the cluster boundaries. Specifically, the continuous distance = maximum depth - minimum depth. This value provides the extension range of the fractured rock mass in the vertical direction.

[0078] At the same time, to ensure the accuracy of these fractured rock masses found through unsupervised methods, the positions and continuous distances of the fractured rock masses obtained from the cluster analysis are compared with the actual coring results to verify the accuracy of the prediction.

[0079] In a further method of the above method, mainly the processes of obtaining the neighborhood radius and the minimum number of points are explained as follows: Calculate the distances from each data point in the historical drilling data to its k nearest neighbors, and construct a k-distance graph; Determine the inflection point region in the curve on the k-distance graph; Determine the inflection point with the maximum slope corresponding to the inflection point region, and based on the inflection point with the maximum slope, determine multiple points near the inflection point, and obtain the slopes corresponding to the points near the inflection point; Based on the determined slopes, perform data clustering on the historical drilling data set to obtain the clustering results, and determine the best slope as the neighborhood radius based on the clustering.

[0080] Under the regional limitation of the determined neighborhood radius, adopt a traversal method with the initial number of points as the starting value, and gradually increase the number of points according to the preset increment; form different data clusters based on different numbers of points; Based on different data clusters, determine the evaluation indicators corresponding to different numbers of points; Based on the evaluation indicators of different numbers of points, count the number of consecutive times of consistent evaluation indicators. When the number of consecutive times reaches the preset number of times, based on the visualization results of the data clusters corresponding to the continuously consistent evaluation indicators, determine one of the numbers of points as the minimum number of points.

[0081] It should be noted that in the present invention, in order to determine the appropriate eps value (i.e., neighborhood radius) and minPts value (i.e., minimum number of points), two methods of KNN distance plot and data traversal are adopted.

[0082] 1. KNN distance plot: (1) Calculate the k-nearest neighbor distance: First, calculate the distance from each data point in the dataset to its k nearest neighbors, and construct a k-distance graph.

[0083] (2) Identify the inflection point region: Observe the k-distance graph and find the inflection point region in the curve, that is, the starting point of the transition region where the k-th NearestDistance value changes from relatively stable to a sharp increase.

[0084] (3) Screen candidate eps values: Determine the point with the largest slope: Since the inflection point may not be accurately determined in the graph, the point with the largest slope is used as a candidate value for the inflection point. The point with the largest slope marks the transition from the high-density region to the low-density region and is an important reference point.

[0085] Screen points near the inflection point: Further screen the k-th Nearest Distance values of the inflection point and 4 points near it to obtain a set containing multiple inflection point slopes.

[0086] (4) Determine the optimal eps value: Apply the screened inflection point slopes as the neighborhood radius to the clustering algorithm respectively, and confirm the optimal inflection point slope as the neighborhood radius by comparing the clustering results and evaluation indicators under different inflection point slopes.

[0087] Taking small fracture data as an example below. First, import the data. According to the data characteristics, under the condition of fixing other structural parameters of the DBSCAN algorithm, set the structural parameter k value of KNN to 2 according to the empirical method, and draw the k-thNearest Distance under different data points, as Figure 4 shown. Through the k-distance graph, it is found that near the 1614 position, the k-th NearestDistance graph shows an obvious inflection point region. Select the corresponding eps values (0.394, 0.402, 0.697, 4.562) of this inflection point region, and perform unsupervised clustering fracture identification accordingly. The results are shown in Figure 5a 、 Figure 5b 、 Figure 5c and Figure 5d . Through analysis of the results, it is concluded that the optimal eps value determined by the k-th nearest neighbor distance graph is 0.697, and the fracture identification effect is the most ideal at this time.

[0088] 2. Data traversal: The value of the parameter minpts is determined by the method of data traversal. The specific steps are as follows: First, according to the determined value of eps, starting from one more than the data dimension, gradually increase the value of minpts for data traversal. Different minpts values will form different data clusters, and these clusters will affect the evaluation indicators and effects of fractured rock mass identification. During the traversal process, it may be found that within certain intervals of minpts values, the numerical values of the evaluation indicators of the identification results remain consistent. This situation indicates that within this interval, different data clusters have the same identification effect on the fractured rock mass. Therefore, only need to continue to increase the value of minpts until it is found that the evaluation indicators no longer change with the increase of the minpts value, then the traversal can be stopped. Finally, combining the evaluation indicators and the data visualization results, jointly analyze and determine the optimal minpts value.

[0089] According to the characteristics of the dataset and the number of unsupervised classifications, the following clustering effect evaluation indicators are adopted: 1) Silhouette coefficient The silhouette coefficient (SS) is an index for evaluating the performance of a clustering algorithm, which can measure the compactness and separation degree of the clustering results. The calculation formula of the silhouette coefficient is as follows: In the formula a represents the average distance between a sample point and other sample points in the same cluster (intra-cluster distance). b represents the average distance between a sample point and all sample points in the nearest cluster (inter-cluster distance). The closer the silhouette coefficient is to 1, the closer the sample point is to other sample points in its cluster and the farther it is from the sample points in the nearest cluster, and the better the clustering effect. A silhouette coefficient of 0 means that the sample point is on the boundary of the cluster and it is difficult to judge which cluster it belongs to. A negative silhouette coefficient means that the sample point is closer to the sample points in other clusters, and the clustering effect is poor.

[0090] 2) Calinski-Harabasz index The calculation formula of the CH index is as follows: In the formula n represents the total number of samples, K represents the number of clusters, n i represents the number of samples in cluster i. S W represents the sum of the intra-cluster distances, S B represents the sum of the inter-cluster distances. CH The larger the index, the closer the sample points within the cluster are and the farther the sample points between the clusters are, and the better the clustering effect.

[0091] Taking the small fracture data as an example again, under the condition that the eps value is determined to be 0.697, the minpts parameter is set to traverse data starting from 3 using the data traversal method (since the dimension of the experimental data is 2, the minpts parameter is set to start traversing from 3), and the evaluation indexes under each different combination of minpts parameters are calculated and the clustering effect diagrams are drawn until the evaluation indexes no longer change when reaching a certain value, terminating the data traversal. In this experiment, when the minPts parameter increases to 8, the values of SS and CH both tend to be stable and no longer change, which indicates that in this case, even if the value of minPts is further increased, the recognition accuracy of fractured rock masses will not be improved, as Figure 6 shown.

[0092] First, according to the data traversal results of minpts, the evaluation indexes under each combination are referred to remove the clustering results with poor evaluation, and then the recognition results of fractured rock masses under different minPts values (5, 8, and 12 respectively) are analyzed, as Figure 7a 、 Figure 7b and Figure 7c shown. It is determined that when the minPts value is set to 8, the best unsupervised clustering evaluation index results are obtained (SS is 0.810 and CH is 357.398).

[0093] In the further method of the above method, it mainly explains the process of preprocessing the drilling data before making the drilling data of the rock mass distributed in the preset relationship diagram, which is specifically as follows: Based on the obtained drilling data, effective drilling process data identification, outlier removal, and drilling data splicing are carried out.

[0094] 1) Effective drilling process data identification: The complete drilling process can be subdivided into three stages: the preparation stage, the stable drilling stage, and the drill rig lifting stage. The drilling displacement in the original data is shown in the following figure. A series of necessary preparation work is carried out in the preparation stage and the drill rig lifting stage, including the positioning of the drill rig, the cleaning of the hole bottom, the assembly of drill pipes, and the replacement of drill bits, etc. The stable drilling stage is the core of the drilling process, and in this stage, the drilling operation is realized through the continuous action of the power head. Therefore, first, the data of the stable drilling stage need to be extracted from the whole drilling process.

[0095] The expression of the method for extracting the process data of the stable drilling stage is as follows: In the formula, Nis the drilling speed in the real-time process of the drill rig, and P is the impact pressure in the real-time process of the drill rig; and are the corresponding drilling displacements (m) at the starting point and the ending point of the stable drilling stage of the displacement sensor respectively; is the actual length of the drill pipe, in m, adjusted according to the telescopic displacement range of the power head of the on-site equipment; is the maximum drilling speed, in r / min, adjusted according to the rotational speed range of the power head of the on-site equipment; is the maximum impact pressure, in MPa, adjusted according to the impact pressure range of the on-site equipment.

[0096] 2) Outlier rejection: Identify and reject the abnormal data points generated by factors such as equipment failures and environmental interferences to ensure the accuracy and reliability of the data.

[0097] Before the formal data preprocessing, it is necessary to identify and reject the obviously abnormal data. The outlier detection adopts a variety of statistical methods and machine learning methods embedded, including box plots, Z-score methods, DBSCAN clustering methods, isolation forest methods, etc. To determine the most suitable method, a certain proportion of outliers are artificially set in the normal data in the early stage. These outliers should represent the abnormal types that may be encountered in actual applications, and use them to detect the ability of different methods to identify outliers. The evaluation index uses the accuracy rate, that is, the proportion of correctly identified normal values and outliers, so as to determine the best method for identifying abnormal data, and then reject them.

[0098] 3) Drilling data splicing Since the drilling displacement is in segments due to the limited length of the drill pipe in a single drilling process, it is necessary to splice the drilling displacement data of different drill pipes first.

[0099] First, ensure that the displacement data of each section of the drill pipe has the same timestamp format for subsequent data splicing. If the timestamps are not aligned, it is necessary to adjust the data through interpolation or other methods to make their time series consistent.

[0100] Then, when replacing the drill pipe, record the ending displacement data of the previous section of the drill pipe and the starting displacement data of the next section of the drill pipe. Connect the starting displacement data of the next section of the drill pipe with the ending displacement data of the previous section of the drill pipe to ensure the continuity of the data.

[0101] Then, to obtain the spliced drilling displacement starting from zero, the following steps can be taken: a. Determine the starting point of the drilling, that is, the position where the first section of the drill pipe starts to drill. b. Set the displacement value at this starting point as the reference point, that is, the displacement value is zero. c. For all subsequent displacement data, subtract the displacement value at the starting point. In this way, the displacement at the starting point becomes zero, and the displacements of other points are the displacements relative to the starting point.

[0102] Then, check whether there are mutations or discontinuities in the spliced displacement data, which may be caused by drill pipe replacement or other reasons. Use methods such as moving average or low-pass filter to smooth the data to reduce noise. The moving average can be achieved by calculating the average value within a certain time window, while the low-pass filter allows data below a specific frequency to pass through and suppresses high-frequency noise.

[0103] Finally, plot the processed displacement data as a curve in chronological order, with the abscissa being time and the ordinate being the drilling depth. The curve obtained in this way shows the depth change of the drilling process over time, that is, the drilling time history curve.

[0104] Through the above steps, a continuous drilling displacement data starting from zero can be obtained, and the drilling time history curve can be plotted. Such a curve helps to analyze information such as drilling efficiency, bit wear, and formation changes.

[0105] The following describes the device for identifying fractured rock masses based on drilling data provided by the present invention. The device for identifying fractured rock masses based on drilling data described below can be correspondingly referred to the method for identifying fractured rock masses based on drilling data described above.

[0106] Figure 8 The flowchart of a device for identifying fractured rock masses based on drilling data provided by the present invention is shown. Refer to Figure 8 This device includes an acquisition module 81, an acquisition module 82, and an identification module 83, where: The acquisition module is used to obtain the drilling data of the rock mass and construct a relationship graph based on the drilling data of the rock mass.

[0107] The acquisition module is used to obtain a pre-configured neighborhood radius and minimum number of points. Among them, the neighborhood radius is determined by the slope inflection point in the k-distance graph generated based on historical drilling data, and the minimum number of points is determined by the identification evaluation index obtained by performing fractured rock mass identification on historical drilling data with regional restrictions based on the determined neighborhood radius.

[0108] The identification module is used to perform data clustering processing on the drilling data of the rock mass in a preset relationship graph based on the neighborhood radius and the minimum number of points, and determine abnormal data clusters. The abnormal data clusters represent the regional information of the fractured rock mass in the preset relationship graph.

[0109] Since the principle of the device in the embodiment of the present invention is the same as that of the method in the above embodiment, no more detailed explanation will be given here.

[0110] It should be noted that in the embodiment of the present invention, the relevant functional modules can be implemented by a hardware processor.

[0111] The device for identifying fractured rock masses based on drilling data provided by the present invention can reasonably optimize the neighborhood radius and the minimum number of points, and then, based on the neighborhood radius and the minimum number of points, perform data clustering processing on the drilling data of the rock mass in the relationship graph to determine abnormal data clusters. It can handle clusters of any shape, enhance the ability to identify fractured rock masses with diverse morphologies, provide more accurate basic data for engineering applications, not only improve the accuracy of identification, but also effectively reduce the dependence on manual intervention and reduce the risk of human errors.

[0112] Figure 9 An example of the physical structure diagram of an electronic device is shown as Figure 9 shown. The electronic device may include: a processor 91, a communication interface 92, a memory 93, and a communication bus 94. Among them, the processor 91, the communication interface 92, and the memory 93 complete communication with each other through the communication bus 94. The processor 91 can call the logical instructions in the memory 93 to execute the method for identifying fractured rock masses based on drilling data. The method includes: obtaining the drilling data of the rock mass, and constructing a relationship graph based on the drilling data of the rock mass; obtaining a pre-configured neighborhood radius and the minimum number of points, where the neighborhood radius is determined by the slope inflection point in the k-distance graph generated based on historical drilling data, and the minimum number of points is determined by the identification evaluation index obtained by performing fractured rock mass identification on the historical drilling data with regional limitation based on the determined neighborhood radius; based on the neighborhood radius and the minimum number of points, perform data clustering processing on the drilling data of the rock mass in the relationship graph to determine abnormal data clusters, and the abnormal data clusters represent the regional information of the fractured rock mass in the preset relationship graph.

[0113] In addition, when the logical instructions in the above-mentioned memory 93 are implemented in the form of software functional units and sold or used as an independent product, they can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or a part of this technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: USB flash drives, mobile hard disks, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical discs, etc., which can store program codes.

[0114] On the other hand, the present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the fracture rock mass identification method based on drilling data provided by the above-mentioned various methods. The method includes: obtaining the drilling data of the rock mass, and constructing a relationship graph based on the drilling data of the rock mass; obtaining a pre-configured neighborhood radius and a minimum number of points. Among them, the neighborhood radius is determined by the slope inflection point in the k-distance graph generated based on historical drilling data, and the minimum number of points is determined by the identification evaluation index obtained by performing fracture rock mass identification on historical drilling data with regional limitation based on the determined neighborhood radius; based on the neighborhood radius and the minimum number of points, performing data clustering processing on the drilling data of the rock mass in the relationship graph to determine abnormal data clusters, and the abnormal data clusters represent the regional information of the fracture rock mass in the preset relationship graph.

[0115] In another aspect, the present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it is implemented to execute the fracture rock mass identification method based on drilling data provided by the above-mentioned various methods. The method includes: obtaining the drilling data of the rock mass, and constructing a relationship graph based on the drilling data of the rock mass; obtaining a pre-configured neighborhood radius and a minimum number of points. Among them, the neighborhood radius is determined by the slope inflection point in the k-distance graph generated based on historical drilling data, and the minimum number of points is determined by the identification evaluation index obtained by performing fracture rock mass identification on historical drilling data with regional limitation based on the determined neighborhood radius; based on the neighborhood radius and the minimum number of points, performing data clustering processing on the drilling data of the rock mass in the relationship graph to determine abnormal data clusters, and the abnormal data clusters represent the regional information of the fracture rock mass in the preset relationship graph.

[0116] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place, or may be distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art can understand and implement it without creative labor.

[0117] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on such an understanding, the essence of the above technical solution, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0118] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features. These modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for identifying fractured rock mass based on drilling data, characterized in that: include: Obtain drilling data of the rock mass, and construct a relationship diagram based on the drilling data of the rock mass; Obtaining a preconfigured neighborhood radius and a minimum number of points, wherein the neighborhood radius is determined by a slope inflection point in a k-distance graph generated based on historical drilling data, and the minimum number of points is determined by performing regional restrictions based on the determined neighborhood radius and identifying fractured rock masses on the historical drilling data to obtain an identification evaluation index; Based on the neighborhood radius and the minimum number of points, data clustering processing is performed on the drilling data of the rock mass in a preset relationship diagram to determine abnormal data clusters, wherein the abnormal data clusters represent regional information of the fractured rock mass in the preset relationship diagram.

2. The method for identifying fractured rock mass based on drilling data according to claim 1, characterized in that: The drilling data of the rock mass includes drilling depth, drilling pressure, rotation speed, torque and drilling speed. Accordingly, a drilling speed normalization parameter is determined according to the drilling pressure, rotation speed, torque and drilling speed, and a relationship diagram is constructed based on the drilling depth and the drilling speed normalization parameter.

3. The method for identifying fractured rock mass based on drilling data according to claim 2, characterized in that: The method of determining a normalized drilling speed parameter according to the drilling pressure, rotation speed, torque and drilling speed comprises: According to the drilling pressure, rotation speed, torque and drilling speed, the drilling speed normalization parameter is determined using the following calculation formula; in, is the normalized parameter of drilling speed, is the drilling speed, is the drilling pressure, is the rotation speed, is the torque, , , and is the preset coefficient.

4. The method for identifying fractured rock mass based on drilling data according to claim 1 or 3, characterized in that: The method further comprises: Perform regional continuity judgment on the abnormal data cluster. When it is determined that the abnormal data cluster has continuity characteristics, identify the center position of the abnormal data cluster and determine the continuity distance of the abnormal data cluster. When it is determined that the abnormal data cluster does not have continuity characteristics, divide the abnormal data cluster again to determine multiple abnormal data sub-clusters, identify the center position of the abnormal data sub-cluster and determine the continuity distance of the abnormal data sub-cluster; wherein the continuity distance represents the continuity depth of the fractured rock mass.

5. The method for identifying fractured rock mass based on drilling data according to claim 1, characterized in that: The method further comprises a step of obtaining a neighborhood radius, comprising: Calculate the distance from each data point in the historical drilling data to its nearest k neighbors and construct a k-distance graph; Determine the inflection point region in the curve on the k-distance graph; Determine the inflection point with the largest slope corresponding to the inflection point area, determine multiple points near the inflection point based on the inflection point with the largest slope, and obtain the slopes corresponding to the points near the inflection point; Based on the determined slopes, data clustering is performed on the historical drilling data set to obtain clustering results, and based on the clustering, an optimal slope is determined as a neighborhood radius.

6. The method for identifying fractured rock mass based on drilling data according to claim 5, characterized in that: The method further comprises a step of obtaining a minimum number of points, comprising: Under the area restriction of the determined neighborhood radius, a traversal method is adopted to take the initial number of points as the starting value, and the number of points is gradually increased according to the preset increment; different data clusters are formed based on different numbers of points; Based on different data clusters, determine the evaluation indicators corresponding to different points; Based on evaluation indicators with different numbers of points, the number of consecutive times when the evaluation indicators are consistent is counted. When it is determined that the number of consecutive times reaches a preset number, one of the points is determined as the minimum number of points based on the visualization results of the data clusters corresponding to the consecutive and consistent evaluation indicators.

7. The method for identifying fractured rock mass based on drilling data according to claim 5, characterized in that: Before distributing the drilling data of the rock mass in a preset relationship diagram, the method further includes: Based on the obtained drilling data, effective drilling process data identification, outlier removal and drilling data splicing are carried out.

8. A fractured rock mass identification device based on drilling data, characterized in that: include: An acquisition module is used to obtain drilling data of the rock mass and to construct a relationship diagram based on the drilling data of the rock mass; An acquisition module, used to acquire a pre-configured neighborhood radius and a minimum number of points, wherein the neighborhood radius is determined by a slope inflection point in a k-distance graph generated based on historical drilling data, and the minimum number of points is determined by an identification evaluation index obtained by performing regional restrictions based on the determined neighborhood radius and performing fracture rock mass identification on the historical drilling data; The identification module is used to perform data clustering processing on the drilling data of the rock mass in a preset relationship diagram based on the neighborhood radius and the minimum number of points to determine abnormal data clusters, wherein the abnormal data clusters represent regional information of the fractured rock mass in the preset relationship diagram.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the fractured rock mass identification method as described in any one of claims 1 to 7 is implemented.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for identifying fractured rock mass as claimed in any one of claims 1 to 7 is implemented.

Images