Advanced geological prediction system and method for shield machine construction

Abstract

The present application relates to the technical field of shield machine construction, and discloses an advanced geological prediction system and method for shield machine construction, which generates a detection wave signal through an excitation device and uses a detection device to receive a reflected wave signal reflected by the stratum. Subsequently, the reflected wave signal is transmitted to a data processing terminal for analog-to-digital conversion and preliminary filtering to obtain a digital reflected wave signal. In order to further improve the signal quality, empirical mode decomposition and adaptive filtering based on feature distillation are used to perform joint denoising and enhancement processing on the signal in the time and frequency domains, and finally the enhanced reflected wave signal features are extracted to judge the geological conditions ahead. In this way, the accuracy and reliability of geological prediction are significantly improved, and it is particularly suitable for shield construction under complex geological conditions.

Description

Technical Field

[0001] The present application relates to the technical field of shield machine construction, and more specifically, to an advanced geological prediction system and method for shield machine construction. Background Art

[0002] In civil engineering, especially during shield tunneling, geological advance prediction technology is crucial for ensuring construction safety and efficiency. While existing technologies, such as the shield tunneling advance prediction system and method described in patent application publication number CN111123351A, offer an effective solution, practical applications still require improvement.

[0003] Specifically, the solution proposed in CN111123351A relies primarily on a system consisting of a drilling device, an excitation device, and a detector device to achieve advanced geological prediction. This system drills holes in the tunnel wall, uses the excitation device to generate vibration signals, and then uses the detector device to receive and analyze the reflected signals. However, in actual operation, interference from background electromagnetic noise is a significant problem. Because shield construction sites typically have multiple mechanical devices operating simultaneously, this can cause strong electromagnetic interference, which affects the quality of the signals received by the detector device and reduces the accuracy of the prediction results.

[0004] Therefore, an optimized advanced geological prediction scheme for shield machine construction is expected. Summary of the Invention

[0005] In order to solve the above technical problems, the present application is proposed. The embodiments of the present application provide an advanced geological prediction system and method for shield machine construction, which improves the accuracy and reliability of geological prediction and is particularly suitable for shield construction under complex geological conditions.

[0006] According to one aspect of the present application, a method for advanced geological prediction for shield machine construction is provided, comprising: using a detection device to receive a reflected wave signal, wherein the reflected wave is formed by reflecting a detection wave signal emitted by an excitation device through a target stratum; the detection device transmits the reflected wave signal to a data processing terminal; the data processing terminal performs analog-to-digital conversion on the reflected wave signal to obtain a digital reflected wave signal; the data processing terminal performs time-frequency domain joint denoising and signal enhancement on the digital reflected wave signal to obtain an enhanced reflected wave signal, wherein the time-frequency domain joint denoising and signal enhancement on the digital reflected wave signal comprises: performing empirical mode decomposition and feature distillation-based adaptive filtering on the digital reflected wave signal; the data processing terminal extracts reflected wave signal features from the enhanced reflected wave signal, and judges the geological conditions ahead based on the reflected wave signal features to achieve advanced prediction.

[0007] In the above-mentioned advanced geological prediction method for shield machine construction, the data processing terminal performs analog-to-digital conversion on the reflected wave signal to obtain a digital reflected wave signal, including: converting the reflected wave signal into a digital signal to obtain an initial digital reflected wave signal; and using a bandpass filter to preliminarily filter the initial digital reflected wave signal to obtain the digital reflected wave signal.

[0008] In the above-mentioned advanced geological prediction method for shield machine construction, the data processing terminal performs time-frequency domain joint denoising and signal enhancement on the digital reflected wave signal to obtain an enhanced reflected wave signal, including: performing empirical mode decomposition on the digital reflected wave signal to obtain a set of signal inherent modal components; performing feature distillation-based adaptive filtering on the set of signal inherent modal components to obtain a set of filtered signal inherent modal components; and performing signal reconstruction on the set of filtered signal inherent modal components to obtain the enhanced reflected wave signal.

[0009] In the advanced geological prediction method for shield machine construction, the set of signal intrinsic modal components is subjected to feature distillation-based adaptive filtering to obtain a set of filtered signal intrinsic modal components, including: extracting signal intrinsic modal components to be distilled from the set of signal intrinsic modal components; determining a signal intrinsic modal feature distillation strength metric based on the signal intrinsic modal components to be distilled and the set of signal intrinsic modal components; and determining whether to distill the signal intrinsic modal components to be distilled based on the signal intrinsic modal feature distillation strength metric.

[0010] In the above-mentioned advanced geological prediction method for shield machine construction, based on the inherent modal components of the signal to be distilled and the set of the signal inherent modal components, the signal inherent modal feature distillation strength measurement value is determined, including: inputting the set of the signal inherent modal components into the prior feature extraction self-organizing network to obtain the signal inherent modal prior feature aggregation coding vector; inputting the signal inherent modal components to be distilled and the signal inherent modal prior feature aggregation coding vector into the feature distillation guided network to obtain the signal inherent modal component distillation potential representation matrix; performing context-aware gain on the signal inherent modal component distillation potential representation matrix to obtain the signal inherent modal component context-aware gain mapping matrix; and determining the signal inherent modal feature distillation strength measurement value based on the signal inherent modal component context-aware gain mapping matrix.

[0011] In the above-mentioned advanced geological prediction method for shield machine construction, the inherent modal component of the signal to be distilled and the inherent modal prior feature aggregation coding vector of the signal are input into the feature distillation guidance network to obtain the signal inherent modal component distillation potential representation matrix, including: multiplying the inherent modal component of the signal to be distilled and the transposed vector of the inherent modal prior feature aggregation coding vector of the signal, and then dividing by the length of the inherent modal prior feature aggregation coding vector of the signal to obtain the signal inherent modal component distillation potential representation matrix.

[0012] In the above-mentioned advanced geological prediction method for shield machine construction, context-aware gain is performed on the signal intrinsic modal component distillation potential representation matrix to obtain a signal intrinsic modal component context-aware gain mapping matrix, including: calculating the spatial proximity similarity factor of the signal intrinsic modal component to be distilled and the signal intrinsic modal prior feature aggregate coding vector; calculating the distribution proximity representation factor of the signal intrinsic modal component to be distilled and the signal intrinsic modal prior feature aggregate coding vector; combining the spatial proximity similarity factor, the distribution proximity representation factor and the signal intrinsic modal component distillation potential representation matrix to construct a signal intrinsic modal component context-aware gain mapping matrix.

[0013] In the above-mentioned advanced geological prediction method for shield machine construction, the signal inherent modal feature distillation strength measurement value is determined based on the signal inherent modal component context-aware gain mapping matrix, including: calculating the trace value of the signal inherent modal component context-aware gain mapping matrix as the signal inherent modal feature distillation strength measurement value.

[0014] In the above-mentioned advanced geological prediction method for shield machine construction, the data processing terminal extracts the reflected wave signal characteristics from the enhanced reflected wave signal, including: using wavelet transform to extract the signal characteristics of the enhanced reflected wave signal to obtain the reflected wave signal characteristics.

[0015] According to another aspect of the present application, an advanced geological prediction system for shield machine construction is also provided, including: a reflected wave signal receiving module, used to receive a reflected wave signal using a detection device, wherein the reflected wave is formed by the detection wave signal emitted by the excitation device being reflected by the target stratum; a reflected wave signal transmission module, used for the detection device to transmit the reflected wave signal to a data processing terminal; an analog-to-digital conversion module, used for the data processing terminal to perform analog-to-digital conversion on the reflected wave signal to obtain a digital reflected wave signal; a reflected wave signal enhancement module, used for the data processing terminal to perform time-frequency domain joint denoising and signal enhancement on the digital reflected wave signal to obtain an enhanced reflected wave signal, wherein the time-frequency domain joint denoising and signal enhancement on the digital reflected wave signal includes: performing empirical mode decomposition and adaptive filtering based on feature distillation on the digital reflected wave signal; a geological condition judgment module, used for the data processing terminal to extract reflection wave signal characteristics from the enhanced reflection wave signal, and judge the geological conditions ahead based on the reflection wave signal characteristics to achieve advanced prediction.

[0016] Compared with the existing technology, the advanced geological prediction system and method for shield machine construction provided by this application generates a detection wave signal through an excitation device, and uses a detection device to receive the reflected wave signal reflected by the stratum. Subsequently, the reflected wave signal is transmitted to the data processing terminal for analog-to-digital conversion and preliminary filtering to obtain a digital reflected wave signal. In order to further improve the signal quality, empirical mode decomposition and adaptive filtering based on feature distillation are used to perform joint denoising and enhancement processing on the signal in the time and frequency domains, and finally the enhanced reflected wave signal features are extracted to judge the geological conditions ahead. In this way, the accuracy and reliability of geological prediction are significantly improved, and it is particularly suitable for shield construction under complex geological conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The above and other purposes, features, and advantages of the present application will become more apparent through a more detailed description of the embodiments of the present application in conjunction with the accompanying drawings. The accompanying drawings are intended 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 drawings, the same reference numerals generally represent the same components or steps.

[0018] Figure 1 This is a schematic flow chart of an advanced geological prediction method for shield machine construction according to an embodiment of the present application.

[0019] Figure 2 This is a schematic flow chart of step 3 in the advanced geological prediction method for shield machine construction according to an embodiment of the present application.

[0020] Figure 3 This is a schematic flow chart of step 4 in the advanced geological prediction method for shield machine construction according to an embodiment of the present application.

[0021] Figure 4 This is a schematic flow chart of step 42 in the method for advanced geological prediction for shield machine construction according to an embodiment of the present application.

[0022] Figure 5 This is a schematic flowchart of step 422 in the advanced geological prediction method for shield machine construction according to an embodiment of the present application.

[0023] Figure 6 This is a schematic block diagram of an advanced geological prediction system for shield machine construction according to an embodiment of the present application. DETAILED DESCRIPTION

[0024] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described herein.

[0025] Figure 1 FIG. 1 is a schematic flow chart of a method for advanced geological prediction for shield machine construction according to an embodiment of the present application. Figure 1 As shown, the advanced geological prediction method for shield machine construction includes: S1, using a detection device to receive a reflected wave signal, wherein the reflected wave is formed by the detection wave signal emitted by the excitation device and reflected by the target stratum; S2, the detection device transmits the reflected wave signal to a data processing terminal; S3, the data processing terminal performs analog-to-digital conversion on the reflected wave signal to obtain a digital reflected wave signal; S4, the data processing terminal performs time-frequency domain joint denoising and signal enhancement on the digital reflected wave signal to obtain an enhanced reflected wave signal; S5, the data processing terminal extracts the reflected wave signal feature from the enhanced reflected wave signal, and judges the geological conditions ahead based on the reflected wave signal feature to achieve advanced prediction.

[0026] Specifically, in step S1, a detector is used to receive a reflected wave signal, which is formed by the detection wave signal emitted by the excitation device and reflected by the target stratum. It should be understood that during shield construction, geological advance prediction is of vital importance to ensuring project safety and preventing potential geological disasters. Using a detector to receive the reflected wave signal formed by the detection wave signal emitted by the excitation device and reflected by the target stratum is one of the key steps to achieve this goal. This method relies on seismic wave theory, that is, when the energy wave generated by the vibration source (excitation device) encounters the interface between different media, it will be reflected. By analyzing the characteristics of these reflected waves, the properties and structure of the underground medium can be inferred.

[0027] Specifically, the excitation device acts as a source of vibration, which can generate vibration signals of specific frequency and intensity. These signals propagate to the surrounding medium in the form of waves and are reflected when encountering stratum boundaries or abnormal bodies. The detector is responsible for capturing these reflected wave signals and converting them into electrical signals for transmission to the data processing terminal for further analysis. The advantage of this method lies in its non-invasiveness and adaptability to complex geological environments. It is especially suitable for underground conditions that cannot be directly observed. For example, in shield construction, since the stratum conditions in front of the tunnel are unknown, the use of this technology based on reflected waves can help engineers understand in advance the possible geological changes ahead, such as weak interlayers, caves, etc., so that they can take corresponding countermeasures.

[0028] In a specific embodiment, holes are first drilled around the support ring of the shield machine. These holes are usually distributed radially and form a circular observation system. For example, 8 or more holes can be drilled, one of which is used to place the excitation device, and the remaining holes are used to install the detection device. Then, the excitation device is placed in the pre-drilled hole, ensuring that it is in close contact with the surrounding rock mass so as to effectively transmit the vibration signal. Once the excitation device is activated, it generates a series of vibration pulses, which propagate in the stratum and are reflected back when encountering different geological interfaces. At the same time, the detection devices distributed in other holes start working. They will record the time, amplitude and other information of these reflected wave signals, and send these data to the data processing terminal via wired or wireless means.

[0029] Specifically, in step S2, the detection device transmits the reflected wave signal to the data processing terminal. It should be understood that the detection wave signal generated by the excitation device is reflected when encountering the interface between different media, and these reflected waves carry important information about the underground geological structure. However, it is difficult to directly interpret the geological information contained in the original signal collected on site, and it must be further processed and analyzed. Therefore, transmitting these reflected wave signals to the data processing terminal becomes a necessary step. The data processing terminal is equipped with advanced algorithms and technologies, which can perform a series of complex operations on the signal, such as analog-to-digital conversion, filtering and denoising, and feature extraction, thereby revealing detailed information of the stratum, such as lithology and structural characteristics.

[0030] In one specific embodiment, the connection between the detector device and the data processing terminal can utilize either wired or wireless communication technologies. For example, in a wired connection, the detector device is typically equipped with a highly sensitive sensor to capture reflected wave signals and convert them into electrical signals. These electrical signals are then transmitted via signal transmission lines (such as shielded cables) to the data processing terminal located on the ground or inside the shield machine. To ensure signal integrity and accuracy, the signal transmission lines must exhibit excellent anti-interference properties to avoid the influence of external electromagnetic noise. Furthermore, to adapt to complex construction environments, the transmission lines must be flexible and durable for ease of installation and maintenance.

[0031] In another specific embodiment, considering the complexity and space limitations of the construction site, a wireless communication solution may be more suitable. In this case, the detection device not only includes a sensor but also needs to integrate a wireless transmission module. This module is responsible for digitizing the received reflected wave signal and transmitting it via radio waves of a specific frequency. Correspondingly, the data processing terminal needs to be equipped with a receiving module to receive data from each detection point. The advantage of a wireless transmission solution is that it reduces the wiring workload and improves the flexibility and scalability of the system. However, to ensure the reliability and real-time performance of data transmission, effective encryption and error correction mechanisms must be implemented to prevent signal loss or bit errors.

[0032] Specifically, in step S3, the data processing terminal performs analog-to-digital conversion on the reflected wave signal to obtain a digital reflected wave signal. It should be understood that in complex shield construction environments, the original reflected wave signal is often affected by various interference factors, such as electromagnetic noise generated by the operation of mechanical equipment and other sound sources in the external environment. Direct analysis using unprocessed analog signals would make it difficult to accurately extract effective information about the underground geological structure. Therefore, digitizing analog signals through analog-to-digital conversion not only facilitates the storage and transmission of large amounts of data but also provides a foundation for subsequent advanced signal processing.

[0033] In one embodiment, Figure 2 As shown, the data processing terminal performs analog-to-digital conversion on the reflected wave signal to obtain a digital reflected wave signal, including: S31, converting the reflected wave signal into a digital signal to obtain an initial reflected wave digital signal; S32, using a bandpass filter to preliminarily filter the initial reflected wave digital signal to obtain the digital reflected wave signal.

[0034] In step S31, the data processing terminal is equipped with a high-precision analog-to-digital converter (ADC) to convert the analog reflected wave signal transmitted by the detector into a digital signal. For example, at a typical shield tunneling construction site, when the vibration signal generated by the excitation device is reflected from different stratum interfaces and captured by the detector, it is first transmitted to the data processing terminal via a shielded cable. Before entering the data processing terminal, the signal may undergo some preprocessing steps, such as amplification or preliminary filtering, to ensure signal quality. Once the analog signal reaches the data processing terminal, the analog-to-digital converter begins operation. In one specific embodiment, the ADC used has 16-bit resolution and a million samples per second (MS / s) capability. It discretizes the continuously varying analog signal into a series of binary values. Each sample point represents the signal amplitude value at a specific moment, forming a time series data set. This process requires not only high-precision hardware support but also a strict synchronization mechanism to ensure sampling consistency across all channels, which is particularly important in multi-channel systems.

[0035] In step S32, a bandpass filter is used to perform preliminary filtering on the initial digital signal of the reflected wave. This step is intended to remove background noise and other interfering factors and improve the signal-to-noise ratio. For example, a bandpass filter can be used with a frequency range set within a specific range to retain useful seismic wave components and eliminate unnecessary high-frequency noise or low-frequency drift.

[0036] The design of a bandpass filter must consider the specific geological exploration requirements and the characteristics of the on-site environment. In one specific embodiment, if the reflected waves from the target formation are primarily concentrated between 10Hz and 100Hz, the passband of the bandpass filter should be set within this range. Furthermore, to adapt to varying construction conditions, the filter parameters can be dynamically adjusted based on actual conditions. For example, in complex geological conditions, a narrower passband may be required to enhance sensitivity to specific frequency bands; whereas in relatively simple geological environments, the passband restriction can be relaxed to obtain more information. The use of a bandpass filter for preliminary filtering is also based on a deep understanding of signal characteristics. As seismic waves propagate through the formation, they experience varying degrees of attenuation and scattering depending on the medium, resulting in a large amount of useless components in the received reflected wave signal. If not filtered, these components can mislead the final geological interpretation. By properly designing the filter parameters, noise can be effectively suppressed, the characteristics of the target signal can be highlighted, and the accuracy of geological forecasts can be improved.

[0037] Specifically, in step S4, the data processing terminal performs joint time-frequency domain denoising and signal enhancement on the digital reflected wave signal to obtain an enhanced reflected wave signal. It should be understood that in a complex shield construction environment, the original reflected wave signal is often affected by various interference factors, such as electromagnetic noise generated by the operation of mechanical equipment and other sound sources in the external environment. Directly analyzing the unprocessed analog signal makes it difficult to accurately extract effective information about the underground geological structure. Therefore, the digital reflected wave signal is subjected to joint time-frequency domain denoising and signal enhancement to obtain an enhanced reflected wave signal.

[0038] In one embodiment, Figure 3 As shown, the data processing terminal performs time-frequency domain joint denoising and signal enhancement on the digital reflected wave signal to obtain an enhanced reflected wave signal, including: S41, performing empirical mode decomposition on the digital reflected wave signal to obtain a set of signal intrinsic modal components; S42, performing feature distillation-based adaptive filtering on the set of signal intrinsic modal components to obtain a set of filtered signal intrinsic modal components; S43, performing signal reconstruction on the set of filtered signal intrinsic modal components to obtain the enhanced reflected wave signal.

[0039] In step S41, the data processing terminal performs empirical mode decomposition (EMD) on the received digital reflected wave signal. This process decomposes the digital reflected wave signal into multiple intrinsic mode components (IMCs), each representing a signal component in a different frequency band. In one specific embodiment, the received reflected wave signal contains multiple frequency components, and EMD can be used to separate these components, each corresponding to a different stratigraphic structure or anomaly. In practice, the EMD algorithm repeatedly filters the signal, removing the local mean with each filter, until a stopping condition is met. This ultimately yields a series of IMCs, arranged in descending order of frequency, to form a set of IMCs.

[0040] Specifically, empirical mode decomposition (EMD), as an adaptive signal processing method, can dynamically adapt to changes in different frequency components, avoiding the distortion problems that traditional fixed filters may introduce. Especially in complex geological conditions, seismic signals often contain multi-band components. Traditional Fourier transforms or other fixed-frequency filtering methods struggle to fully capture all frequency components. EMD, on the other hand, can flexibly decompose the signal into multiple IMFs based on actual signal variations, thereby better capturing the reflection characteristics of the bed interface.

[0041] In step S42, the set of signal intrinsic modal components is adaptively filtered based on feature distillation to obtain a set of filtered signal intrinsic modal components. It should be understood that while empirical mode decomposition (EMD) can decompose the original signal into several intrinsic modal components with different frequency characteristics, these components contain both components representing true geological information and potentially a significant amount of noise. Directly using these unprocessed components for geological analysis can lead to misjudgments or inaccurate forecasts. By employing adaptive filtering technology based on feature distillation, noise interference can be intelligently identified and removed based on the specific characteristics of each intrinsic modal component, thereby retaining those signal components that best reflect the true geological state. The benefits of this approach are obvious: on the one hand, it improves the reliability and accuracy of subsequent geological analysis; on the other hand, by optimizing signal quality, it enables deeper data mining and pattern recognition, enabling a more accurate three-dimensional image of the geological structure ahead, thereby achieving more effective early warning.

[0042] In one embodiment, Figure 4 As shown, the set of signal intrinsic modal components is adaptively filtered based on feature distillation to obtain a set of filtered signal intrinsic modal components, including: S421, extracting the signal intrinsic modal components to be distilled from the set of signal intrinsic modal components; S422, determining the signal intrinsic modal feature distillation strength measurement value based on the signal intrinsic modal components to be distilled and the set of signal intrinsic modal components; S423, determining whether to distill the signal intrinsic modal components to be distilled based on the signal intrinsic modal feature distillation strength measurement value.

[0043] In step S421, the process of extracting the signal intrinsic modal components to be distilled from the set of signal intrinsic modal components can be expressed as follows: ;in, is the set of intrinsic modal components of the signal, and are the first, second, and third components in the set of signal intrinsic modal components. and The intrinsic modal components of the signal, The first in the set of intrinsic modal components of the signal The intrinsic modal components of the signal, is the index, Can be used to refer to any signal intrinsic modal component in the set, Represents the intrinsic modal components of the signal to be distilled. It should be understood that the set of signal intrinsic modal components contains multiple levels of modal information after the decomposition of the original signal. This information covers both the detailed expression of local features and the potential pattern of the global structure. However, not all modal components have the same contribution or importance to the current task. By taking each signal intrinsic modal component in the set of signal intrinsic modal components as the signal intrinsic modal component to be distilled in turn and analyzing them separately to determine whether they need to be distilled, effective distillation of the entire set of signal intrinsic modal components can be achieved.

[0044] It should be understood that in one embodiment, Figure 5 As shown, in step S422, based on the signal intrinsic modal components to be distilled and the set of signal intrinsic modal components, a signal intrinsic modal feature distillation strength measure value is determined, including: S4221, inputting the set of signal intrinsic modal components into a priori feature extraction self-organizing network to obtain a signal intrinsic modal prior feature aggregation coding vector; S4222, inputting the signal intrinsic modal components to be distilled and the signal intrinsic modal prior feature aggregation coding vector into a feature distillation guided network to obtain a signal intrinsic modal component distillation potential representation matrix; S4223, performing context-aware gain on the signal intrinsic modal component distillation potential representation matrix to obtain a signal intrinsic modal component context-aware gain mapping matrix; S4224, determining the signal intrinsic modal feature distillation strength measure value based on the signal intrinsic modal component context-aware gain mapping matrix.

[0045] In step S4221, the set of intrinsic modal components of the signal is input into the prior feature extraction self-organizing network to obtain the signal intrinsic modal prior feature aggregation coding vector. It should be understood that the process of inputting the set of intrinsic modal components of the signal into the prior feature extraction self-organizing network uses a self-supervised learning mechanism to automatically learn and extract high-level feature representations with strong semantic expression capabilities from unlabeled data. Specifically, through this processing method, the potential structural patterns and global context information in the original signal can be captured and converted into a more abstract but meaningful coding form. This coding not only carries the global information of the input feature sequence, but also reveals the hidden relational patterns inside the data through the self-supervisory mechanism, thereby providing a strong prior guidance signal for subsequent feature distillation. Specifically, the process can be expressed by the formula: ;in, and Take The maximum and minimum values ​​in represents the adjustment of hyperparameters, Indicates the The dynamic adjustment coefficient of the signal's inherent mode, Indicates the Dynamically adjust the weight coefficient of the signal's inherent mode, Represents the signal intrinsic modal prior feature aggregation coding vector. The function of adjusting the hyperparameter is to control the sensitivity of the signal intrinsic modal dynamic adjustment coefficient by quantifying the ratio of the signal extreme difference to the dynamic constraint. When it increases, the denominator increases and the The value of , thereby weakening the signal dynamic response capability; on the contrary Zoom in when you zoom out , enhancing the capture of subtle changes. Based on experience, the hyperparameter is preset to 0.5, which can be adjusted according to actual conditions and is not specifically limited in this embodiment.

[0046] In one embodiment, in step S4222, the signal intrinsic modal components to be distilled and the signal intrinsic modal prior feature aggregation coding vector are input into a feature distillation guidance network to obtain a signal intrinsic modal component distillation potential representation matrix, including: multiplying the signal intrinsic modal components to be distilled and the transposed vector of the signal intrinsic modal prior feature aggregation coding vector, and then dividing by the length of the signal intrinsic modal prior feature aggregation coding vector to obtain the signal intrinsic modal component distillation potential representation matrix. Specifically, the process can be expressed as follows: ;in, is matrix multiplication, represents the transpose of the matrix, is the length of the signal intrinsic modal prior feature aggregation coding vector, yes The corresponding signal intrinsic modal components are distilled into the latent representation matrix.

[0047] It should be understandable that feature distillation guides the network to focus on the feature subspace that is most relevant to the distillation task. In this process, the intrinsic modal component of the signal to be distilled acts as a query component and interacts with the prior information aggregate coding vector as a template vector to implicitly model the correlation between the two through the signal intrinsic modal component distillation potential representation matrix. That is, the correlation between the intrinsic modal component of the signal to be distilled and the signal intrinsic modal prior feature aggregate coding vector is dynamically captured through the signal intrinsic modal component distillation potential representation matrix, and this relationship is explicitly expressed through the signal intrinsic modal component distillation potential representation matrix. In particular, the signal intrinsic modal component distillation potential representation matrix not only integrates the deep information of the target feature and the global context, but also filters out irrelevant features, making subsequent feature distillation more accurate.

[0048] In step S4223, the potential representation matrix is ​​distilled considering the intrinsic modal components of the signal When explicitly expressing the correlation between the distillation target and the global prior information, it is expected to further enhance the feature vector as the associated local and Distill the latent representation matrix with respect to the intrinsic modal components of the signal That is, context-aware gain is performed on the signal intrinsic modal component distillation potential representation matrix to obtain a signal intrinsic modal component context-aware gain mapping matrix.

[0049] First, we use the spatial proximity similarity factor To constrain component stability and use the distribution proximity representation factor To control the overall interactivity of the components, based on this, the spatial proximity similarity factor of the intrinsic modal component of the signal to be distilled and the aggregated coding vector of the intrinsic modal prior feature of the signal is first calculated, and the distribution proximity representation factor of the intrinsic modal component of the signal to be distilled and the aggregated coding vector of the intrinsic modal prior feature of the signal is calculated, which is expressed as: ;in, represents the two-norm of the vector, Represents vector subtraction.

[0050] Then, based on the historical knowledge fusion mechanism, the component feature linkage strength parameter mapping is introduced into the latent space representation learning framework to construct the context-aware gain mapping matrix of the signal intrinsic modal component. That is, the spatial proximity similarity factor, the distribution proximity representation factor, and the distilled latent representation matrix of the signal intrinsic modal component are combined to construct the context-aware gain mapping matrix of the signal intrinsic modal component, which is expressed as: ;in, Indicates point multiplication by position, represents the first signal intrinsic mode mapping vector, represents the second signal intrinsic modal mapping vector.

[0051] In this way, on the basis of establishing the spatial rules of dimensional interaction, the distribution difference sensitivity adaptation of local contributions is strengthened, that is, the adaptation of global semantic explicit encoding and local feature gain distribution sensitivity is driven, and the potential representation matrix of the signal intrinsic modal component distillation is realized. Deep self-supervised balancing of global context.

[0052] In one embodiment, in step S4224, determining a signal intrinsic modal feature distillation strength metric based on the signal intrinsic modal component context-aware gain mapping matrix includes calculating a trace measure of the signal intrinsic modal component context-aware gain mapping matrix as the signal intrinsic modal feature distillation strength metric. Specifically, the process can be expressed as follows: ;in, is the trace value of the matrix, for The corresponding context-aware gain mapping matrix of the signal's intrinsic modal components.

[0053] It should be understood that by calculating the trace value of the context-aware gain mapping matrix of the signal's intrinsic modal components to determine the signal's intrinsic modal feature distillation strength measure, we can effectively evaluate the expressive power of each component and its contribution to the overall feature representation. This quantitative metric not only reflects the optimization potential of a specific component in its current state but also measures its consistency and relevance with global prior information. Therefore, based on this signal's intrinsic modal feature distillation strength measure, more accurate decisions can be made.

[0054] In step S423, based on the signal inherent modal feature distillation strength measure value, it is determined whether to distill the signal inherent modal component to be distilled. That is, if the signal inherent modal feature distillation strength measure value of a certain signal inherent modal component indicates that it has a high expressive ability and is highly consistent with the global semantic information, then it can be considered that the signal inherent modal component is related to the advance prediction task, and the signal inherent modal component is retained. If the signal inherent modal feature distillation strength measure value shows that a certain component is significantly different from the global information, then it can be considered that the signal inherent modal component has a low correlation with the advance prediction task, and the distillation program is started to remove the signal inherent modal component. Specifically, the process can be expressed by the formula: ;in, Indicates the preset threshold, Indicates whether to perform distillation on the intrinsic modal component of the signal to be distilled. The threshold is preset to 1.8 based on experience and can be adjusted according to actual conditions. This embodiment does not impose a specific limitation.

[0055] In step S43, the set of the filtered signal intrinsic modal component is carried out signal reconstruction. Although EMD and adaptive filtering have greatly improved signal quality, to really bring into play its value, it is also necessary to reassemble the set of the filtered signal intrinsic modal component into a complete and coherent signal. In this process, the phase relationship and the energy distribution between the filtered signal intrinsic modal component must be considered to ensure that the reconstructed signal has both maintained the characteristic of the original signal and reached higher signal-to-noise ratio. In a specific embodiment, the frequency range and relative intensity of the filtered signal intrinsic modal component are matched to generate the final enhanced reflected wave signal.

[0056] Specifically, in step S5, the data processing terminal extracts reflection wave signal features from the enhanced reflection wave signal and, based on these features, determines the geological conditions ahead, thereby achieving advanced forecasting. It should be understood that the process of extracting reflection wave signal features and determining geological conditions not only reveals the basic physical properties of the underground medium but also allows for further inference of deeper geological information. For example, by analyzing characteristics such as the arrival time, amplitude variation, and frequency distribution of the reflection wave, it is possible to infer the thickness of the stratum, the lithology type, and the presence of anomalies. This is of great significance for formulating reasonable construction plans and selecting appropriate support measures.

[0057] In one embodiment, the data processing terminal extracts reflected wave signal features from the enhanced reflected wave signal, including extracting the signal features of the enhanced reflected wave signal using a wavelet transform. Specifically, the wavelet transform provides high-resolution information in both the time and frequency domains, making it well-suited for processing non-stationary signals. By performing a continuous wavelet transform on the signal, wavelet coefficient maps at different scales can be obtained, from which the main frequency components of the reflected wave and their temporal variation trends can be clearly visualized. This information can be used to further extract key characteristic parameters reflecting geological characteristics, such as the depth of the reflection interface and the location of the lithologic change point.

[0058] Based on the extracted reflection wave signal characteristics, the next step is to determine the geological conditions ahead. This step is usually carried out through a comprehensive analysis combining geological models and historical data. For example, based on the known stratigraphic distribution map and previous exploration results, combined with the currently acquired reflection wave characteristic parameters, the specific geological conditions ahead of the tunnel can be inferred.

[0059] In one embodiment, a geographic information system (GIS) and 3D visualization software are used to assist in the analysis. For example, features extracted through wavelet transforms show a significant increase in the amplitude of reflected waves within a certain depth range. Combined with geological model analysis, this may be due to the presence of a hard rock layer or a karst cave in that area. By constructing a 3D structural map depicting the different geological conditions (such as anomalous rock masses, lithologies, and karst features) ahead of the tunnel working face and above and below the tunnel strike, geological advance prediction can be achieved.

[0060] In summary, the advanced geological prediction method for shield machine construction provided by this application has been explained, which generates a detection wave signal through an excitation device and uses a detection device to receive the reflected wave signal reflected by the stratum. Subsequently, the reflected wave signal is transmitted to the data processing terminal for analog-to-digital conversion and preliminary filtering to obtain a digital reflected wave signal. In order to further improve the signal quality, empirical mode decomposition and adaptive filtering based on feature distillation are used to perform joint denoising and enhancement processing on the signal in the time and frequency domains, and finally the enhanced reflected wave signal features are extracted to judge the geological conditions ahead. In this way, the accuracy and reliability of geological prediction are significantly improved, and it is particularly suitable for shield construction under complex geological conditions.

[0061] This application also provides an advanced geological prediction system for shield machine construction, such as Figure 6 As shown, the advanced geological prediction system 600 for shield machine construction includes: a reflected wave signal receiving module 610, which is used to receive the reflected wave signal using a detector device, wherein the reflected wave is formed by the detection wave signal emitted by the excitation device and reflected by the target stratum; a reflected wave signal transmission module 620, which is used for the detector device to transmit the reflected wave signal to a data processing terminal; an analog-to-digital conversion module 630, which is used for the data processing terminal to perform analog-to-digital conversion on the reflected wave signal to obtain a digital reflected wave signal; a reflected wave signal enhancement module 640, which is used for the data processing terminal to perform time-frequency domain joint denoising and signal enhancement on the digital reflected wave signal to obtain an enhanced reflected wave signal, wherein the time-frequency domain joint denoising and signal enhancement on the digital reflected wave signal includes: performing empirical mode decomposition and adaptive filtering based on feature distillation on the digital reflected wave signal; and a geological condition judgment module 650, which is used for the data processing terminal to extract reflected wave signal features from the enhanced reflected wave signal and judge the geological conditions ahead based on the reflected wave signal features to achieve advanced prediction.

[0062] An embodiment of the present application also provides a computer-readable storage medium, which stores computer program code. When the computer program code runs on a computer, the computer executes the above-mentioned related method steps to implement an injection molded part detection method provided by the above embodiment.

[0063] An embodiment of the present application further provides a computer program product. When the computer program product is run on a computer, the computer is caused to execute the above-mentioned related steps to implement an injection molded part detection method provided in the above embodiment.

[0064] Among them, the system, computer-readable storage medium or computer program product provided in the embodiments of the present application are all used to execute the corresponding methods provided above. Therefore, the beneficial effects that can be achieved can refer to the beneficial effects in the corresponding methods provided above, and will not be repeated here.

[0065] It should be noted that the order of the above embodiments of the present application is only for description and does not represent the advantages or disadvantages of the embodiments.

[0066] The processes depicted in the accompanying drawings do not necessarily require the specific order or sequential order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous. The various embodiments in this specification are described in a progressive manner. Similar parts between the various embodiments can be referenced to each other. Each embodiment focuses on the differences from other embodiments.

Claims

1. A method for advanced geological prediction for shield machine construction, characterized in that: include: Receive a reflected wave signal using a detector device, wherein the reflected wave is formed when the detection wave signal emitted by the excitation device is reflected by the target formation; The detection device transmits the reflected wave signal to the data processing terminal; The data processing terminal performs analog-to-digital conversion on the reflected wave signal to obtain a digital reflected wave signal; the data processing terminal performs time-frequency domain joint denoising and signal enhancement on the digital reflected wave signal to obtain an enhanced reflected wave signal, wherein the time-frequency domain joint denoising and signal enhancement on the digital reflected wave signal includes: performing empirical mode decomposition and adaptive filtering based on feature distillation on the digital reflected wave signal; the data processing terminal extracts reflected wave signal features from the enhanced reflected wave signal, and determines the geological conditions ahead based on the reflected wave signal features to achieve advanced prediction; The data processing terminal performs time-frequency domain joint denoising and signal enhancement on the digital reflected wave signal to obtain an enhanced reflected wave signal, including: performing empirical mode decomposition on the digital reflected wave signal to obtain a set of signal intrinsic modal components; performing feature distillation-based adaptive filtering on the set of signal intrinsic modal components to obtain a set of filtered signal intrinsic modal components; and performing signal reconstruction on the set of filtered signal intrinsic modal components to obtain the enhanced reflected wave signal.

2. The method for advanced geological prediction for shield machine construction according to claim 1, characterized in that: The data processing terminal performs analog-to-digital conversion on the reflected wave signal to obtain a digital reflected wave signal, including: converting the reflected wave signal into a digital signal to obtain an initial reflected wave digital signal; and performing preliminary filtering on the initial reflected wave digital signal using a bandpass filter to obtain the digital reflected wave signal.

3. The method for advanced geological prediction for shield machine construction according to claim 2, characterized in that: Performing feature distillation-based adaptive filtering on the set of signal intrinsic modal components to obtain a set of filtered signal intrinsic modal components, comprising: extracting signal intrinsic modal components to be distilled from the set of signal intrinsic modal components; Based on the intrinsic modal component of the signal to be distilled and the set of the signal intrinsic modal components, a signal intrinsic modal feature distillation strength metric value is determined; based on the signal intrinsic modal feature distillation strength metric value, it is determined whether to distill the signal intrinsic modal component to be distilled.

4. The method for advanced geological prediction for shield machine construction according to claim 3, characterized in that: Based on the signal intrinsic modal components to be distilled and the set of signal intrinsic modal components, a signal intrinsic modal feature distillation strength measure value is determined, including: inputting the set of signal intrinsic modal components into a priori feature extraction self-organizing network to obtain a signal intrinsic modal prior feature aggregation coding vector; inputting the signal intrinsic modal components to be distilled and the signal intrinsic modal prior feature aggregation coding vector into a feature distillation guided network to obtain a signal intrinsic modal component distillation potential representation matrix; performing context-aware gain on the signal intrinsic modal component distillation potential representation matrix to obtain a signal intrinsic modal component context-aware gain mapping matrix; and determining the signal intrinsic modal feature distillation strength measure value based on the signal intrinsic modal component context-aware gain mapping matrix.

5. The method for advanced geological prediction for shield machine construction according to claim 4, characterized in that: Inputting the signal intrinsic modal component to be distilled and the signal intrinsic modal prior feature aggregation coding vector into a feature distillation guided network to obtain a signal intrinsic modal component distillation potential representation matrix, including: multiplying the signal intrinsic modal component to be distilled and the transposed vector of the signal intrinsic modal prior feature aggregation coding vector, and then dividing by the length of the signal intrinsic modal prior feature aggregation coding vector to obtain the signal intrinsic modal component distillation potential representation matrix.

6. The method for advanced geological prediction for shield machine construction according to claim 5, characterized in that: A context-aware gain is performed on the signal intrinsic modal component distillation potential representation matrix to obtain a signal intrinsic modal component context-aware gain mapping matrix, including: calculating the spatial proximity similarity factor of the signal intrinsic modal component to be distilled and the signal intrinsic modal prior feature aggregate coding vector; calculating the distribution proximity representation factor of the signal intrinsic modal component to be distilled and the signal intrinsic modal prior feature aggregate coding vector; and constructing a signal intrinsic modal component context-aware gain mapping matrix by combining the spatial proximity similarity factor, the distribution proximity representation factor and the signal intrinsic modal component distillation potential representation matrix.

7. The method for advanced geological prediction for shield machine construction according to claim 6, characterized in that: Based on the signal intrinsic modal component context-aware gain mapping matrix, determining a signal intrinsic modal feature distillation strength metric value includes: calculating a trace value of the signal intrinsic modal component context-aware gain mapping matrix as the signal intrinsic modal feature distillation strength metric value.

8. The method for advanced geological prediction for shield machine construction according to claim 7, characterized in that: The data processing terminal extracts the reflected wave signal feature from the enhanced reflected wave signal, including: extracting the signal feature of the enhanced reflected wave signal by using wavelet transform to obtain the reflected wave signal feature.

9. An advanced geological prediction system for shield machine construction, used to execute the advanced geological prediction method for shield machine construction according to any one of claims 1 to 8, characterized in that: include: a reflected wave signal receiving module, configured to receive a reflected wave signal by using a detector device, wherein the reflected wave is formed by the detection wave signal emitted by the excitation device and reflected by the target stratum; and a reflected wave signal transmission module, configured to transmit the reflected wave signal to a data processing terminal by the detector device; an analog-to-digital conversion module, configured for a data processing terminal to perform analog-to-digital conversion on the reflected wave signal to obtain a digital reflected wave signal; a reflected wave signal enhancement module, configured for a data processing terminal to perform joint time-frequency domain denoising and signal enhancement on the digital reflected wave signal to obtain an enhanced reflected wave signal, wherein the joint time-frequency domain denoising and signal enhancement on the digital reflected wave signal includes: performing empirical mode decomposition and adaptive filtering based on feature distillation on the digital reflected wave signal; and a geological condition judgment module, configured for a data processing terminal to extract reflected wave signal features from the enhanced reflected wave signal, and judge the geological conditions ahead based on the reflected wave signal features to achieve advanced prediction.

Images