A data processing method and system for mining drilling

Abstract

The present invention discloses a data processing method and system for mining drilling, specifically in the field of drilling data processing, for solving the problem of mining drilling data preprocessing, by calculating the rate of change of formation density and Poisson's ratio data, and comparing its standard deviation with a set fluctuation threshold, to determine the fluctuation level of the data. According to the fluctuation level, the interpolation method or truncation method is flexibly selected to process the data to ensure data reliability and processing flexibility. Reliable signals are generated by comparing the response degree and the data reliability threshold, and the abnormality of the measured data is evaluated, providing accuracy and reliability for geological exploration and drilling engineering. The classification ability of the feature is judged according to the classification uncertainty index, and the response degree is introduced for quality correction, and finally a feature set with high correlation and classification ability is obtained, thereby improving the accuracy and reliability of drilling data analysis.

Description

Technical Field

[0001] The present invention relates to the field of drilling data processing, and more specifically, to a data processing method and system for mining drilling. Background Art

[0002] When conducting mining drilling, analyzing various data during the drilling process can provide key information to guide geological exploration and drilling engineering, thereby promoting the development of mining projects.

[0003] However, there are deficiencies in the existing drilling data preprocessing, especially in feature selection. Traditional methods often lack pertinence and fail to fully consider the volatility and correlation of data, resulting in the selected features may not accurately reflect the formation conditions or have an unclear relationship with the target variable. In addition, the lack of effective evaluation indicators and screening mechanisms can easily lead to unreliable or redundant selected features, affecting the accuracy and reliability of subsequent data analysis. Therefore, it is necessary to improve the drilling data preprocessing method, introduce a more flexible and accurate feature selection strategy, and fully consider the volatility, correlation and reliability of the data to improve the accuracy and effectiveness of drilling data analysis and provide more reliable support for decision-making in mining projects.

[0004] In order to solve the above problems, a technical solution is now provided. Summary of the invention

[0005] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a data processing method and system for mining drilling, which determines the fluctuation level of the data by calculating the rate of change of the formation density and Poisson's ratio data and comparing its standard deviation with the set fluctuation threshold. According to the fluctuation level, the interpolation method or truncation method is flexibly selected to process the data to ensure data reliability and processing flexibility. By comparing the response degree and the data reliability threshold, a reliable signal is generated, and the abnormality of the measured data is evaluated to provide accuracy and reliability for geological exploration and drilling engineering. The classification ability of the feature is judged according to the classification uncertainty index, and the response degree is introduced for quality correction, and finally a feature set with high correlation and classification ability is obtained, which improves the accuracy and reliability of drilling data analysis to solve the problems raised in the above background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions: S1. Based on the corresponding acquisition time point, the corresponding interpolation method is adapted according to the data volatility, and the formation density data and Poisson's ratio data are aligned according to the time point;

[0007] S2. Based on the formation density data and Poisson's ratio data obtained under the alignment conditions, they are merged into a unified data set, a correlation matrix is ​​constructed, and the gray value between pixels is used to calculate the response degree to evaluate the reliability of the measurement data;

[0008] S2. Take the formation density data and Poisson's ratio data with expected reliability as the target variables, analyze each feature of the drilling process in turn with the target variables, obtain the corresponding classification uncertainty index, and screen and select the features.

[0009] In a preferred embodiment, S1 includes the following contents:

[0010] Perform timestamp alignment on formation density data and Poisson's ratio data, including the following:

[0011] According to the pre-set time window of the formation density data and Poisson's ratio data, that is, the time range in which the two data exist together, the selected time window includes the intersection of the two data;

[0012] For the case where the time points are not completely aligned, evaluate the fluctuation level of the data and select the corresponding difference method according to the fluctuation level to interpolate the formation density and Poisson's ratio data to the same time point. The execution steps are as follows:

[0013] Step 1: Calculate the first-order difference of the time series data of formation density and Poisson's ratio in unit time, that is, calculate the change between adjacent time points, perform first-order difference calculation on the formation density data and Poisson's ratio data, and obtain the change rate series of formation density and Poisson's ratio;

[0014] Step 2: Calculate the standard deviation of the change rate series of formation density and Poisson's ratio to evaluate their fluctuation degree;

[0015] Compare the standard deviation with the fluctuation threshold. If the standard deviation is greater than or equal to the fluctuation threshold, it means that the data fluctuates greatly and a high fluctuation signal is generated. The truncation method is suitable for aligning the data to the same time point. If the standard deviation is less than the fluctuation threshold, it means that the data fluctuates less and a low fluctuation signal is generated. The interpolation method is suitable for aligning the data to the same time point.

[0016] In a preferred embodiment, S2 includes the following contents:

[0017] Step 1: Perform minimum-maximum normalization processing on the formation density data and Poisson's ratio data per unit time and convert them into the same scale to eliminate the influence of dimension;

[0018] Step 2: The standardized formation density data and Poisson's ratio data are combined into a correlation matrix, where each element in the correlation matrix represents the degree of correlation between two data points;

[0019] Step three, calculate the response degree between the formation density data and the Poisson's ratio data, using the gray value between pixels to calculate, indicating the response degree between adjacent pixels.

[0020] In a preferred embodiment, the response level is compared with a data reliability threshold interval;

[0021] If the response degree is greater than the maximum value of the data reliability threshold interval, it means that the correlation between the formation density data and the Poisson's ratio data is high, generating a high reliability signal;

[0022] On the contrary, if the response degree is less than the minimum value of the data reliability threshold interval, it means that the correlation between the formation density data and the Poisson's ratio data is low, generating a low reliability signal;

[0023] If the response degree is within the data reliability threshold interval, it means that the correlation between the formation density data and the Poisson's ratio data is within an acceptable range, and a medium reliability signal is generated.

[0024] In a preferred embodiment, S3 includes the following contents:

[0025] After obtaining the non-low reliability signal, the formation density data and Poisson's ratio data obtained per unit time are respectively used as target variables, and the features to be screened for auxiliary analysis are counted. The features to be screened are screened by the target variables. The process is as follows:

[0026] Step 1, determining the timestamp for obtaining the formation density data and Poisson's ratio data per unit time;

[0027] Step 2: Count the characteristics of the drilling process;

[0028] Step 3: Under the condition of the timestamp corresponding to the unit time of obtaining the formation density data and Poisson's ratio data, for each feature, the data set is grouped according to the value of the feature, and for each value of each feature, the category distribution of the target variable under the corresponding value is calculated, and its Gini coefficient is calculated, and the weighted average Gini coefficient of each feature is calculated as a measure of the diversity of the feature to the target variable;

[0029] Step 4: Calculate the absolute value of the difference between the weighted average Gini coefficients of each target variable corresponding to each feature, and then conduct classification discussion based on the degree of response to obtain the classification uncertainty index.

[0030] In a preferred embodiment, in step five, the classification uncertainty index is compared with a screening threshold;

[0031] If the classification uncertainty index is greater than or equal to the screening threshold, it means that the classification effect of the feature on the target variable is uncertain or unreliable, and a rejection signal is generated; conversely, if the classification uncertainty index is less than the screening threshold, it means that the classification effect of the feature on the target variable is relatively certain and reliable, and a recruitment signal is generated;

[0032] Step six, count the characteristics of life recruitment signals, sort them from small to large according to the classification uncertainty index, and delete them according to the preset recruitment number.

[0033] A data processing system for mining drilling, comprising a time alignment module, a data merging module, a target variable selection module, a feature analysis module and a feature screening module;

[0034] The time alignment module is used to select the interpolation method according to the data volatility, align the formation density and Poisson's ratio data, and send the alignment results to the data merging module;

[0035] The data merging module constructs a correlation matrix with the aligned data, calculates the response degree to evaluate the data reliability, and sends the response degree to the target variable selection module;

[0036] The target variable selection module determines the formation density and Poisson's ratio data that meet the expected reliability as the target variable, and sends the target variable to the feature analysis module;

[0037] The feature analysis module combines each feature with the target variable, calculates the classification uncertainty index, evaluates the classification ability of the feature, and sends the classification uncertainty index to the feature screening module;

[0038] The feature screening module screens features based on the classification uncertainty index to determine whether to retain them for further analysis.

[0039] Technical effects and advantages of a data processing method and system for mining drilling of the present invention:

[0040] 1. The present invention calculates the rate of change of formation density and Poisson's ratio data, calculates their standard deviations respectively to evaluate the degree of fluctuation, and compares them with the set fluctuation threshold to determine the fluctuation level of the data. According to the fluctuation level, the interpolation method or truncation method is flexibly selected to process the data to ensure the reliability and processing flexibility of the aligned data. By comparing the response level and data reliability threshold of formation density data and Poisson's ratio data, a corresponding reliable signal is generated to analyze the abnormality of the measured data. It can effectively evaluate and process drilling data, ensure the quality and authenticity of the data measured by drilling, provide important reference information for geological exploration and drilling engineering, and help improve the accuracy and reliability of data analysis, thereby promoting the progress and development of related projects.

[0041] 2. The present invention determines the formation density data and Poisson's ratio data that can be used as the target variable based on the comparison results of the response degree and the data reliability threshold interval, and calculates the classification uncertainty index for each feature under the condition that the feature data and the target variable are in the same timestamp to evaluate the classification ability of the feature for the target variable. By comparing the classification uncertainty index and the preset screening threshold, it can be determined whether the classification ability of the feature is sufficiently reliable, thereby deciding whether to retain the feature for further analysis. The response degree is introduced in a timely manner and taken into consideration in the feature screening process to correct the screening quality of the classification uncertainty index. The features are sorted according to the classification uncertainty index and deleted according to the preset number of admissions to obtain the final auxiliary analysis feature set. In this way, features with high relevance and good classification ability to the target variable can be effectively identified, thereby improving the accuracy and reliability of drilling data analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 A schematic flow chart of a data processing method for mining drilling according to the present invention;

[0043] Figure 2 The present invention is a structural schematic diagram of a data processing system for mining drilling. DETAILED DESCRIPTION

[0044] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0045] Embodiment 1: Figure 1 The present invention provides a data processing method for mining drilling, comprising:

[0046] S1. Based on the corresponding acquisition time point, the corresponding interpolation method is adapted according to the data volatility, and the formation density data and Poisson's ratio data are aligned according to the time point;

[0047] S2. Based on the formation density data and Poisson's ratio data obtained under the alignment conditions, they are merged into a unified data set, a correlation matrix is ​​constructed, and the gray value between pixels is used to calculate the response degree to evaluate the reliability of the measurement data;

[0048] S2. Take the formation density data and Poisson's ratio data with expected reliability as the target variables, analyze each feature of the drilling process in turn with the target variables, obtain the corresponding classification uncertainty index, and screen and select the features.

[0049] S1 includes the following:

[0050] During the drilling process, logging tools are used to obtain various logging curve data, including natural gamma logging (GR), density logging (DEN), acoustic logging (AC), neutron logging (CNL), etc. These logging curves provide information about the physical properties of the formation.

[0051] Formation density is calculated from a density log (DEN) curve. The density log provides the bulk density of the formation in grams per cubic centimeter (g / cm³).

[0052] The calculation formula of formation density is as follows: ;

[0053] in, is the formation density, is the reading of the measured density log curve, is the density of the drilling fluid, is the dry rock density of the formation, is the dry rock density of the formation.

[0054] Formation density is used to indicate the material density of underground rocks and is one of the important physical properties of rocks. The size of the formation density reflects the compactness of the formation rock, that is, the compactness of the particles or porosity inside the rock. A formation with a higher density indicates that the rock is denser, the particles are tighter, and the porosity is smaller; while a formation with a lower density may indicate that there are more pores or low-density components in the rock. Therefore, changes in formation density can be used to identify the properties of geological layers, distinguish different lithologies or lithofacies, and evaluate the structure, properties, and oil and gas content of underground rocks.

[0055] Poisson's ratio is calculated from acoustic logging (AC) and density logging (DEN) data. The acoustic logging provides the elastic modulus (Young's modulus) of the formation, while the density logging provides the density of the formation. Poisson's ratio is one of the elastic properties of the formation and is not measured directly, but is estimated from the acoustic velocity and the formation density.

[0056] The calculation formula of Poisson's ratio is as follows: ;

[0057] in, is the Poisson's ratio, is the longitudinal wave velocity (P-wave velocity), is the shear wave velocity (S wave velocity).

[0058] Poisson's ratio is used to indicate the deformation characteristics of rocks or materials under stress, that is, when a material is subjected to external forces, the contraction or extension along one direction will cause deformation perpendicular to that direction. The larger the Poisson's ratio, the more sensitive the material is to external forces and the greater the deformation capacity of the material, while the smaller the Poisson's ratio, the smaller the deformation capacity of the material.

[0059] The acquisition of formation density data and Poisson's ratio data based on the same timestamp is of great significance for drilling analysis. Doing so ensures that the data acquired under the same geological conditions are consistent, which is conducive to the accurate evaluation and prediction of formation characteristics. Formation density data can provide the compactness and physical properties of the formation, while Poisson's ratio data reflects the elastic properties of the formation. By acquiring these two types of data at the same time, a more comprehensive understanding of the nature and composition of the formation can be achieved, helping engineers and geologists make more accurate drilling decisions. For example, in rock mechanics modeling and wellbore stability analysis, formation density and Poisson's ratio are important parameters that can help predict the stability of the wellbore and the fracture of the rock. Therefore, the acquisition of formation density data and Poisson's ratio data based on the same timestamp helps to improve the efficiency and success rate of drilling projects and provide reliable data support for geological exploration and resource development.

[0060] Perform timestamp alignment on formation density data and Poisson's ratio data, including the following:

[0061] According to the pre-set time window of the formation density data and Poisson's ratio data, that is, the time range in which the two data exist together, the selected time window contains the intersection of the two data.

[0062] For the case where the time points are not completely aligned, evaluate the fluctuation level of the data and select the corresponding difference method according to the fluctuation level to interpolate the formation density and Poisson's ratio data to the same time point. The execution steps are as follows:

[0063] Step 1: Calculate the first-order difference of the time series data of formation density and Poisson's ratio in unit time, that is, calculate the change between adjacent time points. Calculate the first-order difference of formation density data and Poisson's ratio data to obtain the change rate series of formation density and Poisson's ratio. and .

[0064] ;

[0065] ;

[0066] Step 2: Calculate the standard deviation of the change rate series of formation density and Poisson's ratio to evaluate their fluctuation degree. The standard deviation reflects the degree of dispersion of the data. The larger the standard deviation, the greater the fluctuation degree of the data.

[0067] ;

[0068] ;

[0069] in, is the length of the change sequence, and are the means of the change rate series of formation density and Poisson's ratio, respectively.

[0070] Compare the standard deviation with the fluctuation threshold. If the standard deviation is greater than or equal to the fluctuation threshold, it means that the data fluctuates greatly and a high fluctuation signal is generated. The truncation method is suitable for aligning the data to the same time point. If the standard deviation is less than the fluctuation threshold, it means that the data fluctuates less and a low fluctuation signal is generated. The interpolation method is suitable for aligning the data to the same time point.

[0071] For cases where time points are not completely aligned, the choice of interpolation or truncation should be evaluated based on the specific situation. The following is an evaluation of the two methods:

[0072] Interpolation fills in missing data by interpolating between known time points, which can maintain the continuity and smoothness of the data. This method is suitable for situations where the data changes relatively smoothly, so it is suitable for data with low fluctuations.

[0073] The truncation method cuts off the part with data in the time window and discards the missing data, which can ensure that the retained data is true and reliable. This method is suitable for situations where the data changes drastically or there is a lot of noise, so it is suitable for data with high volatility.

[0074] The correlation between formation density data and Poisson's ratio data can be used to analyze the degree of abnormality of measured data, because formation density and Poisson's ratio usually have a certain correlation in geology and geophysics. In geological bodies, different rocks or strata usually have specific density and Poisson's ratio characteristics, which are detected and measured during underground drilling. Therefore, if there is a high correlation between formation density data and Poisson's ratio data, that is, their change trends are similar or the change rules are consistent, it usually indicates that the properties of underground rocks or strata are relatively stable, and the degree of abnormality of measured data is low. On the contrary, if the correlation between formation density data and Poisson's ratio data is low, it may mean that the properties of underground strata have changed or there are abnormalities, resulting in reduced reliability of measured data. Therefore, by analyzing the correlation between formation density data and Poisson's ratio data, it can provide important reference information for geological exploration and drilling engineering, and help evaluate the degree of abnormality of measured data and the properties of underground rocks.

[0075] S2 includes the following:

[0076] Step 1: Perform minimum-maximum normalization processing on the formation density data and Poisson's ratio data per unit time and convert them into the same scale to eliminate the influence of dimension;

[0077] Step 2: The standardized formation density data and Poisson's ratio data are combined into a correlation matrix. Each element in the correlation matrix represents the degree of correlation between two data points;

[0078] Step 3: Calculate the response between the formation density data and the Poisson's ratio data. Use the grayscale value between pixels to calculate, which represents the response between adjacent pixels. The calculation formula is as follows: ;

[0079] in, is the degree of response, is the number of data points, is the first Line and The value of the column element indicates the The formation density data and The response level of a Poisson's ratio data.

[0080] The response degree is used to evaluate the abnormality of drilling measurement data. The higher the response degree, the lower the abnormality of drilling measurement data. The higher the response degree, the higher the correlation between the two, which means that there is a certain pattern or law between the two, which usually indicates that the changes between these data are coordinated with each other, rather than caused by random factors. This is a normal phenomenon of drilling data, because formation density and Poisson's ratio usually have a certain correlation in geology and geophysics. In geological bodies, different rocks or formations usually have specific density and Poisson's ratio characteristics, which will be detected and measured during underground drilling. Therefore, if there is a high correlation between formation density data and Poisson's ratio data, that is, their change trends are similar or their change laws are consistent, it indicates that the abnormality of the measurement data is low and the measured data is relatively reliable.

[0081] On the contrary, if the response between the formation density data and the Poisson's ratio data is low, it implies that the measured data is highly abnormal. A higher response indicates a low correlation between the two. Low correlation data means that there is no obvious pattern or regularity in their changes, suggesting that there is a problem with the measurement equipment, and the measurement equipment needs to be checked and calibrated to ensure the accuracy and reliability of the data.

[0082] Compare the response level with the data reliability threshold interval;

[0083] If the response degree is greater than the maximum value of the data reliability threshold interval, it means that the correlation between the formation density data and the Poisson's ratio data is high, and the reliability of the measured data is high. This means that the abnormality of the measured data is low, and the data can be used more reliably for geological exploration and drilling engineering analysis, indicating that the data distortion is low and a high reliability signal is generated;

[0084] On the contrary, if the response degree is less than the minimum value of the data reliability threshold interval, it means that the correlation between the formation density data and the Poisson's ratio data is low, there are measurement anomalies, the reliability of the data is poor, indicating that the data distortion is high, and a low reliability signal is generated;

[0085] If the response level is within the data reliability threshold range, it means that the correlation between the formation density data and the Poisson's ratio data is within an acceptable range. This means that there is a certain degree of correlation between the two, but it is not very significant or very low. Therefore, it is necessary to further consider using some auxiliary features to improve the reliability and abnormality of the data and generate a medium reliability signal.

[0086] The present invention calculates the rate of change of formation density and Poisson's ratio data, calculates their standard deviations respectively to evaluate the degree of fluctuation, and compares them with the set fluctuation threshold to determine the fluctuation level of the data. According to the fluctuation level, the interpolation method or truncation method is flexibly selected to process the data to ensure the reliability and processing flexibility of the aligned data. By comparing the response level and data reliability threshold of formation density data and Poisson's ratio data, a corresponding reliable signal is generated to analyze the abnormality of the measured data. It can effectively evaluate and process drilling data, ensure the quality and authenticity of the data measured by drilling, provide important reference information for geological exploration and drilling engineering, and help improve the accuracy and reliability of data analysis, thereby promoting the progress and development of related projects.

[0087] S3 includes the following:

[0088] After obtaining the non-low reliability signal, the formation density data and Poisson's ratio data obtained per unit time are respectively used as target variables, and the features to be screened for auxiliary analysis are counted. The features to be screened are screened by the target variables. The process is as follows:

[0089] Step 1, determining the timestamp for obtaining the formation density data and Poisson's ratio data per unit time;

[0090] Step 2: Count the characteristics of the drilling process;

[0091] During the drilling process, the features that can be counted include but are not limited to:

[0092] Drilling depth: The depth during the drilling process, that is, the drilling depth, usually measured in meters.

[0093] Drill speed: The speed at which the drill bit rotates per minute, usually measured in revolutions per minute.

[0094] Weight on Bit: The pressure exerted by the drill bit on the formation, usually measured in force per unit area (e.g., Pascals).

[0095] Drilling speed: The speed at which the drill bit drills into the formation, usually measured in meters per minute.

[0096] Core sample: A rock sample taken from the bottom of a well for geological analysis and physical property testing.

[0097] Drilling fluid properties: The properties of drilling fluid, including density, viscosity, pH value, etc.

[0098] Drilling fluid circulation rate: The circulation rate of drilling fluid in the wellbore, usually measured in volume per minute.

[0099] Well temperature: The temperature at the bottom of the well, which can affect the performance of the drill bit and drilling tools.

[0100] Formation pressure: The pressure conditions in the formation can affect the stability of drilling.

[0101] Step three, under the condition of a timestamp corresponding to the unit time for obtaining the formation density data and Poisson's ratio data, for each feature, the data set is grouped according to the value of the feature, and for each value of each feature, the category distribution of the target variable under the value is calculated, and its Gini coefficient is calculated, and the weighted average Gini coefficient of each feature is calculated as a measure of the diversity of the feature to the target variable.

[0102] Step 4: Calculate the absolute value of the difference between the weighted average Gini coefficients of each target variable corresponding to each feature, and then conduct classification discussion based on the response degree to obtain the classification uncertainty index. For example, it can be obtained by the following calculation formula: ;

[0103] in: The classification uncertainty index of each feature indicates the classification ability of each feature for the two target variables; Indicates that under the given feature value, the target variable A category of The sum of squares of the conditional probabilities, is the target variable The number of categories, where the target variable Used to refer to the density of a formation; Indicates that under the given feature value, the target variable A category of The sum of squares of the conditional probabilities, Target variable The number of categories, where the target variable Used to refer to Poisson's ratio;

[0104] The purpose of calculating the absolute value of the difference between the sum of squares of the conditional probabilities of the same feature for different target variables is to analyze the degree of difference in the classification ability of the same feature for two target variables. By comparing the absolute value of the difference between the sum of squares of the conditional probabilities of the two target variables, the difference in the classification ability of the feature between different target variables can be evaluated. If the difference is small, it means that the classification ability of the feature for the two target variables is similar; if the difference is large, it means that there is a large difference in the classification ability of the feature between different target variables. This analysis helps to determine which features have a higher discriminatory ability for the prediction of different target variables, so that features can be better selected and the performance of the model can be optimized.

[0105] and They are respectively the case where the response level is within the data reliability threshold range. and The preset scaling factor of

[0106] The preset proportionality factor is a parameter used to adjust and calibrate the relative weights of variables under different conditions to ensure that the model or calculation results are balanced and accurate under different circumstances. Specifically, and They correspond to the correction of the difference in the sum of squares of classification probabilities and the proportional factor when the response level is within the data reliability threshold range. The function is to adjust the proportion of different influencing factors so that the calculation results have appropriate sensitivity and responsiveness in specific situations, avoiding calculation distortion due to imbalance of different factors. These coefficients are determined by experience or model preset values ​​to ensure the robustness of calculation under different data fluctuations or reliability conditions.

[0107] It is the correction factor when the response level is not within the data reliability threshold range;

[0108] The correction coefficient is a key parameter used to adjust the calculation results when the response level is not within the data reliability value range. It is used to adaptively correct the model or formula derivation originally based on reliable data to ensure that relatively accurate and stable results can still be obtained in abnormal or unreliable data scenarios. Specifically, by correcting the deviation caused by data fluctuations or anomalies, the balance and rationality of the calculation process are maintained, so that even in the case of deviation from standard conditions, the output of the model still has practical reference value and operability.

[0109] and are the upper limits of the response level and data reliability threshold interval, is the data reliability threshold interval.

[0110] Reliable threshold interval: a numerical interval used to quantify and evaluate the range of data reliability. It means that within this interval, the credibility and accuracy of the data can meet the preset standards or requirements, and can be used as an effective input for the model or algorithm to ensure the validity of the final calculation results. This interval is determined through experiments or experience to clarify when the data can be considered to be sufficiently reliable, thereby guiding relevant decisions.

[0111] Upper limit of the reliable threshold interval: As the maximum value of the reliable threshold interval, it indicates that the data reliability is no longer considered reliable above this value. This value is used to define when the data deviates from the expected reliable range. When it exceeds this upper limit, it indicates that the data may be affected by external interference or noise, so the data needs to be corrected or re-collected to ensure that it meets the expected reliability standards.

[0112] The response level is introduced to address the impact of the reliability of the obtained formation density and Poisson's ratio data on the final screening of features for auxiliary analysis when medium reliability signals are obtained.

[0113] When a medium reliability signal is obtained, that is, when the formation density number and Poisson's ratio data are relatively reliable, the response degree is introduced to consider the impact of the reliability of these two data on the final screening of features for auxiliary analysis. The response degree measures the degree of correlation between the formation density data and the Poisson's ratio data, that is, whether their changing trends are similar or consistent. If the response degree is high, it means that there is a strong correlation between the two data and the changing trends are consistent, thereby improving the reliability of the data. On the contrary, if the response degree is low, it means that the correlation between the data is weak and there may be certain inconsistencies, thereby reducing the reliability of the data.

[0114] The purpose of introducing the response degree is to comprehensively consider the reliability factors of the data in the feature screening process. By analyzing the response degree, the reliability information can be incorporated into the feature screening evaluation process to further optimize the accuracy of feature selection. Specifically, if the correlation between the formation density data and the Poisson's ratio data is high, the information of these two data can be more trusted, and the relevant features can be selected more confidently for auxiliary analysis; if the correlation is low, it may be necessary to treat these data with caution and take more verification and calibration measures to ensure that the final selected features have high reliability and accuracy. Therefore, the introduction of the response degree helps to improve the accuracy and reliability of feature screening, thereby improving the quality and effect of data analysis.

[0115] The smaller the classification uncertainty index is, the smaller the uncertainty of the feature's classification ability for the two target variables is, that is, the feature value can better distinguish the different categories of the target variable, and therefore has a higher correlation with the target variable; on the contrary, the larger the classification uncertainty index is, that is, the feature value cannot well distinguish the different categories of the target variable, and therefore has a lower correlation with the target variable.

[0116] Step 5, the classification uncertainty index is compared with the screening threshold;

[0117] If the classification uncertainty index is greater than or equal to the screening threshold, it means that the classification ability of the feature is poor, that is, the classification effect of the feature on the target variable is uncertain or unreliable. This may mean that the feature cannot distinguish different categories well or does not have strong predictive ability, so it is not suitable for further analysis or modeling, and therefore has no positive significance for drilling data analysis and needs to be discarded to generate a discard signal.

[0118] On the contrary, if the classification uncertainty index is less than the screening threshold, it means that the classification ability of the feature is good, that is, the classification effect of the feature on the target variable is relatively certain and reliable. This means that the feature can effectively distinguish different categories or has strong predictive ability, so it is suitable for further data analysis and modeling to generate a hiring signal.

[0119] Step six, count the characteristics of life recruitment signals, sort them from small to large according to the classification uncertainty index, and delete them according to the preset recruitment number.

[0120] The present invention determines the formation density data and Poisson's ratio data that can be used as target variables based on the comparison results of the response degree and the data reliability threshold interval, and calculates the classification uncertainty index for each feature under the condition that the feature data and the target variable are in the same timestamp to evaluate the classification ability of the feature to the target variable. By comparing the classification uncertainty index and the preset screening threshold, it can be judged whether the classification ability of the feature is sufficiently reliable, so as to decide whether to retain the feature for further analysis. The response degree is introduced in a timely manner and taken into consideration in the feature screening process to correct the screening quality of the classification uncertainty index. The features are sorted according to the classification uncertainty index and deleted according to the preset number of admissions to obtain the final auxiliary analysis feature set. In this way, the features with high correlation and good classification ability to the target variable can be effectively identified, thereby improving the accuracy and reliability of drilling data analysis.

[0121] Embodiment 2: Figure 2 The present invention provides a data processing system for mining drilling, including a time alignment module, a data merging module, a target variable selection module, a feature analysis module and a feature screening module;

[0122] The time alignment module is used to select the interpolation method according to the data volatility, align the formation density and Poisson's ratio data, and send the alignment results to the data merging module;

[0123] The data merging module constructs a correlation matrix with the aligned data, calculates the response degree to evaluate the data reliability, and sends the response degree to the target variable selection module;

[0124] The target variable selection module determines the formation density and Poisson's ratio data that meet the expected reliability as the target variable, and sends the target variable to the feature analysis module;

[0125] The feature analysis module combines each feature with the target variable, calculates the classification uncertainty index, evaluates the classification ability of the feature, and sends the classification uncertainty index to the feature screening module;

[0126] The feature screening module screens features based on the classification uncertainty index to determine whether to retain them for further analysis.

[0127] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters and thresholds in the formula are set by technicians in this field according to actual conditions.

[0128] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination thereof. When implemented by software, the above embodiments may be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a computer-readable storage medium, or may be transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions may be transmitted from one website, computer, server or data center to another website, computer, server or data center by wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium may be any available medium that a computer can access or a data storage device such as a server or data center that contains one or more available media sets. The available medium may be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium may be a solid-state hard disk.

[0129] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0130] In the several embodiments provided in the present application, it should be understood that the disclosed systems and devices can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0131] 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 distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0132] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0133] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, and other media that can store program codes.

[0134] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0135] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A data processing method for mining drilling, characterized in that: S1. Based on the corresponding acquisition time point, the corresponding interpolation method is adapted according to the data volatility, and the formation density data and Poisson's ratio data are aligned according to the time point; S2. Based on the formation density data and Poisson's ratio data obtained under the alignment conditions, they are merged into a unified data set, and the correlation matrix is ​​constructed. The gray value between pixels is used to calculate the response degree to evaluate the reliability of the measurement data. The response degree calculation formula is as follows: ; in, is the degree of response, is the number of data points, is the first Line and The value of the column element indicates the The formation density data and The degree of response of Poisson's ratio data; S3. Taking the formation density data and Poisson's ratio data with expected reliability as the target variables, each feature of the drilling process is analyzed in combination with the target variables in turn to obtain the corresponding classification uncertainty index, and the features are screened and selected; S3 includes the following: Step 1, determining the timestamp for obtaining the formation density data and Poisson's ratio data per unit time; Step 2: Count the characteristics of the drilling process; Step 3: Under the condition of the timestamp corresponding to the unit time of obtaining the formation density data and Poisson's ratio data, for each feature, the data set is grouped according to the value of the feature, and for each value of each feature, the category distribution of the target variable under the corresponding value is calculated, and its Gini coefficient is calculated, and the weighted average Gini coefficient of each feature is calculated as a measure of the diversity of the feature to the target variable; Step 4: Calculate the absolute value of the difference between the weighted average Gini coefficients of each target variable corresponding to each feature, and then conduct classification discussion based on the response degree to obtain the classification uncertainty index. ; Obtained by the following calculation formula: ; in: Indicates that under the given feature value, the target variable A category of The sum of squares of the conditional probabilities, is the target variable The number of categories, where the target variable Used to refer to the density of a formation; Indicates that under the given feature value, the target variable A category of The sum of squares of the conditional probabilities, is the target variable The number of categories, where the target variable Used to refer to Poisson's ratio; and They are the preset proportional coefficients when the response level is within the data reliability threshold range; It is the correction factor when the response level is not within the data reliability threshold range; and are the upper limits of the response level and data reliability threshold interval, is the data reliability threshold interval; Step 5, the classification uncertainty index is compared with the screening threshold; If the classification uncertainty index is greater than or equal to the screening threshold, it means that the classification effect of the feature on the target variable is uncertain or unreliable, and a rejection signal is generated; conversely, if the classification uncertainty index is less than the screening threshold, it means that the classification effect of the feature on the target variable is relatively certain and reliable, and a recruitment signal is generated; Step six: Count the features of the acceptance signal, sort them from small to large according to the classification uncertainty index, and delete them according to the preset number of acceptances.

2. A data processing method for mining drilling according to claim 1, characterized in that: S1 includes the following: Perform timestamp alignment on formation density data and Poisson's ratio data, including the following: According to the pre-set time window of the formation density data and Poisson's ratio data, that is, the time range in which the two data exist together, the selected time window includes the intersection of the two data; For the case where the time points are not completely aligned, evaluate the fluctuation level of the data and select the corresponding difference method according to the fluctuation level to interpolate the formation density and Poisson's ratio data to the same time point. The execution steps are as follows: Step 1: Calculate the first-order difference of the time series data of formation density and Poisson's ratio in unit time, that is, calculate the change between adjacent time points, perform first-order difference calculation on the formation density data and Poisson's ratio data, and obtain the change rate series of formation density and Poisson's ratio; Step 2: Calculate the standard deviation of the change rate series of formation density and Poisson's ratio to evaluate their fluctuation degree; Compare the standard deviation with the fluctuation threshold. If the standard deviation is greater than or equal to the fluctuation threshold, it means that the data fluctuates greatly, generating a high fluctuation signal. The truncation method is suitable for aligning the data to the same time point. If the standard deviation value is less than the fluctuation threshold, it means that the data fluctuation is small, generating a low fluctuation signal, which is suitable for interpolation method to align the data to the same time point.

3. A data processing method for mining drilling according to claim 2, characterized in that: S2 includes the following: Step 1: Perform minimum-maximum normalization processing on the formation density data and Poisson's ratio data per unit time and convert them into the same scale to eliminate the influence of dimension; Step 2: The standardized formation density data and Poisson's ratio data are combined into a correlation matrix, where each element in the correlation matrix represents the degree of correlation between two data points; Step three, calculate the response degree between the formation density data and the Poisson's ratio data, using the gray value between pixels to calculate, indicating the response degree between adjacent pixels.

4. A data processing method for mining drilling according to claim 3, characterized in that: Compare the response level with the data reliability threshold interval; If the response degree is greater than the maximum value of the data reliability threshold interval, it means that the correlation between the formation density data and the Poisson's ratio data is high, generating a high reliability signal; On the contrary, if the response degree is less than the minimum value of the data reliability threshold interval, it means that the correlation between the formation density data and the Poisson's ratio data is low, generating a low reliability signal; If the response degree is within the data reliability threshold interval, it means that the correlation between the formation density data and the Poisson's ratio data is within an acceptable range, and a medium reliability signal is generated.

5. A data processing system for mining drilling, used to implement the data processing method according to any one of claims 1 to 4, comprising a time alignment module, a data merging module, a target variable selection module, a feature analysis module and a feature screening module; The time alignment module is used to select the interpolation method according to the data volatility, align the formation density and Poisson's ratio data, and send the alignment results to the data merging module; The data merging module constructs a correlation matrix with the aligned data, calculates the response degree to evaluate the data reliability, and sends the response degree to the target variable selection module; The target variable selection module determines the formation density and Poisson's ratio data that meet the expected reliability as the target variable, and sends the target variable to the feature analysis module; The feature analysis module combines each feature with the target variable, calculates the classification uncertainty index, evaluates the classification ability of the feature, and sends the classification uncertainty index to the feature screening module; The feature screening module screens features based on the classification uncertainty index to determine whether to retain them for further analysis.

Images