Intelligent identification method for unfavorable geologic body while drilling based on multi-modal fusion

Abstract

The invention provides an intelligent identification method for unfavorable geologic bodies while drilling based on multi-modal fusion. The method comprises the steps of data collection, data preprocessing, feature extraction, modal common and private feature extraction, modal fusion and unfavorable geologic body identification. According to the method, the multi-modal while-drilling data set is formed by collecting the four modal data of the drilling machine operation parameters, the while-drilling vibration parameters, the while-drilling acoustic signals and the while-drilling resistivity imaging graph, the complementarity of different data sources can be fully utilized by the multi-modal data, and more comprehensive analysis is provided for recognition of the unfavorable geologic body; according to the technology, the accuracy and reliability of identification can be improved, the anti-noise capability can be enhanced through cross validation of different data modes, and environmental interference signals are filtered out; besides, multi-modal fusion can adapt to more complex geological environments, and particularly under the condition that multiple geological anomalies coexist, potential unfavorable geologic bodies can be better identified, real-time and accurate feedback can be provided, and technical support is provided for construction under complex geological conditions.

Description

Technical Field

[0001] The present invention relates to the technical field of bad geological body recognition, and in particular to a method for intelligent recognition of bad geological bodies while drilling based on multimodal fusion. Background Art

[0002] With the rapid development of modern infrastructure construction, underground engineering has become increasingly important in the fields of transportation, energy development, and urban underground space utilization. However, due to the complexity of underground geological conditions, various unfavorable geological bodies are often encountered during the construction process, such as faults, caves, weak interlayers, karst, expansive soils, and groundwater-rich areas. These unfavorable geological bodies not only increase the difficulty of construction, but may also lead to a series of serious engineering problems such as collapse, water gushing, and deformation, endangering construction safety and engineering quality. Therefore, how to identify these unfavorable geological bodies in real time and accurately during the construction process, especially during the drilling process, has become an important technical problem that needs to be solved in the field of underground engineering.

[0003] At present, in the process of identifying poor geological bodies, traditional geophysical methods such as seismic wave detection and geological radar are widely used, but they have obvious limitations. First, the detection depth of geophysical methods is limited. Especially in water-bearing or high-density rock areas, the signal is easily attenuated, making it difficult to accurately identify deep geological bodies. Secondly, the spatial resolution of these methods is insufficient, and it is difficult to accurately locate poor geological bodies under complex geological conditions. In addition, geophysical data are usually obtained before construction and cannot reflect geological changes during construction in real time, resulting in delayed decision-making.

[0004] The identification while drilling technology has strong advantages in identifying bad geological bodies by collecting data such as drilling rig operating parameters and drilling vibration parameters in real time. It has strong real-time performance, can synchronously obtain underground deep geological information, and adjust the construction plan in time according to real-time feedback. However, the inventors of this application have found through research that the identification while drilling technology that relies on single modal data, such as only using drilling rig parameters or acoustic signals, still has the following problems: the information of single modal data is limited, and it is easily affected by noise and environmental interference, resulting in inaccurate identification results. Especially under complex geological conditions, it has poor adaptability and is prone to misjudgment. Summary of the invention

[0005] In view of the technical problems that the existing while-drilling identification technology relies on single-modal data, the information of the single-modal data is limited, and it is easily affected by noise and environmental interference, resulting in inaccurate identification results, especially under complex geological conditions, the adaptability is poor, and misjudgment is prone to occur. The present invention provides a while-drilling adverse geological body intelligent identification method based on multi-modal fusion.

[0006] In order to solve the above technical problems, the present invention adopts the following technical solutions:

[0007] A method for intelligently identifying unfavorable geological bodies while drilling based on multimodal fusion, comprising the following steps:

[0008] Data collection: Collect multi-modal while-drilling data and corresponding adverse geological body information, wherein the multi-modal while-drilling data includes drilling rig operation parameters, while-drilling vibration parameters, while-drilling acoustic signals and while-drilling resistivity imaging images;

[0009] Data preprocessing: Four modal data, namely, drilling rig operating parameters, drilling vibration parameters, drilling acoustic signals and drilling resistivity imaging, are selected to form a multimodal drilling dataset. The adverse geological body information corresponding to the drilling data is selected to form a dataset label. The dataset label is used to guide the model to learn how to identify different adverse geological bodies based on the input multimodal drilling data. Through the correspondence between the model prediction results and the actual geological conditions, the model can learn by comparing the predictions and actual labels, and continuously optimize and improve the model.

[0010] Feature extraction: The input of the multi-modal LWD dataset is four modes, namely, the drilling rig operation parameter mode r, the LWD vibration parameter mode v, the LWD acoustic signal mode a, and the LWD resistivity imaging mode i. The corresponding original modal data are m∈{r,v,a,i}; where, T m Input sample number for different modalities, d m are the feature dimensions of different modalities, and m is the modal category;

[0011] For the three types of modes, namely, drilling rig operating parameters, drilling vibration parameters, and drilling acoustic signals, we use stacked bidirectional long short-term memory (BiLSTM) to map different original modal data into a unified dimension to obtain features. m∈{r,v,a}:x m =BiLSTM(o m ),m∈{r,v,a}; the image feature extractor residual neural network, i.e. ResNet, is used for the resistivity imaging mode while drilling to map the original modal data to a unified dimension to obtain the feature m∈{i};x m =ResNet(o m ),m∈{i}; the unified dimensional features extracted from the four modalities are x m ,m∈{r,v,a,i};

[0012] Modality common and private feature extraction: Each modality feature is mapped twice to obtain the modality common feature space and modality private feature space. Specifically, different modality features are mapped through a common encoder E c Mapping to the common hidden features of the modalities m∈{r,v,a,i}: m∈{r,v,a,i}, the different modal features are passed through the private encoder E p Mapping to modality private hidden features m∈{r,v,a,i}: m∈{r,v,a,i}; among them, there are a total of encoders E c and private encoder E p They are all composed of a feedforward neural network and a Sigmoid activation function; the four modal features are respectively encoded by a common encoder E c Mapping to the modality-shared feature space and private encoder E p After mapping to the modality-private feature space, the four modalities have common hidden features. and four modal private hidden features

[0013] Modal fusion: Consider the features of the common subspace of modalities as the common modality Consider the features of the modality-private subspace as private modalities Stack the shared mode and private mode to get the matrix According to the Cross-modal Transformer model, the shared modal features are enhanced using private modal data, and the stacking matrix M is transformed into Next, use the Transformer model to extract features and get the matrix

[0014] Identification of unfavorable geological bodies: Matrix vector concatenation to get matrix After the matrix The final unfavorable geological body recognition result is obtained through a feedforward neural network G(·) Indicates whether there are unfavorable geological bodies in the drilling area and the type of unfavorable geological bodies.

[0015] Furthermore, in the data preprocessing step, the unfavorable geological bodies corresponding to the drilling data include faults, caves, weak interlayers, karst, expansive soils and groundwater-rich areas.

[0016] Furthermore, in the modal fusion step, the Cross-modal Transformer model is a deep learning model for processing multimodal data. It improves the performance of the task by combining information between different modalities. Each Cross-modal Transformer consists of D layers of cross-modal attention blocks. The calculation formula of the Cross-modal Transformer is:

[0017]

[0018] Among them, fθ is a feed-forward neural network parameterized by θ, represents the multi-head cross-modal attention mechanism, and LN represents layer normalization.

[0019] Furthermore, in the step of identifying the bad geological body, the feedforward neural network G(·) includes an input layer, a hidden layer and an output layer connected in sequence, and in a series of transformations of the hidden layer, each layer is provided with an activation function that enhances the nonlinear processing capability of the model, thereby being able to capture the complex relationship in the data.

[0020] Compared with the prior art, the intelligent identification method of unfavorable geological bodies while drilling based on multimodal fusion provided by the present invention collects four modal data including drilling rig operation parameters, drilling vibration parameters, drilling acoustic signals and drilling resistivity imaging images to form a multimodal drilling data set, and uses BiLSTM and ResNet to map the modal features of the four modal data to a unified dimension; on this basis, the four modal features are respectively mapped to a common encoder E c Mapping to the modality-shared feature space and private encoder E p Mapped to the modal private feature space, four modal common hidden features and four modal private hidden features are obtained; then the Cross-modal Transformer model is used to combine private modal data with common modal features to improve the performance of the task, and then the Transformer model is used for feature extraction, and finally the intelligent identification of poor geological bodies while drilling is realized through the feedforward neural network. Traditional identification while drilling mainly uses single modal data, but the amount of information carried by one modal data is limited and cannot contain all geological body information. Multimodal data can make full use of the complementarity of different data sources to provide a more comprehensive analysis for the identification of poor geological bodies. This technology can not only improve the accuracy and reliability of identification, but also enhance the anti-noise ability and filter out environmental interference signals through cross-validation of different data modalities. In addition, multimodal fusion can adapt to more complex geological environments, especially when multiple geological anomalies coexist, it can better identify potential poor geological bodies and provide real-time and accurate feedback. Therefore, the use of multimodal identification while drilling technology will effectively improve the safety and efficiency of underground engineering construction and provide solid technical support for construction under complex geological conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 It is a flow chart of the method for intelligent identification of adverse geological bodies while drilling based on multimodal fusion provided by the present invention.

[0022] Figure 2 It is a structural diagram of the Cross-modal Transformer model provided by the present invention. DETAILED DESCRIPTION

[0023] In order to make the technical means, creative features, objectives and effects achieved by the present invention easy to understand, the present invention is further described below with reference to specific diagrams.

[0024] Please refer to Figure 1 and Figure 2 As shown, the present invention provides a method for intelligently identifying unfavorable geological bodies while drilling based on multimodal fusion, comprising the following steps:

[0025] Data collection: Collect multimodal downhole data and corresponding adverse geological body information, the multimodal downhole data including drilling rig operation parameters, downhole vibration parameters, downhole acoustic signals and downhole resistivity imaging; specifically, the drilling rig operation parameters collect the working status data of the drilling rig in real time through built-in sensors or external monitoring equipment, including rotation speed, thrust, torque, drilling speed, etc.; the downhole vibration parameters monitor the vibration of the drill pipe in real time by installing an acceleration sensor near the drill bit; the downhole acoustic signal collects the acoustic signal generated by the drill bit breaking the rock in real time by installing an acoustic sensor near the drill bit; the downhole resistivity imaging is obtained in real time by installing a downhole resistivity imaging logging instrument near the drill bit.

[0026] Data preprocessing: Four modal data including drilling rig operating parameters, drilling vibration parameters, drilling acoustic signals and drilling resistivity imaging are selected to form a multimodal drilling data set. The adverse geological body information corresponding to the drilling data is selected to form a data set label. The data set label is used to guide the model to learn how to identify different adverse geological bodies based on the input multimodal drilling data. Through the correspondence between the model prediction results and the actual geological conditions, the model can learn by comparing the predictions and actual labels, and continuously optimize and improve the model. Specifically, the adverse geological bodies include faults, caves, weak interlayers, karst, expansive soils and groundwater-rich areas.

[0027] Feature extraction: The input of the multi-modal LWD dataset is four modes, namely, the drilling rig operation parameter mode r, the LWD vibration parameter mode v, the LWD acoustic signal mode a, and the LWD resistivity imaging mode i. The corresponding original modal data are m∈{r,v,a,i}; where, T m Input sample number for different modalities, d m are the feature dimensions of different modalities, and m is the modal category;

[0028] For the drilling rig operating parameters, drilling vibration parameters, and drilling acoustic signal modes, these three modes belong to low-dimensional modes, and the drilling data has time series, so the stacked bidirectional long short-term memory, namely BiLSTM, is used to map different original modal data to a unified dimension to obtain features. m∈{r,v,a}:x m =BiLSTM(om ),m∈{r,v,a}; for the LWD resistivity imaging mode, the image feature extractor residual neural network, i.e. ResNet, is used to map the original modal data to a unified dimension to obtain the feature m∈{i};x m =ResNet(o m ),m∈{i}; the unified dimensional features extracted from the four modalities are x m ,m∈{r,v,a,i}.

[0029] Extraction of common and private features of modalities: The characteristics of the geological body while drilling are manifested in different modes such as drilling rig operation parameters, drilling vibration parameters, drilling acoustic signals, and drilling resistivity imaging. Each modal information contains useful information for judging the characteristics of the geological body, which is the information shared by each mode. At the same time, the data structure and data form of each modal data are very different, so the information of different modes contains the private information of each mode. The features extracted from the four modes of drilling rig operation parameters, drilling vibration parameters, drilling acoustic signals, and drilling resistivity imaging are x m ,m∈{r,v,a,i}, each modal feature is mapped twice to obtain the modal common feature space and the modal private feature space. Specifically, the different modal features are mapped through the shared encoder E c Mapping to the common hidden features of the modalities m∈{r,v,a,i}: m∈{r,v,a,i}, with a total encoder E c It is composed of a feedforward neural network and a Sigmoid activation function, with a total of hidden features It is considered to be useful information for judging geological body characteristics in different modal features, because this part of information has the same tendency in geological body characteristic analysis, and is expressed as a similar distance in the common feature space. The common encoders of different modes use the same parameters.

[0030] The different modal features are passed through the private encoder E p Mapping to modality private hidden features m∈{r,v,a,i}: m∈{r,v,a,i}, private encoder E p It is also composed of a feedforward neural network and a Sigmoid activation function, with private hidden features It is considered to be the private information of geological body characteristics in different modal features. Because the private features of each modality are very different and have a long distance in the private feature space, the private encoders of different modalities use different parameters.

[0031] The four modal features are respectively encoded by the common encoder E c Mapping to the modality-shared feature space and private encoder E p After mapping to the modality-private feature space, the four modalities have common hidden features. and four modal private hidden features

[0032] Modal fusion: Consider the features of the common subspace of modalities as the common modality Consider the features of the modality-private subspace as private modalities Stack the shared mode and private mode to get the matrix According to the Cross-modal Transformer model, the shared modal features are enhanced using private modal data, and the stacking matrix M is transformed into

[0033] The Cross-modal Transformer model is a deep learning model for processing multimodal data. It improves the performance of tasks by combining information between different modalities. Each Cross-modal Transformer consists of D layers of cross-modal attention blocks. The calculation formula of Cross-modal Transformer is:

[0034]

[0035] Among them, f θ is a feed-forward neural network parameterized by θ, represents the multi-head cross-modal attention mechanism, and LN represents layer normalization;

[0036] Next, use the Transformer model to extract features and get the matrix

[0037] Identification of unfavorable geological bodies: Matrix vector concatenation to get matrix After the matrix The final unfavorable geological body recognition result is obtained through a feedforward neural network G(·) Indicates whether there are bad geological bodies in the drilling area and the type of bad geological bodies. The feedforward neural network G(·) transmits and processes information through a series of hierarchical neurons. In the input layer, the original feature data is input into the network, and then these data are transformed through a series of hidden layers. Each layer uses an activation function to enhance the nonlinear processing ability of the model, so that the complex relationship in the data can be captured, and finally the bad geological body identification result is obtained in the output layer. That is, the feedforward neural network G(·) includes an input layer, a hidden layer and an output layer connected in sequence, and in a series of transformations of the hidden layer, each layer is provided with an activation function that enhances the nonlinear processing capability of the model, thereby being able to capture complex relationships in the data.

[0038] Compared with the prior art, the intelligent identification method of unfavorable geological bodies while drilling based on multimodal fusion provided by the present invention collects four modal data including drilling rig operation parameters, drilling vibration parameters, drilling acoustic signals and drilling resistivity imaging images to form a multimodal drilling data set, and uses BiLSTM and ResNet to map the modal features of the four modal data to a unified dimension; on this basis, the four modal features are respectively mapped to a common encoder E c Mapping to the modality-shared feature space and private encoder E p Mapped to the modal private feature space, four modal common hidden features and four modal private hidden features are obtained; then the Cross-modal Transformer model is used to combine private modal data with common modal features to improve the performance of the task, and then the Transformer model is used for feature extraction, and finally the intelligent identification of poor geological bodies while drilling is realized through the feedforward neural network. Traditional identification while drilling mainly uses single modal data, but the amount of information carried by one modal data is limited and cannot contain all geological body information. Multimodal data can make full use of the complementarity of different data sources to provide a more comprehensive analysis for the identification of poor geological bodies. This technology can not only improve the accuracy and reliability of identification, but also enhance the anti-noise ability and filter out environmental interference signals through cross-validation of different data modalities. In addition, multimodal fusion can adapt to more complex geological environments, especially when multiple geological anomalies coexist, it can better identify potential poor geological bodies and provide real-time and accurate feedback. Therefore, the use of multimodal identification while drilling technology will effectively improve the safety and efficiency of underground engineering construction and provide solid technical support for construction under complex geological conditions.

[0039] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the purpose and scope of the technical solution of the present invention, which should be included in the scope of the claims of the present invention.

Claims

1. A method for intelligent identification of unfavorable geological bodies while drilling based on multimodal fusion, characterized in that: The following steps are involved: Data collection: Collect multi-modal while-drilling data and corresponding adverse geological body information, wherein the multi-modal while-drilling data includes drilling rig operation parameters, while-drilling vibration parameters, while-drilling acoustic signals and while-drilling resistivity imaging images; Data preprocessing: Four modal data, namely, drilling rig operating parameters, drilling vibration parameters, drilling acoustic signals, and drilling resistivity imaging, are selected to form a multimodal drilling dataset. The unfavorable geological bodies corresponding to the drilling data are selected to form dataset labels. The dataset labels are used to guide the model to learn how to identify different unfavorable geological bodies based on the input multimodal drilling data. Through the correspondence between the model prediction results and the actual geological conditions, the model can learn by comparing the predictions and actual labels, and continuously optimize and improve the model. Feature extraction: The input of the multi-modal LWD dataset is four modes, namely, the drilling rig operation parameter mode r, the LWD vibration parameter mode v, the LWD acoustic signal mode a, and the LWD resistivity imaging mode i. The corresponding original modal data are Among them, T m The number of input samples for different modalities, d m are the feature dimensions of different modalities, and m is the modal category; For the three types of modes, namely, drilling rig operating parameters, drilling vibration parameters, and drilling acoustic signals, we use stacked bidirectional long short-term memory (BiLSTM) to map different original modal data into a unified dimension to obtain features. x m =BiLSTM(o m ),m∈{r,v,a}; the image feature extractor residual neural network, i.e. ResNet, is used for the resistivity imaging mode while drilling to map the original modal data to a unified dimension to obtain the feature x m =ResNet(o m ),m∈{i}; the unified dimensional features extracted from the four modalities are x m ,m∈{r,v,a,i}; Modality common and private feature extraction: Each modality feature is mapped twice to obtain the modality common feature space and modality private feature space. Specifically, different modality features are mapped through a common encoder E c Mapping to the common hidden features of the modalities The different modal features are passed through the private encoder E p Mapping to modality private hidden features Among them, there are a total of encoders E c and private encoder E p They are all composed of a feedforward neural network and a Sigmoid activation function; the four modal features are respectively encoded by a common encoder E c Mapping to the modality common feature space and private encoder E p After mapping to the modality-private feature space, the four modalities have common hidden features. and four modal private hidden features Modal fusion: Consider the features of the common subspace of modalities as the common modality Consider the features of the modality-private subspace as private modalities Stack the shared mode and private mode to get the matrix According to the Cross-modal Transformer model, the shared modal features are enhanced using private modal data, and the stacking matrix M is transformed into Next, use the Transformer model to extract features and get the matrix Identification of unfavorable geological bodies: Matrix vector concatenation to get matrix After the matrix The final unfavorable geological body recognition result is obtained through a feedforward neural network G(·) Indicates whether there is any adverse geological body information in the drilling area and the type of adverse geological body.

2. The method for intelligent identification of unfavorable geological bodies while drilling based on multimodal fusion according to claim 1 is characterized in that: In the data preprocessing step, the unfavorable geological bodies corresponding to the drilling data include faults, caves, weak interlayers, karst, expansive soils and groundwater-rich areas.

3. The method for intelligent identification of unfavorable geological bodies while drilling based on multimodal fusion according to claim 1, characterized in that: In the modal fusion step, the Cross-modal Transformer model is a deep learning model for processing multimodal data. It improves the performance of the task by combining information between different modalities. Each Cross-modal Transformer consists of D layers of cross-modal attention blocks. The calculation formula of the Cross-modal Transformer is: Among them, f θ is a feed-forward neural network parameterized by θ, represents the multi-head cross-modal attention mechanism, and LN represents layer normalization.

4. The method for intelligent identification of unfavorable geological bodies while drilling based on multimodal fusion according to claim 1, characterized in that: In the step of identifying the bad geological body, the feedforward neural network G(·) includes an input layer, a hidden layer and an output layer connected in sequence, and in a series of transformations of the hidden layer, each layer is provided with an activation function that enhances the nonlinear processing capability of the model, thereby being able to capture the complex relationship in the data.

Images