Real-time monitoring system and method for core drilling production

Abstract

The present application relates to the field of intelligent monitoring, and specifically discloses a real-time monitoring system and method for core drilling production, which collects the time series of drill position data, the time series of drilling pressure value, the time series of drill speed value and the time series of drill temperature value, and uses AI-based data processing and analysis technology to perform time series analysis and association on the drill position data, the drilling pressure value, the drill speed value and the drill temperature value, so as to automatically judge whether the working state of the drill bit is normal based on the time series dynamic interaction characteristics of the drill bit travel state and the drill bit process. In this way, the drilling process can be fully monitored so as to detect abnormal situations in time to ensure the safety of operators.

Description

Technical Field

[0001] The present application relates to the field of intelligent monitoring, and more specifically, to a real-time monitoring system and method for core drilling production. Background Art

[0002] Core drilling is a geological exploration technology that uses a core drill to drill cylindrical samples (cores) of underground rock formations. Since the drilling environment is often dangerous, monitoring the core drilling process can monitor the operating status, prevent accidents, and ensure the safety of operators.

[0003] However, the traditional monitoring system for core drilling production lacks sufficient sensors to collect data in real time, resulting in an inability to respond promptly to abnormal situations that occur during the drilling process. At the same time, there is a lack of comprehensive monitoring and analysis of the drilling process, which increases the occurrence of dangerous accidents.

[0004] Therefore, an optimized real-time monitoring system for core drilling production is desired. Summary of the invention

[0005] In order to solve the above technical problems, the present application is proposed. The embodiment of the present application provides a real-time monitoring system and method for core drilling production, which collects the time series of drill bit position data, the time series of drilling pressure value, the time series of drill bit speed value and the time series of drill bit temperature value, and uses AI-based data processing and analysis technology to perform time series analysis and association on the drill bit position data, the drilling pressure value, the drill bit speed value and the drill bit temperature value, so as to automatically judge whether the working state of the drill bit is normal based on the time series dynamic interaction characteristics of the drill bit travel state and the drill bit process. In this way, the drilling process can be fully monitored so as to detect abnormal situations in time to ensure the safety of operators.

[0006] According to one aspect of the present application, a real-time monitoring system for core drilling production is provided, which includes: a drill bit parameter acquisition module, which is used to obtain the time series of drill bit position data, the time series of drilling pressure values, the time series of drill bit rotation speed values, and the time series of drill bit temperature values ​​collected by a sensor group; a drill bit travel state feature extraction module, which is used to extract the drill bit travel state features of the time series of the drill bit position data to obtain a drill bit travel state time series associated feature vector; a drill bit travel process feature extraction module, which is used to arrange the time series of drilling pressure values, the time series of drill bit rotation speed values, and the time series of drill bit temperature values ​​into a drill bit travel process multimodal time series matrix according to parameter sample dimensions and time dimensions, and then input the matrix into a drill bit travel process feature extractor based on a hole convolutional network to obtain a drill bit travel process multimodal time series associated feature map; a feature enhancement module, which is used to extract the drill bit travel process multimodal time series associated feature map. The state temporal association feature map is input into the feature enhancement module based on significance-globality to obtain the multimodal temporal association feature map of the enhanced drill bit travel process; the association focusing module is used to input the multimodal temporal association feature map of the enhanced drill bit travel process into the feature multiscale focusing module to obtain the multimodal temporal association focusing feature vector of the enhanced drill bit travel process; the feature dynamic interaction module is used to input the multimodal temporal association focusing feature vector of the enhanced drill bit travel process and the drill bit travel state temporal association feature vector into the feature dynamic interaction module based on prior distribution to obtain the drill bit travel state-drill bit process feature dynamic interaction representation vector as the drill bit travel state-drill bit process feature dynamic interaction representation feature; the monitoring result generation module is used to obtain the monitoring result based on the drill bit travel state-drill bit process feature dynamic interaction representation feature, and the monitoring result is used to indicate whether the working state of the drill bit is normal.

[0007] In the above-mentioned real-time monitoring system for core drilling production, the drill bit travel state feature extraction module is used to: input the time series of the drill bit position data into the drill bit travel state feature extractor based on the GRU unit to obtain the drill bit travel state time series correlation feature vector.

[0008] In the above-mentioned real-time monitoring system for core drilling production, the feature enhancement module includes: a feature expansion unit, which is used to expand the multimodal time series correlation feature map of the drill bit travel process into a sequence of multimodal time series correlation feature vectors of the drill bit travel process; a prominent feature calculation unit, which is used to extract the maximum value of each multimodal time series correlation feature vector of the drill bit travel process in the sequence of multimodal time series correlation feature vectors of the drill bit travel process to obtain a prominent multimodal time series correlation feature vector of the drill bit travel process; and a global feature calculation unit, which is used to extract the average value of each multimodal time series correlation feature vector of the drill bit travel process in the sequence of multimodal time series correlation feature vectors of the drill bit travel process to obtain a global drilling A multi-modal temporal correlation feature vector of the head moving process; a prominent feature one-dimensional convolution encoding unit, which is used to input the multi-modal temporal correlation convolution encoding vector of the protruding drill bit moving process obtained after one-dimensional convolution encoding of the multi-modal temporal correlation feature vector of the protruding drill bit moving process into the ReLU function to obtain a multi-modal temporal correlation activation encoding feature vector of the protruding drill bit moving process; a prominent feature activation unit, which is used to perform matrix multiplication of the prominent point convolution feature vector obtained after point convolution encoding of the multi-modal temporal correlation activation encoding feature vector of the protruding drill bit moving process with the first weight matrix to obtain a multi-modal temporal correlation activation feature vector of the protruding feature drill bit moving process; a global feature one-dimensional convolution encoding unit, which is used to The global drill bit travel process multimodal temporal association convolution coding vector obtained after one-dimensional convolution coding of the global drill bit travel process multimodal temporal association feature vector is input into the ReLU function to obtain the global drill bit travel process multimodal temporal association activation coding feature vector; a global feature activation unit is used to perform matrix multiplication of the global point convolution feature vector obtained after point convolution coding of the global drill bit travel process multimodal temporal association activation coding feature vector with a second weight matrix to obtain the global feature drill bit travel process multimodal temporal association activation feature vector; a feature fusion unit is used to fuse the prominent feature drill bit travel process multimodal temporal association activation feature vector and the global feature drill bit After the multimodal temporal association activation feature vector of the traveling process is obtained to obtain the global-prominent feature multimodal temporal association feature vector of the traveling process of the drill bit, the global-prominent feature multimodal temporal association feature vector of the traveling process of the drill bit is nonlinearly activated to obtain the global-prominent feature multimodal temporal association weight feature vector of the traveling process of the drill bit; a feature weighting unit is used to use the global-prominent feature multimodal temporal association weight feature vector of the traveling process of the drill bit as a weight, perform position point multiplication on the sequence of the multimodal temporal association feature vectors of the traveling process of the drill bit, and add the sequence of the multimodal temporal association feature vectors of the traveling process of the drill bit to obtain the sequence of the multimodal temporal reinforcement association feature vectors of the traveling process of the drill bit;and a feature aggregation unit, configured to aggregate the sequence of the multi-modal temporal enhanced associated feature vectors of the drill bit travel process in accordance with the manner of expanding the multi-modal temporal associated feature graph of the drill bit travel process to obtain the enhanced multi-modal temporal associated feature graph of the drill bit travel process. ;

[0009] In the above-mentioned real-time monitoring system for core drilling production, the feature fusion unit is used to: perform feature fusion on the prominent feature drill bit travel process multimodal temporal association activation feature vector and the global feature drill bit travel process multimodal temporal association activation feature vector to obtain the global-prominent feature drill bit travel process multimodal temporal association feature vector; input the global-prominent feature drill bit travel process multimodal temporal association feature vector into the tanh function to obtain a first activated global-prominent feature drill bit travel process multimodal temporal association feature vector; input the global-prominent feature drill bit travel process multimodal temporal association feature vector into the sigmoid function to obtain a second activated global-prominent feature drill bit travel process multimodal temporal association feature vector; and, multiply the first activated global-prominent feature drill bit travel process multimodal temporal association feature vector and the second activated global-prominent feature drill bit travel process multimodal temporal association feature vector by position point to obtain the global-prominent feature drill bit travel process multimodal temporal association weight feature vector.

[0010] In the above-mentioned real-time monitoring system for core drilling production, the correlation focusing module includes: an upsampling unit, which is used to upsample the multimodal temporal correlation feature map of the enhanced drill bit travel process to obtain the upsampled multimodal temporal correlation feature map of the enhanced drill bit travel process; a global mean pooling processing unit, which is used to perform global mean pooling processing on each feature matrix along the channel dimension in the upsampled multimodal temporal correlation feature map of the enhanced drill bit travel process to obtain the upsampled multimodal temporal correlation channel feature vector of the enhanced drill bit travel process; a temporal correlation global processing unit, which is used to perform point convolution encoding on the multimodal temporal correlation channel feature vector of the upsampled multimodal temporal correlation channel of the enhanced drill bit travel process to obtain the multimodal temporal correlation channel of the enhanced drill bit travel process. A sequence-related global feature vector; a channel modulation unit, used for performing point convolution encoding on the multi-modal temporal correlation feature map of the up-sampled enhanced drill bit travel process to obtain a channel-modulated up-sampled enhanced drill bit travel process multi-modal temporal correlation feature map; a temporal correlation local processing unit, used for performing two-dimensional convolution encoding on the channel-modulated up-sampled enhanced drill bit travel process multi-modal temporal correlation feature map to obtain a multi-modal temporal correlation local feature vector of the enhanced drill bit travel process; and a multi-scale correlation unit, used for fusing the multi-modal temporal correlation global feature vector of the enhanced drill bit travel process and the multi-modal temporal correlation local feature vector of the enhanced drill bit travel process to obtain a multi-modal temporal correlation focused feature vector of the enhanced drill bit travel process.

[0011] In the above-mentioned real-time monitoring system for core drilling production, the feature dynamic interaction module includes: a feature corresponding calculation unit, which is used to calculate the position-by-position division of the multi-modal time series correlation focusing feature vector of the enhanced drill bit travel process and the time series correlation feature vector of the drill bit travel state to obtain the drill bit travel process-travel state response interaction feature vector; a first feature weighting unit, which is used to calculate the product of the drill bit travel process-travel state response interaction feature vector and the first weight coefficient to obtain the first weight feature vector, and then use the feature value of each position in the first weight feature vector as the exponent of the natural constant to calculate the exponential function value with the natural constant as the base to obtain the first drill bit travel process-travel state response interaction exponential function feature vector; a first Gaussian calculation unit, which is used to calculate the product of the first drill bit travel process-travel state response interaction exponential function feature vector and the first Gaussian distribution function coefficient to obtain the first drill bit travel process-travel state response interaction feature vector; the first feature weighting unit is used to calculate the drill bit After the product of the travel process-travel state response interaction characteristic vector and the second weight coefficient is obtained to obtain the second weight characteristic vector, the characteristic value of each position in the second weight characteristic vector is used as the exponent of the natural constant to calculate the exponential function value with the natural constant as the base to obtain the second drill bit travel process-travel state response interaction exponential function characteristic vector; a second Gaussian calculation unit is used to calculate the product of the second drill bit travel process-travel state response interaction exponential function characteristic vector and the second Gaussian distribution function coefficient to obtain the second drill bit travel process-travel state response interaction characteristic vector; and a dynamic interaction unit is used to calculate the positional addition between the first drill bit travel process-travel state response interaction characteristic vector and the second drill bit travel process-travel state response interaction characteristic vector to obtain the drill bit travel state-drill bit process characteristic dynamic interaction representation vector; wherein the first Gaussian distribution function coefficient and the second Gaussian distribution function coefficient are both generated by a Gaussian distribution random number function with a mean of 0 and a variance of 1.

[0012] In the above-mentioned real-time monitoring system for core drilling production, the monitoring result generation module is used to: input the drill bit travel status-drill bit process feature dynamic interaction representation vector into the classifier-based real-time monitoring result generator to obtain the monitoring result.

[0013] In the above-mentioned real-time monitoring system for core drilling production, it also includes a training module for training the drill bit travel state feature extractor based on the GRU unit, the drill bit travel process feature extractor based on the void convolution network, the feature enhancement module based on significance-globality, the feature multi-scale focusing module, the feature dynamic interaction module based on prior distribution and the real-time monitoring result generator based on classifier.

[0014] In the above-mentioned real-time monitoring system for core drilling production, the training module includes: a training data acquisition unit, used to acquire training data, the training data includes a time series of training drill bit position data, a time series of training drilling pressure values, a time series of training drill bit rotation speed values ​​and a time series of training drill bit temperature values, and a real monitoring result, the real monitoring result is a real value of whether the working state of the drill bit is normal; a training drill bit travel state time series association unit, used to input the time series of the training drill bit position data into the drill bit travel state feature extractor based on the GRU unit to obtain a training drill bit travel state time series association feature vector; a drill bit travel state time series association unit, used to input the time series of the training drill bit position data into the drill bit travel state feature extractor based on the GRU unit to obtain a training drill bit travel state time series association feature vector; a multimodal time series association unit for the drill bit travel process, used for arranging the time series of the training drilling pressure value, the time series of the training drill bit speed value and the time series of the training drill bit temperature value according to the parameter sample dimension and the time dimension into a multimodal time series matrix for the drill bit travel process, and then inputting the matrix into the drill bit travel process feature extractor based on the hole convolution network to obtain a multimodal time series association feature map for the drill bit travel process; a training feature enhancement unit, used for inputting the multimodal time series association feature map for the drill bit travel process into the feature enhancement module based on significance-globality to obtain a multimodal time series association feature map for the drill bit travel process; a training multimodal time series association unit, used for inputting the multimodal time series association feature map for the drill bit travel process into the feature enhancement module based on significance-globality to obtain a multimodal time series association feature map for the drill bit travel process; a training multimodal time series association unit, used for The training monitoring result generating unit is used to input the training enhanced drill bit travel process multimodal temporal correlation feature map into the feature multiscale focusing module to obtain the training enhanced drill bit travel process multimodal temporal correlation focusing feature vector; the training feature dynamic interaction unit is used to input the training enhanced drill bit travel process multimodal temporal correlation focusing feature vector and the training drill bit travel state temporal correlation feature vector into the feature dynamic interaction module based on prior distribution to obtain the training drill bit travel state-drill bit process feature dynamic interaction representation vector; the training monitoring result generating unit is used to input the training drill bit travel state-drill bit process feature dynamic interaction representation vector into the classifier based on A real-time monitoring result generator is provided to obtain a training monitoring result; a classification loss unit is used to calculate the cross entropy loss function value between the training monitoring result and the real monitoring result to obtain a classification loss function value; and a model training unit is used to train the drill travel state feature extractor based on the GRU unit, the drill travel process feature extractor based on the void convolution network, the feature enhancement module based on significance-globality, the feature multi-scale focusing module, the feature dynamic interaction module based on the prior distribution and the real-time monitoring result generator based on the classifier based on the classification loss function value and through back propagation of gradient descent.

[0015] According to another aspect of the present application, a real-time monitoring method for core drilling production is provided, which includes: acquiring a time series of drill bit position data, a time series of drill bit pressure values, a time series of drill bit rotation speed values, and a time series of drill bit temperature values ​​collected by a sensor group; performing drill bit travel state feature extraction on the time series of drill bit position data to obtain a drill bit travel state time series associated feature vector; arranging the time series of drill bit pressure values, the time series of drill bit rotation speed values, and the time series of drill bit temperature values ​​into a drill bit travel process multimodal time series matrix according to parameter sample dimensions and time dimensions, and then inputting the matrix into a drill bit travel process feature extractor based on a hole convolutional network to obtain a drill bit travel process multimodal time series associated feature map; and inputting the drill bit travel process multimodal time series associated feature map into a drill bit travel process multimodal time series associated feature map. The feature enhancement module based on significance-globality is input to obtain a multi-modal temporal correlation feature map of the enhanced drill bit travel process; the multi-modal temporal correlation feature map of the enhanced drill bit travel process is input into the feature multi-scale focusing module to obtain a multi-modal temporal correlation focused feature vector of the enhanced drill bit travel process; the multi-modal temporal correlation focused feature vector of the enhanced drill bit travel process and the drill bit travel state temporal correlation feature vector are input into the feature dynamic interaction module based on prior distribution to obtain a drill bit travel state-drill bit process feature dynamic interaction representation vector as a drill bit travel state-drill bit process feature dynamic interaction representation feature; based on the drill bit travel state-drill bit process feature dynamic interaction representation feature, a monitoring result is obtained, and the monitoring result is used to indicate whether the working state of the drill bit is normal.

[0016] Compared with the prior art, the present application provides a real-time monitoring system and method for core drilling production, which collects the time series of drill bit position data, the time series of drilling pressure value, the time series of drill bit speed value and the time series of drill bit temperature value, and uses AI-based data processing and analysis technology to perform time series analysis and correlation on the drill bit position data, the drilling pressure value, the drill bit speed value and the drill bit temperature value, so as to automatically judge whether the working state of the drill bit is normal based on the time series dynamic interaction characteristics of the drill bit travel state and the drill bit process. In this way, the drilling process can be fully monitored so as to detect abnormal situations in time to ensure the safety of operators. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0018] Figure 14 is a block diagram of a real-time monitoring system for core drilling production according to an embodiment of the present application.

[0019] Figure 2 The flowchart is a real-time monitoring method for core drilling production according to an embodiment of the present application. DETAILED DESCRIPTION

[0020] The embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as being limited to the embodiments described herein, but rather these embodiments are provided for a more thorough and complete understanding of the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are only for exemplary purposes and are not intended to limit the scope of protection of the present disclosure.

[0021] It should be understood that the various steps described in the method embodiments of the present disclosure may be performed in different orders and / or in parallel. In addition, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present disclosure is not limited in this respect.

[0022] Core drilling is a geological exploration technology that uses a core drill to drill cylindrical samples (cores) of underground rock formations. Since the drilling environment is often dangerous, monitoring the core drilling process can monitor the operating status, prevent accidents, and ensure the safety of operators.

[0023] However, the traditional monitoring system for core drilling production lacks sufficient sensors to collect data in real time, resulting in an inability to respond promptly to abnormal situations that occur during the drilling process. At the same time, there is a lack of comprehensive monitoring and analysis of the drilling process, which increases the occurrence of dangerous accidents.

[0024] Therefore, in response to the above technical problems, the technical concept of this application is to collect the time series of drill bit position data, the time series of drilling pressure value, the time series of drill bit speed value and the time series of drill bit temperature value, and use AI-based data processing and analysis technology to perform time series analysis and correlation on the drill bit position data, the drilling pressure value, the drill bit speed value and the drill bit temperature value, so as to automatically judge whether the working state of the drill bit is normal based on the time series dynamic interaction characteristics of the drill bit travel state and the drilling process. In this way, the drilling process can be fully monitored so as to detect abnormal situations in time to ensure the safety of operators.

[0025] Figure 1 FIG. 1 is a block diagram of a real-time monitoring system for core drilling production according to an embodiment of the present application. Figure 1As shown, according to the embodiment of the present application, the real-time monitoring system 100 for core drilling production includes: a drill bit parameter acquisition module 110, which is used to obtain the time series of drill bit position data, the time series of drilling pressure values, the time series of drill bit speed values, and the time series of drill bit temperature values ​​collected by the sensor group; a drill bit travel state feature extraction module 120, which is used to extract the drill bit travel state feature of the time series of the drill bit position data to obtain the drill bit travel state time series associated feature vector; a drill bit travel process feature extraction module 130, which is used to arrange the time series of the drilling pressure values, the time series of the drill bit speed values, and the time series of the drill bit temperature values ​​into a drill bit travel process multimodal time series matrix according to the parameter sample dimension and the time dimension, and then input the matrix into a drill bit travel process feature extractor based on a void convolutional network to obtain a drill bit travel process multimodal time series associated feature map; a feature enhancement module 140, which is used to extract the drill bit travel process multimodal time series associated feature map; The time series correlation feature map is input into a feature enhancement module based on significance-globality to obtain a multimodal time series correlation feature map of the enhanced drill bit travel process; the correlation focusing module 150 is used to input the multimodal time series correlation feature map of the enhanced drill bit travel process into a feature multiscale focusing module to obtain a multimodal time series correlation focusing feature vector of the enhanced drill bit travel process; the feature dynamic interaction module 160 is used to input the multimodal time series correlation focusing feature vector of the enhanced drill bit travel process and the drill bit travel state time series correlation feature vector into a feature dynamic interaction module based on prior distribution to obtain a drill bit travel state-drill bit process feature dynamic interaction representation vector as a drill bit travel state-drill bit process feature dynamic interaction representation feature; and the monitoring result generation module 170 is used to obtain a monitoring result based on the drill bit travel state-drill bit process feature dynamic interaction representation feature, and the monitoring result is used to indicate whether the working state of the drill bit is normal.

[0026] In the embodiment of the present application, the drill bit parameter acquisition module 110 is used to obtain the time series of the drill bit position data, the time series of the drilling pressure value, the time series of the drill bit speed value, and the time series of the drill bit temperature value collected by the sensor group. It should be understood that the drill bit position data, the drilling pressure value, the drill bit speed value, and the drill bit temperature value are important monitoring parameters in the drilling process. Based on this, in the technical solution of the present application, the time series of the drill bit position data, the time series of the drilling pressure value, the time series of the drill bit speed value, and the time series of the drill bit temperature value collected by the sensor group are obtained, which can realize the intelligent management of the drilling process and improve the accuracy and safety of the operation.

[0027] In the embodiment of the present application, the drill bit travel state feature extraction module 120 is used to extract the drill bit travel state feature of the time series of the drill bit position data to obtain the drill bit travel state time series correlation feature vector. Specifically, in the embodiment of the present application, the drill bit travel state feature extraction module is used to: input the time series of the drill bit position data into the drill bit travel state feature extractor based on the GRU unit to obtain the drill bit travel state time series correlation feature vector. Accordingly, considering that the time series dependencies of the drill bit position data between different time spans in the time series of the drill bit position data are different, and GRU is a recurrent neural network (RNN) variant suitable for processing time series data, which can effectively capture the time dependencies of different time lengths in the data. Based on this, in the technical solution of the present application, the time series of the drill bit position data is input into the drill bit travel state feature extractor based on the GRU unit to capture and extract the time series correlation of the drill bit position data between different time lengths, thereby obtaining the drill bit travel state time series correlation feature vector.

[0028] In the embodiment of the present application, the drill bit travel process feature extraction module 130 is used to arrange the time series of the drilling pressure value, the time series of the drill bit speed value and the time series of the drill bit temperature value according to the parameter sample dimension and the time dimension into a multi-modal time series matrix of the drill bit travel process, and then input the matrix into the drill bit travel process feature extractor based on the hole convolution network to obtain a multi-modal time series correlation feature map of the drill bit travel process. It should be understood that considering that the time series of the drilling pressure value, the time series of the drill bit speed value and the time series of the drill bit temperature value have dynamic change information and mutual correlation in time series. Therefore, in the technical solution of the present application, the time series of the drilling pressure value, the time series of the drill bit speed value and the time series of the drill bit temperature value are arranged according to the parameter sample dimension and the time dimension into a multi-modal time series matrix of the drill bit travel process to form a more comprehensive multi-modal time series data representation. Next, considering that the multimodal timing matrix of the drill bit travel process has mutual timing correlation features, the dilated convolutional network introduces a dilated convolution kernel to better capture the timing correlation features between different parameters. Therefore, in the technical solution of the present application, the multimodal timing matrix of the drill bit travel process is input into the drill bit travel process feature extractor based on the dilated convolutional network to capture and mine the mutual dependence correlation of each parameter in the timing of the drill bit parameters, so as to obtain the multimodal timing correlation feature map of the drill bit travel process.

[0029] In an embodiment of the present application, the feature enhancement module 140 is used to input the multimodal temporal association feature map of the drill bit travel process into a feature enhancement module based on significance-globality to obtain an enhanced multimodal temporal association feature map of the drill bit travel process. Accordingly, it is considered that the multimodal temporal association feature map of the drill bit travel process contains significant feature information and overall global information about the drill bit travel process. Based on this, in order to be able to enhance and characterize the multimodal temporal association features of the drill bit travel process based on the global and significant information in the multimodal temporal association feature map of the drill bit travel process, in the technical solution of the present application, the multimodal temporal association feature map of the drill bit travel process is input into a feature enhancement module based on significance-globality to obtain an enhanced multimodal temporal association feature map of the drill bit travel process. That is to say, the saliency-global feature enhancement module can help the system identify and emphasize the key significant multimodal temporal correlation features of the drill bit travel process in the multimodal temporal correlation feature map of the drill bit travel process, so that the model pays more attention to the key significant information in the drill bit travel process and reduces interference, thereby improving the importance and influence of the key significant multimodal temporal correlation features, which can better capture the important features of the drill bit travel state, improve the discrimination and expression ability of the enhanced multimodal temporal correlation feature map of the drill bit travel process, and provide a better basis for the subsequent evaluation of the working state of the drill bit.

[0030] Specifically, in the embodiment of the present application, the feature enhancement module includes: a feature expansion unit, which is used to expand the multimodal temporal correlation feature graph of the drill bit travel process into a sequence of multimodal temporal correlation feature vectors of the drill bit travel process; a prominent feature calculation unit, which is used to extract the maximum value of each multimodal temporal correlation feature vector of the drill bit travel process in the sequence of multimodal temporal correlation feature vectors of the drill bit travel process to obtain a prominent multimodal temporal correlation feature vector of the drill bit travel process; and a global feature calculation unit, which is used to extract the average value of each multimodal temporal correlation feature vector of the drill bit travel process in the sequence of multimodal temporal correlation feature vectors of the drill bit travel process to obtain a global feature vector of the drill bit travel process. A multi-modal temporal correlation feature vector; a prominent feature one-dimensional convolution encoding unit, used to input the multi-modal temporal correlation convolution encoding vector of the protruding drill bit travel process obtained after one-dimensional convolution encoding of the multi-modal temporal correlation feature vector of the protruding drill bit travel process into the ReLU function to obtain a multi-modal temporal correlation activation encoding feature vector of the protruding drill bit travel process; a prominent feature activation unit, used to perform matrix multiplication of the prominent point convolution feature vector obtained after point convolution encoding of the multi-modal temporal correlation activation encoding feature vector of the protruding drill bit travel process with the first weight matrix to obtain a multi-modal temporal correlation activation feature vector of the protruding feature drill bit travel process; a global feature one-dimensional convolution encoding unit, used to The global drill bit travel process multimodal temporal association convolution coding vector obtained after one-dimensional convolution coding of the global drill bit travel process multimodal temporal association feature vector is input into the ReLU function to obtain the global drill bit travel process multimodal temporal association activation coding feature vector; a global feature activation unit is used to perform matrix multiplication of the global point convolution feature vector obtained after point convolution coding of the global drill bit travel process multimodal temporal association activation coding feature vector with the second weight matrix to obtain the global feature drill bit travel process multimodal temporal association activation feature vector; a feature fusion unit is used to fuse the prominent feature drill bit travel process multimodal temporal association activation feature vector and the global feature drill bit travel process multimodal temporal association activation feature vector. After the multimodal temporal association activation feature vector of the process is obtained to obtain the global-prominent feature multimodal temporal association feature vector of the drill bit travel process, the global-prominent feature multimodal temporal association feature vector of the drill bit travel process is nonlinearly activated to obtain the global-prominent feature multimodal temporal association weight feature vector of the drill bit travel process; a feature weighting unit is used to use the global-prominent feature multimodal temporal association weight feature vector of the drill bit travel process as a weight, perform position point multiplication on the sequence of the multimodal temporal association feature vectors of the drill bit travel process, and add the sequence of the multimodal temporal association feature vectors of the drill bit travel process to obtain the sequence of the multimodal temporal reinforcement association feature vectors of the drill bit travel process;and a feature aggregation unit, configured to aggregate the sequence of the multi-modal temporal enhanced associated feature vectors of the drill bit travel process in accordance with the manner of expanding the multi-modal temporal associated feature graph of the drill bit travel process to obtain the enhanced multi-modal temporal associated feature graph of the drill bit travel process. ;

[0031] More specifically, in an embodiment of the present application, the feature fusion unit is used to: perform feature fusion on the prominent feature drill bit travel process multimodal temporal association activation feature vector and the global feature drill bit travel process multimodal temporal association activation feature vector to obtain the global-prominent feature drill bit travel process multimodal temporal association feature vector; input the global-prominent feature drill bit travel process multimodal temporal association feature vector into a tanh function to obtain a first activated global-prominent feature drill bit travel process multimodal temporal association feature vector; input the global-prominent feature drill bit travel process multimodal temporal association feature vector into a sigmoid function to obtain a second activated global-prominent feature drill bit travel process multimodal temporal association feature vector; and, perform position point multiplication on the first activated global-prominent feature drill bit travel process multimodal temporal association feature vector and the second activated global-prominent feature drill bit travel process multimodal temporal association feature vector to obtain the global-prominent feature drill bit travel process multimodal temporal association weight feature vector.

[0032] In the embodiment of the present application, specifically, the feature enhancement module is used to: use the saliency-globality-based feature enhancement module to process the multimodal temporal correlation feature graph of the drill bit travel process with the following enhancement formula to obtain a sequence of enhanced multimodal temporal correlation feature vectors of the drill bit travel process; wherein the enhancement formula is: ;in, is a sequence of multi-modal temporal correlation feature vectors of the drill bit travel process obtained by expanding the multi-modal temporal correlation feature graph of the drill bit travel process, and are the maximum value and the average value of each eigenvector in the sequence of eigenvectors, and are respectively the multi-modal temporal correlation feature vector of the protruding drill bit travel process and the multi-modal temporal correlation feature vector of the global drill bit travel process, is a one-dimensional convolutional code, yes function, is point convolutional coding, and are the first weight matrix and the second weight matrix respectively, and are respectively the multi-modal temporal correlation activation feature vector of the prominent feature drill bit travel process and the multi-modal temporal correlation activation feature vector of the global feature drill bit travel process, is the function, yes function, It means point multiplication by position. is the multi-modal temporal association weight feature vector of the global-prominent feature drill bit travel process, is a sequence of multimodal temporal enhanced associated feature vectors of the drill bit travel process; and, the sequence of multimodal temporal enhanced associated feature vectors of the drill bit travel process is aggregated in a manner of expanding the multimodal temporal associated feature graph of the drill bit travel process to obtain the enhanced multimodal temporal associated feature graph of the drill bit travel process.

[0033] In the embodiment of the present application, the association focusing module 150 is used to input the multimodal temporal association feature map of the enhanced drill bit travel process into the feature multiscale focusing module to obtain the multimodal temporal association focusing feature vector of the enhanced drill bit travel process. It should be understood that, considering that the multimodal temporal association feature map of the enhanced drill bit travel process describes the overall background information and local temporal association feature information about the multimodal temporal association feature of the enhanced drill bit travel process. Therefore, in order to capture and focus on the multimodal temporal association information of different scales in the multimodal temporal association feature map of the enhanced drill bit travel process, so as to conduct a more in-depth analysis of the multimodal temporal association feature of the enhanced drill bit travel process, in the technical solution of the present application, the multimodal temporal association feature map of the enhanced drill bit travel process is input into the feature multiscale focusing module to obtain the multimodal temporal association focusing feature vector of the enhanced drill bit travel process. It should be understood that in the multimodal temporal association feature map of the drill bit travel process, the feature representation of different scales is crucial to capture the multimodal temporal association feature information of the drill bit travel process at different levels. The feature multi-scale focusing module can analyze and extract the multi-modal temporal correlation feature graph of the enhanced drill bit travel process from the local and global scales, so as to focus on the important features of the multi-modal temporal correlation features of the drill bit travel process at different scales, and comprehensively utilize and effectively integrate the features of these two different scales to make the model more comprehensively consider the multi-modal temporal correlation features of the drill bit travel process, thereby further improving the accuracy and efficiency of the working status judgment system.

[0034] Specifically, in the embodiment of the present application, the correlation focusing module includes: an upsampling unit, which is used to upsample the multimodal temporal correlation feature map of the enhanced drill bit travel process to obtain an upsampled multimodal temporal correlation feature map of the enhanced drill bit travel process; a global mean pooling processing unit, which is used to perform global mean pooling processing on each feature matrix along the channel dimension in the upsampled multimodal temporal correlation feature map of the enhanced drill bit travel process to obtain an upsampled multimodal temporal correlation channel feature vector of the enhanced drill bit travel process; a temporal correlation global processing unit, which is used to perform point convolution encoding on the multimodal temporal correlation channel feature vector of the upsampled multimodal temporal correlation channel of the enhanced drill bit travel process to obtain an upsampled multimodal temporal correlation channel feature vector of the enhanced drill bit travel process. a local feature vector; a channel modulation unit, used for performing point convolution encoding on the multi-modal temporal correlation feature map of the up-sampled enhanced drill bit travel process to obtain a channel-modulated up-sampled enhanced drill bit travel process multi-modal temporal correlation feature map; a temporal correlation local processing unit, used for performing two-dimensional convolution encoding on the channel-modulated up-sampled enhanced drill bit travel process multi-modal temporal correlation feature map to obtain a multi-modal temporal correlation local feature vector of the enhanced drill bit travel process; and a multi-scale correlation unit, used for fusing the multi-modal temporal correlation global feature vector of the enhanced drill bit travel process and the multi-modal temporal correlation local feature vector of the enhanced drill bit travel process to obtain a multi-modal temporal correlation focused feature vector of the enhanced drill bit travel process.

[0035] In an embodiment of the present application, the feature dynamic interaction module 160 is used to input the enhanced drill bit travel process multimodal temporal association focused feature vector and the drill bit travel state temporal association feature vector into a feature dynamic interaction module based on a prior distribution to obtain a drill bit travel state-drill bit process feature dynamic interaction representation vector as a drill bit travel state-drill bit process feature dynamic interaction representation feature. Accordingly, considering that the enhanced drill bit travel process multimodal temporal association focused feature vector and the drill bit travel state temporal association feature vector have a temporally correlated interactive relationship. Therefore, in order to better understand and analyze the dynamic interactive relationship between the two in temporal sequence, in the technical solution of the present application, the enhanced drill bit travel process multimodal temporal association focused feature vector and the drill bit travel state temporal association feature vector are input into a feature dynamic interaction module based on a prior distribution to obtain a drill bit travel state-drill bit process feature dynamic interaction representation vector. Specifically, the feature dynamic interaction module can promote the effective interaction between the multimodal temporal correlation focused feature vector of the drill bit travel process and the temporal correlation feature vector of the drill bit travel state, thereby improving the temporal correlation and information transmission efficiency between features, so that the model can better capture the complex relationship between the drill bit travel state and the drill bit process characteristics.

[0036] Specifically, in an embodiment of the present application, the feature dynamic interaction module includes: a feature corresponding calculation unit, which is used to calculate the position-based division of the multimodal temporal correlation focusing feature vector of the enhanced drill bit travel process and the temporal correlation feature vector of the drill bit travel state to obtain a drill bit travel process-travel state response interaction feature vector; a first feature weighting unit, which is used to calculate the product of the drill bit travel process-travel state response interaction feature vector and a first weight coefficient to obtain a first weight feature vector, and then use the feature value of each position in the first weight feature vector as the exponent of a natural constant to calculate an exponential function value with the natural constant as the base to obtain a first drill bit travel process-travel state response interaction exponential function feature vector; a first Gaussian calculation unit, which is used to calculate the product of the first drill bit travel process-travel state response interaction exponential function feature vector and a first Gaussian distribution function coefficient to obtain a first drill bit travel process-travel state response interaction feature vector; a first feature weighting unit, which is used to calculate the drill bit travel process -After multiplying the product of the drill bit travel process-travel state response interaction characteristic vector and the second weight coefficient to obtain the second weight characteristic vector, the characteristic value of each position in the second weight characteristic vector is used as the exponent of the natural constant to calculate the exponential function value with the natural constant as the base to obtain the second drill bit travel process-travel state response interaction exponential function characteristic vector; a second Gaussian calculation unit, used to calculate the product of the second drill bit travel process-travel state response interaction exponential function characteristic vector and the second Gaussian distribution function coefficient to obtain the second drill bit travel process-travel state response interaction characteristic vector; and a dynamic interaction unit, used to calculate the positional addition between the first drill bit travel process-travel state response interaction characteristic vector and the second drill bit travel process-travel state response interaction characteristic vector to obtain the drill bit travel state-drill bit process characteristic dynamic interaction representation vector; wherein, the first Gaussian distribution function coefficient and the second Gaussian distribution function coefficient are both generated by a Gaussian distribution random number function with a mean of 0 and a variance of 1.

[0037] In the embodiment of the present application, specifically, the feature dynamic interaction module is used to: use the feature dynamic interaction module based on prior distribution to process the multimodal temporal association focused feature vector of the enhanced drill bit travel process and the temporal association feature vector of the drill bit travel state with the following dynamic interaction formula to obtain the drill bit travel state-drill bit process feature dynamic interaction representation vector; wherein the dynamic interaction formula is: ;in, and are respectively the multi-modal time series correlation focused feature vector of the enhanced drill bit travel process and the time series correlation feature vector of the drill bit travel state, and are the first weight coefficient and the second weight coefficient, and The Gaussian distribution function coefficient is a Gaussian distribution random number function with a mean of 0 and a variance of 1. It is a vector representing the dynamic interaction of the drill bit's travel state and the drill bit's process characteristics.

[0038] In an embodiment of the present application, the monitoring result generation module 170 is used to obtain a monitoring result based on the dynamic interaction representation feature of the drill bit travel state-drill bit process characteristics, and the monitoring result is used to indicate whether the working state of the drill bit is normal. Specifically, in an embodiment of the present application, the monitoring result generation module is used to: input the drill bit travel state-drill bit process characteristics dynamic interaction representation vector into a real-time monitoring result generator based on a classifier to obtain the monitoring result. That is, the drill bit travel state-drill bit process characteristics dynamic interaction representation obtained by dynamic interaction of the enhanced drill bit travel process multimodal temporal association focusing feature vector and the drill bit travel state temporal association feature vector is used for classification processing to automatically determine whether the working state of the drill bit is normal. In this way, the drilling process can be fully monitored so as to detect abnormal situations in time to ensure the safety of the operators.

[0039] Here, the time-series correlation feature vector of the drill bit travel state expresses the time-series correlation features of the drill bit position data, while the enhanced multimodal time-series correlation focused feature vector of the drill bit travel process expresses the time-series-sample cross-dimensional local correlation features of the drilling pressure value, the drill bit speed value and the drill bit temperature value, and the full time domain-full sample domain significance enhancement and multi-scale focusing features. Therefore, when the enhanced multimodal time-series correlation focused feature vector of the drill bit travel process and the time-series correlation feature vector of the drill bit travel state are input into the feature dynamic interaction module based on the prior distribution, the obtained drill bit travel state-drill bit process feature dynamic interaction representation vector will also have local feature distribution discretization due to the feature distribution difference between the enhanced multimodal time-series correlation focused feature vector of the drill bit travel process and the time-series correlation feature vector of the drill bit travel state.

[0040] Based on this, in a preferred embodiment, the drill bit travel state-drill bit process feature dynamic interaction representation vector is input into a real-time monitoring result generator based on a classifier to obtain a monitoring result, including: calculating the sum of the absolute values ​​of each eigenvalue of the drill bit travel state-drill bit process feature dynamic interaction representation vector to obtain a first drill bit travel state-drill bit process feature dynamic interaction and modulation value, and calculating the square root of the sum of the squares of each eigenvalue of the drill bit travel state-drill bit process feature dynamic interaction representation vector to obtain a second drill bit travel state-drill bit process feature dynamic interaction and modulation value; performing dot-wise subtraction of the drill bit travel state-drill bit process feature dynamic interaction representation vector and the second drill bit travel state-drill bit process feature dynamic interaction and modulation value, and performing dot-wise multiplication with the number of eigenvalues ​​of the drill bit travel state-drill bit process feature dynamic interaction representation vector and the reciprocal of the first drill bit travel state-drill bit process feature dynamic interaction and modulation value, and taking the reciprocal of each eigenvalue to obtain the first drill bit travel state-drill bit process feature dynamic interaction and modulation value. state-drill bit process characteristic dynamic interaction phase conversion vector; after dot-wise subtraction of the drill bit travel state-drill bit process characteristic dynamic interaction representation vector and the first drill bit travel state-drill bit process characteristic dynamic interaction and modulation value, respectively multiply them with the square root of the number of eigenvalues ​​of the drill bit travel state-drill bit process characteristic dynamic interaction representation vector and the reciprocal of the second drill bit travel state-drill bit process characteristic dynamic interaction and modulation value, and take the reciprocal of each eigenvalue to obtain the second drill bit travel state-drill bit process characteristic dynamic interaction phase conversion vector; subtract the dot product vector of the second drill bit travel state-drill bit process characteristic dynamic interaction phase conversion vector from the first drill bit travel state-drill bit process characteristic dynamic interaction phase conversion vector to obtain the optimized drill bit travel state-drill bit process characteristic dynamic interaction representation vector; the optimized drill bit travel state-drill bit process characteristic dynamic interaction representation vector is passed through a classifier-based real-time monitoring result generator to obtain a monitoring result.

[0041] Here, the drill bit travel state-drill bit process feature dynamic interaction representation vector is recorded as The optimization expression is: ;in, is a vector representing the dynamic interaction between the drill bit travel state and the drill bit process characteristics, represents the set of real numbers, The first vector representing the dynamic interaction between the drill bit travel state and the drill bit process characteristics is represented by The eigenvalues ​​at the positions, The number of eigenvalues ​​representing the dynamic interaction between the drill bit moving state and the drill bit process characteristics, Indicates the first drill bit travel state - the dynamic interaction and modulation value of the drill bit process characteristics, Indicates the second drill bit travel state - the dynamic interaction and modulation value of the drill bit process characteristics, It means point multiplication by position. It means subtracting by position. It means to calculate the inverse of each eigenvalue of the eigenvector. Represents the first drill bit travel state-drill bit process characteristic dynamic interaction phase conversion vector, Represents the second drill bit travel state-drill bit process characteristic dynamic interaction phase conversion vector, represents the weighted hyperparameter, A vector representing the dynamic interaction of the optimized drill bit travel state and the drill bit process characteristics.

[0042] Accordingly, in this preferred embodiment, the difference in the difference and modulation representation of the eigenvalue of the drill bit travel state-drill bit process feature dynamic interaction representation vector relative to the feature set of the vector as a whole of the drill bit travel state-drill bit process feature dynamic interaction representation vector is used as the semantic change intensity information, and a phase-like conversion corresponding to the position-based intensity modulation is performed through different and modulation representation forms, so that the aggregation enhancement of the semantic change phase perception can improve the axial aggregation receptive field along the feature aggregation direction by performing a spatial translation operation based on alternating stacking under the scale balance of the vector set of the drill bit travel state-drill bit process feature dynamic interaction representation vector, so as to improve the classification iteration effect of the real-time monitoring result generator based on the classifier of the feature set input of the drill bit travel state-drill bit process feature dynamic interaction representation vector, that is, to improve the speed of classification training and the accuracy of monitoring results.

[0043] It is worth mentioning that those skilled in the art should be aware that before applying a deep neural network model for inference, the deep neural network model must be trained so that the deep neural network can implement specific functional capabilities.

[0044] Specifically, in the technical solution of the present application, the real-time monitoring system for core drilling production also includes a training module for training the drill bit travel state feature extractor based on the GRU unit, the drill bit travel process feature extractor based on the void convolutional network, the saliency-globality based feature enhancement module, the feature multi-scale focusing module, the feature dynamic interaction module based on the prior distribution and the real-time monitoring result generator based on the classifier.

[0045] Specifically, in an embodiment of the present application, the training module includes: a training data acquisition unit, used to acquire training data, the training data includes a time series of training drill bit position data, a time series of training drill pressure values, a time series of training drill bit speed values, and a time series of training drill bit temperature values, as well as a real monitoring result, the real monitoring result is a real value of whether the working state of the drill bit is normal; a training drill bit travel state time series association unit, used to input the time series of the training drill bit position data into the drill bit travel state feature extractor based on the GRU unit to obtain a training drill bit travel state time series association feature vector; a multimodal drill bit travel process A time series association unit is used to arrange the time series of the training drilling pressure value, the time series of the training drill bit speed value and the time series of the training drill bit temperature value according to the parameter sample dimension and the time dimension into a multi-modal time series matrix of the training drill bit travel process, and then input the matrix into the drill bit travel process feature extractor based on the hole convolution network to obtain a multi-modal time series association feature map of the training drill bit travel process; a training feature enhancement unit is used to input the multi-modal time series association feature map of the training drill bit travel process into the feature enhancement module based on significance-globality to obtain a training enhanced multi-modal time series association feature map of the drill bit travel process; the training multi-modal time series association unit is used to The multi-modal temporal correlation feature graph of the training enhanced drill bit travel process is input into the feature multi-scale focusing module to obtain the multi-modal temporal correlation focusing feature vector of the training enhanced drill bit travel process; the training feature dynamic interaction unit is used to input the multi-modal temporal correlation focusing feature vector of the training enhanced drill bit travel process and the training drill bit travel state temporal correlation feature vector into the feature dynamic interaction module based on prior distribution to obtain the training drill bit travel state-drill bit process feature dynamic interaction representation vector; the training monitoring result generation unit is used to input the training drill bit travel state-drill bit process feature dynamic interaction representation vector into the real-time classifier based on the classifier A monitoring result generator is provided to obtain a training monitoring result; a classification loss unit is used to calculate the cross entropy loss function value between the training monitoring result and the real monitoring result to obtain a classification loss function value; and a model training unit is used to train the drill travel state feature extractor based on the GRU unit, the drill travel process feature extractor based on the void convolution network, the feature enhancement module based on significance-globality, the feature multi-scale focusing module, the feature dynamic interaction module based on the prior distribution and the real-time monitoring result generator based on the classifier based on the classification loss function value and through back propagation of gradient descent.

[0046] In this preferred embodiment, when the drill travel state feature extractor based on the GRU unit, the drill travel process feature extractor based on the void convolution network, the saliency-globality-based feature enhancement module, the feature multi-scale focusing module, the feature dynamic interaction module based on the prior distribution and the classifier-based real-time monitoring result generator are trained based on the classification loss function value, wherein, in each iteration of the training, the training drill travel state-drill process feature dynamic interaction representation vector is iteratively optimized.

[0047] Wherein, the iterative optimization process includes: firstly, clustering the training drill bit travel state-drill bit process feature dynamic interaction representation vector, for example, clustering based on the distance between eigenvalues, and, considering the inconsistent convergence logic of probability density distribution of intra-class features and out-of-class features that may be caused by the clustering operation, further optimizing the clustered training drill bit travel state-drill bit process feature dynamic interaction representation vector, including: determining the number of intra-class features of the training drill bit travel state-drill bit process feature dynamic interaction representation vector; dividing the total number of eigenvalues ​​of the training drill bit travel state-drill bit process feature dynamic interaction representation vector by the number of intra-class features to obtain a class importance value, and calculating the inverse of the class importance value to obtain a class constraint value; calculating the difference between the total number of eigenvalues ​​and the number of intra-class features, and dividing the total number of eigenvalues ​​by the difference to obtain an out-of-class proportion value; Calculate each eigenvalue of the training drill bit travel state-drill bit process feature dynamic interaction representation vector as a power function with the class importance value as the exponent, add the exponential value with the natural function as the base and the class constraint value as the exponent, and then multiply by the out-of-class proportion value to obtain the training drill bit travel state-drill bit process feature dynamic interaction modulation vector; add the weight parameter and the class importance value to obtain the class importance weight value; perform dot multiplication on the training drill bit travel state-drill bit process feature dynamic interaction representation vector and the class importance weight value to obtain the training drill bit travel state-drill bit process feature dynamic interaction ontology vector; and calculate the weighted sum of the training drill bit travel state-drill bit process feature dynamic interaction modulation vector and the training drill bit travel state-drill bit process feature dynamic interaction ontology vector to obtain the optimized training drill bit travel state-drill bit process feature dynamic interaction representation vector.

[0048] Specifically, in this preferred embodiment, the clustered training drill bit travel state-drill bit process feature dynamic interaction representation vector is optimized to obtain an optimized training drill bit travel state-drill bit process feature dynamic interaction representation vector, and the process formula is expressed as follows: ;in, is the training drill bit travel state-drill bit process feature dynamic interaction representation vector, is the total number of eigenvalues ​​of the training drill bit travel state-drill bit process feature dynamic interaction representation vector, is the number of intra-class features of the training drill bit travel state-drill bit process feature dynamic interaction representation vector, It is the point multiplication by position. is the weight parameter, is added by position, It means to perform exponential operation on the eigenvalues ​​at each position in the eigenvector. It is the optimized training drill bit travel state-drill bit process feature dynamic interaction representation vector.

[0049] That is, while clustering the training drill bit travel state-drill bit process feature dynamic interaction representation vector, the cluster importance measure of the feature value of the training drill bit travel state-drill bit process feature dynamic interaction representation vector is combined with the class constraint measure of the clustering operation to further perform responsive modulation with the out-of-class scale factor, and based on the clustering scaling of the feature ontology representation of the training drill bit travel state-drill bit process feature dynamic interaction representation vector, the clustering operation is used to establish the simplicity and effectiveness of the overall feature purification of the feature set of the training drill bit travel state-drill bit process feature dynamic interaction representation vector, and the probability density distribution convergence logic inconsistency caused by the clustering operation is suppressed, so as to improve the classification iteration effect of the real-time monitoring result generator based on the classifier of the feature set input of the training drill bit travel state-drill bit process feature dynamic interaction representation vector, that is, to improve the speed of classification training and the accuracy of monitoring results. In this way, the drilling process can be fully monitored so as to detect abnormal situations in time to ensure the safety of operators.

[0050] In summary, the real-time monitoring system 100 for core drilling production based on the embodiment of the present application is explained, which collects the time series of drill bit position data, the time series of drilling pressure value, the time series of drill bit speed value and the time series of drill bit temperature value, and uses AI-based data processing and analysis technology to perform time series analysis and association on the drill bit position data, the drilling pressure value, the drill bit speed value and the drill bit temperature value, so as to automatically judge whether the working state of the drill bit is normal based on the time series dynamic interaction characteristics of the drill bit travel state and the drill bit process. In this way, the drilling process can be fully monitored so as to detect abnormal situations in time to ensure the safety of the operators.

[0051] As described above, the real-time monitoring system 100 for core drilling production according to the embodiment of the present application can be implemented in various wireless terminals, such as a server with a real-time monitoring algorithm for core drilling production. In a possible implementation, the real-time monitoring system 100 for core drilling production according to the embodiment of the present application can be integrated into the wireless terminal as a software module and / or a hardware module. For example, the real-time monitoring system 100 for core drilling production can be a software module in the operating system of the wireless terminal, or can be an application developed for the wireless terminal; of course, the real-time monitoring system 100 for core drilling production can also be one of the many hardware modules of the wireless terminal.

[0052] Alternatively, in another example, the real-time monitoring system 100 for core drilling production and the wireless terminal may also be separate devices, and the real-time monitoring system 100 for core drilling production may be connected to the wireless terminal via a wired and / or wireless network and transmit interactive information in accordance with an agreed data format.

[0053] Figure 2 Flow chart of a real-time monitoring method for core drilling production according to an embodiment of the present application. Figure 2As shown, according to the embodiment of the present application, the real-time monitoring method for core drilling production includes: S110, acquiring the time series of drill bit position data, the time series of drilling pressure values, the time series of drill bit speed values, and the time series of drill bit temperature values ​​collected by the sensor group; S120, extracting the drill bit travel state feature of the time series of the drill bit position data to obtain the drill bit travel state time series associated feature vector; S130, arranging the time series of drilling pressure values, the time series of drill bit speed values, and the time series of drill bit temperature values ​​into a drill bit travel process multimodal time series matrix according to the parameter sample dimension and the time dimension, and then inputting the matrix into the drill bit travel process feature extractor based on the void convolution network to obtain the drill bit travel process multimodal time series associated feature map; S140, inputting the drill bit travel process multimodal time series associated feature map into the base In the saliency-global feature enhancement module to obtain the enhanced multimodal temporal correlation feature map of the drill bit travel process; S150, the enhanced multimodal temporal correlation feature map of the drill bit travel process is input into the feature multi-scale focusing module to obtain the enhanced multimodal temporal correlation focusing feature vector of the drill bit travel process; S160, the enhanced multimodal temporal correlation focusing feature vector of the drill bit travel process and the drill bit travel state temporal correlation feature vector are input into the feature dynamic interaction module based on the prior distribution to obtain the drill bit travel state-drill bit process feature dynamic interaction representation vector as the drill bit travel state-drill bit process feature dynamic interaction representation feature; and, S170, based on the drill bit travel state-drill bit process feature dynamic interaction representation feature, a monitoring result is obtained, and the monitoring result is used to indicate whether the working state of the drill bit is normal.

[0054] Here, those skilled in the art will appreciate that the specific operations of each step in the above-mentioned real-time monitoring method for core drilling production have been described in detail above. Figure 1 The present invention has been described in detail in the description of the real-time monitoring system for core drilling production, and therefore, its repeated description will be omitted.

[0055] The above descriptions of various implementations of the present disclosure are exemplary and not exhaustive. The disclosure is not limited to the disclosed implementations, and many modifications and variations are obvious to those of ordinary skill in the art without departing from the scope and spirit of the described implementations. The terms used herein are selected to best explain the principles of the implementations, practical applications, or improvements to the technology in the market, or to enable other persons of ordinary skill in the art to understand the various practical methods disclosed herein.

Claims

1. A real-time monitoring system for core drilling production, characterized in that: include: A drill bit parameter acquisition module, used to obtain the time series of drill bit position data, the time series of drilling pressure value, the time series of drill bit speed value and the time series of drill bit temperature value collected by the sensor group; a drill bit travel state feature extraction module, used to extract the drill bit travel state feature of the time series of the drill bit position data to obtain a drill bit travel state time series correlation feature vector; A drill bit travel process feature extraction module, used for arranging the time series of the drilling pressure value, the time series of the drill bit speed value and the time series of the drill bit temperature value into a drill bit travel process multimodal time series matrix according to the parameter sample dimension and the time dimension, and then inputting the matrix into a drill bit travel process feature extractor based on a hole convolutional network to obtain a drill bit travel process multimodal time series correlation feature map; A feature enhancement module, used for inputting the multi-modal temporal correlation feature graph of the drill bit travel process into a feature enhancement module based on saliency-globality to obtain an enhanced multi-modal temporal correlation feature graph of the drill bit travel process; A correlation focusing module, used for inputting the multi-modal temporal correlation feature map of the enhanced drill bit travel process into a feature multi-scale focusing module to obtain a multi-modal temporal correlation focusing feature vector of the enhanced drill bit travel process; A feature dynamic interaction module, used for inputting the multi-modal time series correlation focused feature vector of the enhanced drill bit travel process and the time series correlation feature vector of the drill bit travel state into a feature dynamic interaction module based on a priori distribution to obtain a drill bit travel state-drill bit process feature dynamic interaction representation vector as a drill bit travel state-drill bit process feature dynamic interaction representation feature; A monitoring result generating module, used for obtaining monitoring results based on the dynamic interactive representation characteristics of the drill bit travel state and the drill bit process characteristics, wherein the monitoring results are used to indicate whether the working state of the drill bit is normal; The feature enhancement module includes: a feature expansion unit, which is used to expand the multimodal temporal correlation feature graph of the drill bit travel process into a sequence of multimodal temporal correlation feature vectors of the drill bit travel process; a prominent feature calculation unit, which is used to extract the maximum value of each multimodal temporal correlation feature vector of the drill bit travel process in the sequence of multimodal temporal correlation feature vectors of the drill bit travel process to obtain a prominent multimodal temporal correlation feature vector of the drill bit travel process; and a global feature calculation unit, which is used to extract the average value of each multimodal temporal correlation feature vector of the drill bit travel process in the sequence of multimodal temporal correlation feature vectors of the drill bit travel process to obtain a global multimodal temporal correlation feature vector of the drill bit travel process. a prominent feature one-dimensional convolution encoding unit, used to input the prominent drill bit travel process multimodal temporal association convolution encoding vector obtained after one-dimensional convolution encoding of the multimodal temporal association feature vector of the prominent drill bit travel process into the ReLU function to obtain a prominent drill bit travel process multimodal temporal association activation encoding feature vector; a prominent feature activation unit, used to perform matrix multiplication of the prominent point convolution feature vector obtained after point convolution encoding of the multimodal temporal association activation encoding feature vector of the prominent drill bit travel process with the first weight matrix to obtain a prominent feature drill bit travel process multimodal temporal association activation feature vector; a global feature one-dimensional convolution encoding unit, used to convert the global drill bit travel process into a global multimodal temporal association activation feature vector; The global drill bit travel process multimodal temporal association convolution coding vector obtained after one-dimensional convolution coding of the multimodal temporal association feature vector of the drill bit travel process is input into the ReLU function to obtain the global drill bit travel process multimodal temporal association activation coding feature vector; a global feature activation unit is used to perform matrix multiplication of the global point convolution feature vector obtained after point convolution coding of the global drill bit travel process multimodal temporal association activation coding feature vector with the second weight matrix to obtain the global feature drill bit travel process multimodal temporal association activation feature vector; a feature fusion unit is used to fuse the prominent feature drill bit travel process multimodal temporal association activation feature vector and the global feature drill bit travel process multimodal temporal association activation feature vector. After the modal time series association activation feature vector is obtained to obtain the global-prominent feature drill bit travel process multimodal time series association feature vector, the global-prominent feature drill bit travel process multimodal time series association feature vector is nonlinearly activated to obtain the global-prominent feature drill bit travel process multimodal time series association weight feature vector; a feature weighting unit is used to use the global-prominent feature drill bit travel process multimodal time series association weight feature vector as a weight, perform position point multiplication on the sequence of the drill bit travel process multimodal time series association feature vectors, and add the sequence of the drill bit travel process multimodal time series association feature vectors to obtain the sequence of the drill bit travel process multimodal time series reinforcement association feature vectors;A feature aggregation unit, configured to aggregate the sequence of the multi-modal temporal enhanced correlation feature vectors of the drill bit travel process in accordance with the manner of expanding the multi-modal temporal associated feature graph of the drill bit travel process to obtain the enhanced multi-modal temporal associated feature graph of the drill bit travel process; The correlation focusing module includes: an upsampling unit, which is used to upsample the multimodal temporal correlation feature map of the enhanced drill bit travel process to obtain an upsampled multimodal temporal correlation feature map of the enhanced drill bit travel process; a global mean pooling processing unit, which is used to perform global mean pooling processing on each feature matrix along the channel dimension in the upsampled multimodal temporal correlation feature map of the enhanced drill bit travel process to obtain a multimodal temporal correlation channel feature vector of the upsampled multimodal temporal correlation process; a temporal correlation global processing unit, which is used to perform point convolution encoding on the multimodal temporal correlation channel feature vector of the upsampled multimodal temporal correlation process to obtain a multimodal temporal correlation global feature map of the enhanced drill bit travel process; vector; a channel modulation unit, used for performing point convolution encoding on the multi-modal temporal correlation feature map of the up-sampled enhanced drill bit travel process to obtain a channel-modulated up-sampled enhanced drill bit travel process multi-modal temporal correlation feature map; a temporal correlation local processing unit, used for performing two-dimensional convolution encoding on the channel-modulated up-sampled enhanced drill bit travel process multi-modal temporal correlation feature map to obtain a multi-modal temporal correlation local feature vector of the enhanced drill bit travel process; a multi-scale correlation unit, used for fusing the multi-modal temporal correlation global feature vector of the enhanced drill bit travel process and the multi-modal temporal correlation local feature vector of the enhanced drill bit travel process to obtain a multi-modal temporal correlation focused feature vector of the enhanced drill bit travel process; Among them, the feature dynamic interaction module includes: a feature corresponding calculation unit, which is used to calculate the position-based division of the multi-modal temporal correlation focusing feature vector of the enhanced drill bit travel process and the temporal correlation feature vector of the drill bit travel state to obtain the drill bit travel process-travel state response interaction feature vector; a first feature weighting unit, which is used to calculate the product of the drill bit travel process-travel state response interaction feature vector and the first weight coefficient to obtain the first weight feature vector, and then use the feature value of each position in the first weight feature vector as the exponent of the natural constant to calculate the exponential function value with the natural constant as the base to obtain the first drill bit travel process-travel state response interaction exponential function feature vector; a first Gaussian calculation unit, which is used to calculate the product of the first drill bit travel process-travel state response interaction exponential function feature vector and the first Gaussian distribution function coefficient to obtain the first drill bit travel process-travel state response interaction feature vector; the first feature weighting unit is used to calculate the drill bit travel process-travel state response interaction exponential function feature vector. After the state response interaction feature vector is multiplied by the second weight coefficient to obtain the second weight feature vector, the feature value of each position in the second weight feature vector is used as the exponent of the natural constant to calculate the exponential function value with the natural constant as the base to obtain the second drill bit travel process-travel state response interaction exponential function feature vector; a second Gaussian calculation unit is used to calculate the product of the second drill bit travel process-travel state response interaction exponential function feature vector and the second Gaussian distribution function coefficient to obtain the second drill bit travel process-travel state response interaction feature vector; a dynamic interaction unit is used to calculate the positional addition between the first drill bit travel process-travel state response interaction feature vector and the second drill bit travel process-travel state response interaction feature vector to obtain the drill bit travel state-drill bit process feature dynamic interaction representation vector; wherein the first Gaussian distribution function coefficient and the second Gaussian distribution function coefficient are both generated by a Gaussian distribution random number function with a mean of 0 and a variance of 1.

2. The real-time monitoring system for core drilling production according to claim 1, characterized in that: The drill bit travel state feature extraction module is used to: input the time series of the drill bit position data into the drill bit travel state feature extractor based on the GRU unit to obtain the drill bit travel state time series correlation feature vector.

3. The real-time monitoring system for core drilling production according to claim 2, characterized in that: The feature fusion unit is used to: perform feature fusion on the prominent feature multimodal temporal association activation feature vector of the drill bit's travel process and the global feature multimodal temporal association activation feature vector of the drill bit's travel process to obtain the global-prominent feature multimodal temporal association feature vector of the drill bit's travel process; input the global-prominent feature multimodal temporal association feature vector of the drill bit's travel process into a tanh function to obtain a first activated global-prominent feature multimodal temporal association feature vector of the drill bit's travel process; input the global-prominent feature multimodal temporal association feature vector of the drill bit's travel process into a sigmoid function to obtain a second activated global-prominent feature multimodal temporal association feature vector of the drill bit's travel process; and perform position point multiplication on the first activated global-prominent feature multimodal temporal association feature vector of the drill bit's travel process and the second activated global-prominent feature multimodal temporal association feature vector of the drill bit's travel process to obtain the global-prominent feature multimodal temporal association weight feature vector of the drill bit's travel process.

4. The real-time monitoring system for core drilling production according to claim 3, characterized in that: The monitoring result generating module is used to input the drill bit travel state-drill bit process feature dynamic interaction representation vector into a classifier-based real-time monitoring result generator to obtain the monitoring result.

5. The real-time monitoring system for core drilling production according to claim 4, characterized in that: It also includes a training module for training the drill travel state feature extractor based on the GRU unit, the drill travel process feature extractor based on the void convolution network, the saliency-globality based feature enhancement module, the feature multi-scale focusing module, the feature dynamic interaction module based on the prior distribution and the real-time monitoring result generator based on the classifier.

6. The real-time monitoring system for core drilling production according to claim 5, characterized in that: The training module includes: a training data acquisition unit, used to acquire training data, wherein the training data includes a time series of training drill bit position data, a time series of training drilling pressure values, a time series of training drill bit rotation speed values, and a time series of training drill bit temperature values, as well as a real monitoring result, wherein the real monitoring result is a real value of whether the working state of the drill bit is normal; a training drill bit travel state time series association unit, used to input the time series of the training drill bit position data into the drill bit travel state feature extractor based on the GRU unit to obtain a training drill bit travel state time series association feature vector; a drill bit travel process multimodal time series association unit, used to The time series of the training drilling pressure value, the time series of the training drill bit speed value and the time series of the training drill bit temperature value are arranged into a multi-modal time series matrix of the training drill bit travel process according to the parameter sample dimension and the time dimension, and then input into the drill bit travel process feature extractor based on the void convolution network to obtain a multi-modal time series correlation feature map of the training drill bit travel process; a training feature enhancement unit is used to input the multi-modal time series correlation feature map of the training drill bit travel process into the feature enhancement module based on significance-globality to obtain a training enhanced multi-modal time series correlation feature map of the drill bit travel process; a training multi-modal time series association unit is used to input the training drill bit travel process multi-modal time series correlation feature map into the feature enhancement module based on significance-globality to obtain a training enhanced drill bit travel process multi-modal time series correlation feature map; The multi-modal temporal correlation feature map of the enhanced drill bit travel process is input into the feature multi-scale focusing module to obtain the multi-modal temporal correlation focusing feature vector of the training enhanced drill bit travel process; the training feature dynamic interaction unit is used to input the multi-modal temporal correlation focusing feature vector of the training enhanced drill bit travel process and the training drill bit travel state temporal correlation feature vector into the feature dynamic interaction module based on prior distribution to obtain the training drill bit travel state-drill bit process feature dynamic interaction representation vector; the training monitoring result generation unit is used to input the training drill bit travel state-drill bit process feature dynamic interaction representation vector into the real-time monitoring based on classifier A control result generator is used to obtain a training monitoring result; a classification loss unit is used to calculate the cross entropy loss function value between the training monitoring result and the real monitoring result to obtain a classification loss function value; a model training unit is used to train the drill travel state feature extractor based on the GRU unit, the drill travel process feature extractor based on the void convolution network, the feature enhancement module based on significance-globality, the feature multi-scale focusing module, the feature dynamic interaction module based on prior distribution and the real-time monitoring result generator based on the classifier based on the classification loss function value and through back propagation of gradient descent.

7. A method for using the real-time monitoring system for core drilling production according to claim 1, characterized in that: include: The time series of drill bit position data, the time series of drilling pressure values, the time series of drill bit speed values ​​and the time series of drill bit temperature values ​​collected by the sensor group are obtained; the time series of drill bit position data is subjected to drill bit travel state feature extraction to obtain a drill bit travel state time series associated feature vector; the time series of drilling pressure values, the time series of drill bit speed values ​​and the time series of drill bit temperature values ​​are arranged into a multimodal time series matrix of the drill bit travel process according to the parameter sample dimension and the time dimension, and then input into a drill bit travel process feature extractor based on a hole convolution network to obtain a multimodal time series associated feature map of the drill bit travel process; the multimodal time series associated feature map of the drill bit travel process is input into a feature enhancement module based on significance-globality to obtain an enhanced multimodal time series associated feature map of the drill bit travel process; the enhanced multimodal time series associated feature map of the drill bit travel process is input into a feature multi-scale focusing module to obtain an enhanced multimodal time series associated focused feature vector of the drill bit travel process; Inputting the enhanced drill bit travel process multimodal time series correlation focused feature vector and the drill bit travel state time series correlation feature vector into a feature dynamic interaction module based on prior distribution to obtain a drill bit travel state-drill bit process feature dynamic interaction representation vector as a drill bit travel state-drill bit process feature dynamic interaction representation feature; Based on the dynamic interactive representation feature of the drill bit travel status-drill bit process feature, a monitoring result is obtained, and the monitoring result is used to indicate whether the working state of the drill bit is normal.

Images