Blasting vibration waveform feature extraction and damage range intelligent calibration method and system

Abstract

The application provides a blasting vibration waveform feature extraction and damage range intelligent calibration system. In the application, a waveform irregularity sequence is calculated based on original vibration waveform data; then, a structure surface strain vibration wave is collected by a fiber grating sensor network, and the spatial distribution information of the strain change rate is mapped to a feature enhancement channel; then, a coding and decoding structure is used to learn a feature distribution pattern based on the processed waveform data of the feature enhancement channel, and a reconstruction operation is performed on the current blasting vibration waveform accordingly; by comparing the actual waveform irregularity sequence with the reconstructed sequence, an amplified abnormal feature difference is identified, and a reconstructed error peak value distribution is generated; finally, when the strain gradient mutation point position detected by the fiber sensor network and the reconstructed error peak point space coincide, a blasting damage area heat map is generated based on the coincidence point space density. The application improves the accuracy and reliability of blasting damage range calibration.

Description

Technical Field

[0001] This application relates to the field of feature extraction and damage range calibration technology, and in particular to a method and system for feature extraction and intelligent calibration of blasting vibration waveforms. Background Technology

[0002] In engineering scenarios such as open-pit mine blasting and tunnel excavation, accurate determination of the damage range of blasting vibrations to adjacent rock masses and building structures is a core requirement for safety control. Due to the high randomness and complex propagation path of blasting vibrations, and the concealed nature of rock mass damage, it is urgent to achieve millimeter-level spatial positioning of damage boundaries through multi-source dynamic data collaborative analysis, while eliminating misjudgments caused by geological structural interference.

[0003] The current mainstream approach uses a distributed accelerometer network to collect vibration waveforms, extracts energy accumulation characteristics through time-frequency transformation, constructs a vibration energy contour map based on a preset damage threshold, and infers the damage boundary based on abrupt changes in the energy attenuation gradient. This approach utilizes the physical correlation between high-frequency vibration energy and structural damage to achieve semi-automatic calibration of the damage range.

[0004] However, the existing scheme has inherent defects: the vibration energy characteristics are easily affected by the fluctuation of blasting parameters and the reflection waves from rock fissures, resulting in the appearance of false attenuation bands in the energy contour map in the weak interlayer area; at the same time, there is a lack of real-time verification of the strain state of the structural surface. When there is a spatial offset between the sensor deployment position and the potential damage area, boundary positioning errors will occur, which cannot meet the requirements of high-precision support design. Summary of the Invention

[0005] This application provides a method and system for extracting blasting vibration waveform features and intelligently calibrating damage range, in order to solve the problem of insufficient accuracy and reliability in the calibration of blasting damage range in the prior art.

[0006] Firstly, this application provides a method for extracting blasting vibration waveform features and intelligently calibrating damage range, including:

[0007] During the blasting operation, based on the original vibration waveform data collected from the blasting vibration monitoring points, a waveform irregularity sequence characterizing the irregular changes in waveform morphology is calculated.

[0008] By deploying a fiber optic grating sensor network on the surface of the blasting target area, strain vibration waves of the structural surface caused by the blast are collected in real time, and the spatial distribution information of the strain change rate in the strain vibration waves is mapped into a feature enhancement channel to enhance the information of the waveform irregularity sequence.

[0009] Using the encoding and decoding structure, based on the waveform irregularity sequence data processed by the feature enhancement channel, the characteristic distribution pattern of the blasting vibration waveform under non-damage conditions is learned, and the vibration waveform data generated by the current blasting is reconstructed based on the characteristic distribution pattern.

[0010] By comparing the actual waveform irregularity sequence of the blasting vibration waveform with the corresponding sequence obtained by reconstruction operation, the characteristic differences of the abnormal vibration area are identified and amplified, and a reconstruction error peak distribution reflecting the degree of waveform abnormality is generated.

[0011] When the location of the abrupt change point of the strain gradient on the structural surface detected by the fiber Bragg grating sensor network coincides spatially with the location of the significant peak point in the peak distribution of the reconstruction error, the association rule engine is triggered to generate a thermal map of the blast damage area based on the spatial density of the coincident point. The thermal map of the blast damage area defines the spatial range of damage caused by the blast to the target structure.

[0012] Optionally, the step of acquiring strain vibration waves of the structural surface caused by the blast in real time through a fiber optic grating sensor network deployed on the surface of the blasting target area, and mapping the spatial distribution information of the strain change rate in the strain vibration waves to a feature enhancement channel for enhancing the information of the waveform irregularity sequence, includes:

[0013] The spatial grid cells of the fiber optic grating sensor network are used to collect the strain vibration waves on the surface of the structure caused by the blast in real time, and the rise and fall rate of the waveform of the strain vibration waves on the surface of the structure is extracted.

[0014] The waveform rise and fall rate of change is time-aligned with the corresponding time node of the waveform irregularity sequence according to the position information of the spatial grid cell to establish a mapping relationship;

[0015] Based on the mapping relationship, the rise and fall rate of the waveform is converted into a local gain factor of the waveform irregularity sequence, and the feature enhancement channel covering the entire time axis is formed by superimposing the local gain factor.

[0016] Optionally, when the location of the abrupt change in strain gradient on the structural surface detected by the fiber Bragg grating sensor network spatially coincides with the location of a significant peak point in the peak distribution of the reconstruction error, an association rule engine is triggered to generate a thermal map of the blast damage region based on the spatial density of the overlapping points. The thermal map of the blast damage region defines the spatial range of damage caused by the blast to the target structure, including:

[0017] Obtain the set of coordinates of the locations of abrupt changes in strain gradient on the structural surface detected by the fiber Bragg grating sensor network;

[0018] Establish a spatial proximity matching rule between the coordinate set and the significant peak point positions in the reconstruction error peak distribution, and mark two points as a matching point pair when the distance between them is less than a preset proximity distance threshold.

[0019] The distribution density values ​​of all matching point pairs on the surface of the blasting target area are statistically analyzed. The distribution density values ​​are converted into heat map color level values ​​and filled into the corresponding spatial positions to generate a heat map of the blasting damage area. The heat map of the blasting damage area defines the spatial range of damage caused by the blast to the target structure.

[0020] Optionally, the step of using an encoding / decoding structure to learn the characteristic distribution pattern of blasting vibration waveforms under non-damaging conditions based on the waveform irregularity sequence data processed by the feature enhancement channel, and reconstructing the vibration waveform data generated by the current blasting based on the characteristic distribution pattern, includes:

[0021] The waveform irregularity sequence data after the feature enhancement channel is processed is divided into multiple undamaged state sample blocks according to the time sequence of historical blasting events.

[0022] The undamaged sample block is processed by the compression feature extraction layer of the encoding and decoding structure to generate a feature distribution pattern library that represents the essential laws of the waveform.

[0023] The current blasting vibration waveform data is input into the recovery feature reconstruction layer of the encoding and decoding structure for reconstruction operation, and then the reconstructed waveform sequence is output based on the feature distribution pattern library.

[0024] Optionally, the step of identifying and amplifying the characteristic differences of abnormal vibration regions by comparing the actual waveform irregularity sequence of the blasting vibration waveform with the corresponding sequence obtained by reconstruction operation, and generating a reconstruction error peak distribution reflecting the degree of waveform abnormality, includes:

[0025] The original difference between the actual waveform irregularity sequence and the reconstructed waveform sequence of the blasting vibration waveform is calculated at each time node.

[0026] Within a preset continuous time interval, the original difference values ​​are calculated in a directional manner to generate an interval difference accumulation amount;

[0027] By selecting prominent points in the cumulative difference between the intervals that exceed the average level of adjacent regions, a peak distribution of reconstruction error reflecting the degree of waveform anomaly is constructed.

[0028] Optionally, the step of converting the waveform's rise and fall rate of change into a local gain factor of the waveform irregularity sequence based on the mapping relationship, and forming a feature enhancement channel covering the entire time axis by superimposing the local gain factors, includes:

[0029] According to the mapping relationship, the waveform rise and fall rate of change of each spatial grid cell is input into the proportional converter, and the waveform gain scaling factor corresponding to the time node is output.

[0030] The waveform gain scaling factor is applied to the corresponding time node of the waveform irregularity sequence, and the waveform irregularity value at the corresponding time node is amplified by multiplication to generate a local gain factor.

[0031] For all time points, the local gain factor is accumulated along the time axis to form a time-continuous characteristic enhancement channel covering the entire blasting vibration process.

[0032] Optionally, establishing a spatial proximity matching rule between the coordinate set and the locations of significant peak points in the reconstruction error peak distribution, marking two points as a matching point pair when the distance between them is less than a preset proximity distance threshold, includes:

[0033] Read the first set of spatial location points of the coordinate set, and at the same time read the second set of spatial location points of significant peak points in the peak distribution of the reconstruction error;

[0034] Traverse each point in the first set of spatial points and calculate its planar straight-line distance from all points in the second set of spatial points;

[0035] When the straight-line distance in the plane is less than the preset neighbor distance value, the corresponding strain gradient abrupt change point and the reconstruction error peak point are marked as a set of matching point pairs.

[0036] Secondly, this application provides a system for extracting blasting vibration waveform features and intelligently calibrating damage range, comprising:

[0037] The calculation module is used to calculate the waveform irregularity sequence, which characterizes the irregular changes in waveform morphology, based on the original vibration waveform data collected from the blasting vibration monitoring points during the blasting engineering process.

[0038] The acquisition module is used to acquire strain vibration waves on the surface of the structure caused by the blast in real time through a fiber optic grating sensor network deployed on the surface of the blasting target area, and to map the spatial distribution information of the strain change rate in the strain vibration waves into a feature enhancement channel for information enhancement of the waveform irregularity sequence.

[0039] The learning module is used to learn the characteristic distribution pattern of the blasting vibration waveform under non-damage conditions based on the waveform irregularity sequence data processed by the feature enhancement channel using the encoding and decoding structure, and to perform reconstruction operation on the vibration waveform data generated by the current blasting based on the characteristic distribution pattern.

[0040] The identification module is used to identify and amplify the characteristic differences of the abnormal vibration area by comparing the actual waveform irregularity sequence of the blasting vibration waveform with the corresponding sequence obtained by reconstruction operation, and to generate a reconstruction error peak distribution that reflects the degree of waveform abnormality.

[0041] The generation module is used to trigger the association rule engine when the location of the abrupt change point of the strain gradient on the structural surface detected by the fiber Bragg grating sensor network coincides spatially with the location of the significant peak point in the peak distribution of the reconstruction error. Based on the spatial density of the coincident point, the module generates a thermal map of the blast damage area, which defines the spatial range of damage caused by the blast to the target structure.

[0042] Thirdly, this application provides a computing device, including a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are to be invoked and executed by the processing component to realize the method for extracting blasting vibration waveform features and intelligently calibrating damage range as described in the first aspect above.

[0043] Fourthly, this application provides a computer storage medium storing a computer program, which, when executed by a computer, implements a method for extracting blasting vibration waveform features and intelligently calibrating damage range as described in the first aspect.

[0044] In this embodiment, during the blasting operation, based on the original vibration waveform data collected from blasting vibration monitoring points, a waveform irregularity sequence characterizing irregular waveform changes is calculated. A fiber optic grating sensor network deployed on the surface of the blasting target area is used to collect strain vibration waves on the structural surface caused by the blast in real time. The spatial distribution information of the strain rate of change in the strain vibration waves is mapped to a feature enhancement channel that enhances the waveform irregularity sequence. Using an encoding / decoding structure, based on the waveform irregularity sequence data processed by the feature enhancement channel, the characteristic distribution pattern of the blasting vibration waveform under non-damage conditions is learned. The feature distribution pattern reconstructs the vibration waveform data generated by the current blast; by comparing the actual waveform irregularity sequence of the blast vibration waveform with the corresponding sequence obtained from the reconstruction operation, the feature differences of the abnormal vibration area are identified and amplified, and a reconstruction error peak distribution reflecting the degree of waveform abnormality is generated; when the location of the abrupt change point of the strain gradient on the structural surface detected by the fiber optic grating sensor network coincides spatially with the location of the significant peak point in the reconstruction error peak distribution, the association rule engine is triggered, and a blast damage area heat map is generated based on the spatial density of the coincident point. The blast damage area heat map defines the spatial range of damage caused by the blast to the target structure.

[0045] The technical solution of this application has the following beneficial effects:

[0046] This study quantifies the dynamic fluctuation characteristics of blasting vibration energy release, providing highly sensitive waveform morphology indicators for anomaly detection. It achieves physical coupling between strain gradient changes and irregular vibration waveform characteristics, enhancing the engineering characterization capability of waveform complexity. A baseline library of waveform features under non-damaged conditions is established, and background noise interference is removed through reconstruction operations. Local differences between actual waveforms and baseline features are amplified to accurately locate abnormal vibration energy release areas. Through spatial dual verification using strain abrupt changes and waveform anomalies, misjudgments based on single-source data are eliminated, achieving millimeter-level calibration of damage boundaries.

[0047] Furthermore, strain vibration waves on the structural surface are acquired through spatial grid cells of a fiber optic grating sensor network, and the rise and fall rates of change of the waveform are extracted. Based on the mapping relationship between the spatial grid position and the waveform time node, the rate of change is converted into a local gain factor of the waveform irregularity sequence. The gain factor is superimposed along the time axis to form a feature enhancement channel. The dynamic change characteristics of the structural surface strain are embedded into waveform complexity analysis to eliminate feature distortion caused by sensor placement deviations and improve the spatial resolution of abnormal areas.

[0048] These or other aspects of this application will become more apparent in the following description of the embodiments. Attached Figure Description

[0049] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0050] Figure 1 The flowchart of a method for extracting blasting vibration waveform features and intelligently calibrating damage range provided in this application is shown.

[0051] Figure 2 The illustration shows a scenario diagram of a method for extracting blasting vibration waveform features and intelligently calibrating damage range provided in this application.

[0052] Figure 3 This paper presents a schematic diagram of the structure of an intelligent calibration system for blasting vibration waveform feature extraction and damage range determination provided in this application.

[0053] Figure 4 A schematic diagram of the structure of a computing device provided in this application is shown. Detailed Implementation

[0054] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings.

[0055] In some of the processes described in the specification, claims, and accompanying drawings of this application, multiple operations appearing in a specific order are included. However, it should be clearly understood that these operations may not be executed in the order they appear herein, or may be executed in parallel. The operation numbers, such as 101, 102, etc., are merely used to distinguish different operations and do not themselves represent any execution order. Furthermore, these processes may include more or fewer operations, and these operations may be executed sequentially or in parallel. It should be noted that the descriptions such as "first," "second," etc., in this document are used to distinguish different messages, devices, modules, etc., and do not represent a chronological order, nor do they limit "first" and "second" to different types.

[0056] Existing blasting damage calibration schemes rely on single vibration energy characteristic analysis. Their inherent flaws lie in two aspects: vibration energy contour maps are susceptible to interference from abrupt changes in blasting parameters and reflected waves from rock strata, forming pseudo-attenuation zones in weak interlayer regions; simultaneously, due to the lack of a dynamic verification mechanism for structural surface strain, when the placement of accelerometers deviates from the actual damage area, a systematic spatial offset occurs between the energy gradient abrupt change point and the true damage boundary, leading to positioning errors. Both of these flaws stem from the lack of multi-source collaborative verification of physical damage, failing to meet the requirements of millimeter-level support design.

[0057] To address the aforementioned shortcomings, this invention proposes a dual-source intelligent calibration method that integrates waveform and strain. Its core lies in fusing the complexity characteristics of vibration waveforms with the dynamic response of strain gradients on the structural surface. By mapping the spatial distribution of strain rate of change as an enhanced channel for waveform complexity characteristics, a joint strain-waveform characterization system is constructed. Then, abnormal waveforms are reconstructed based on the non-damage characteristic distribution pattern, amplifying characteristic differences. Finally, a damage thermogram is generated through the spatial overlap rule between strain gradient abrupt change points and waveform reconstruction error peaks. This method innovatively uses the strain physical response as a dynamic modulation factor and spatial verification benchmark for vibration characteristic analysis—eliminating false attenuation misjudgments caused by geological interference and correcting positioning errors caused by sensor placement offsets. This improves the accuracy of damage boundary calibration to the millimeter level, providing technical support for high-precision safety support.

[0058] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0059] Figure 1 This application provides a flowchart of a method for extracting blasting vibration waveform features and intelligently calibrating damage range, as shown in the embodiments. Figure 1 As shown, the method includes:

[0060] 101. During the implementation of blasting engineering, based on the original vibration waveform data collected from blasting vibration monitoring points, calculate the waveform irregularity sequence that characterizes the irregular changes in waveform morphology;

[0061] In the above scheme, the original vibration waveform data is a continuous record of the vibration signal collected from the blasting vibration monitoring point over time. Irregular changes in waveform morphology refer to irregular fluctuations such as non-periodic undulations, steep rises and falls, etc., that appear in the waveform curve. The waveform irregularity sequence is a set of values ​​arranged in chronological order, with each value representing the degree of distortion of the waveform curve at the corresponding time point.

[0062] In this embodiment, firstly, the original vibration signal generated by the blasting is acquired in real time by a vibration sensor installed on the rock surface. This signal is a continuous waveform curve that varies with time; for example, the sensor records the vibration amplitude value every millisecond, forming a waveform data stream lasting several seconds. Secondly, the complete waveform is divided into multiple analysis units according to fixed time segments, and the distortion degree of the waveform curve within each unit is calculated: multiple reference points are selected at equal intervals within the unit; the absolute value of the waveform height difference between each reference point and its adjacent points is calculated; the average of all the absolute values ​​of height differences is used to obtain the waveform irregularity value of that unit. Finally, the waveform irregularity values ​​of all units are combined into a sequence in chronological order. This sequence dynamically reflects the intensity of fluctuations in the release of blasting energy.

[0063] 102. By deploying a fiber optic grating sensor network on the surface of the blasting target area, the strain vibration wave of the structural surface caused by the blast is collected in real time, and the spatial distribution information of the strain change rate in the strain vibration wave is mapped into a feature enhancement channel for information enhancement of the waveform irregularity sequence.

[0064] Optionally, step 102 may specifically include the following steps:

[0065] 1021. Through the spatial grid cells of the fiber optic grating sensor network, the strain vibration wave of the structural surface caused by the blast is collected in real time, and the rise and fall rate of the waveform of the strain vibration wave of the structural surface is extracted.

[0066] 1022. The rise and fall rate of the waveform is aligned with the corresponding time nodes of the waveform irregularity sequence according to the position information of the spatial grid cell to establish a mapping relationship;

[0067] 1023. Based on the mapping relationship, the rise and fall rate of the waveform is converted into a local gain factor of the waveform irregularity sequence, and the feature enhancement channel covering the entire time axis is formed by superimposing the local gain factor.

[0068] Specifically, step 1023 may include the following process: according to the mapping relationship, the waveform rise and fall rate of change of each spatial grid unit is input into a proportional converter, and the waveform gain scaling factor corresponding to the time node is output; the waveform gain scaling factor is applied to the corresponding time node of the waveform irregularity sequence, and the waveform irregularity value of the corresponding time node is amplified by multiplication operation to generate a local gain factor; for all time nodes, the local gain factor is accumulated along the time axis to form a time-continuous feature enhancement channel covering the complete blasting vibration process.

[0069] In the above scheme, the spatial grid cell divides the surface of the blasting target area into equal-area monitoring blocks, with a fiber optic sensor probe deployed at the center of each block. The waveform rise and fall rate of change is the numerical change in strain value per unit time from trough to peak (rise) or peak to trough (fall). The mapping relationship is the rule that establishes the correspondence between the spatial location of the sensor and the time nodes of the vibration waveform. The local gain factor is a waveform characteristic adjustment coefficient generated based on the intensity of strain change. The feature enhancement channel is a sequence of gain coefficients arranged continuously along the blasting duration.

[0070] In this embodiment of the application, firstly through 1021:

[0071] A fiber optic sensor array is deployed on the surface of the blasting target at a preset grid density. Each sensor continuously records the waveform of the strain value on the structural surface as a function of time. The difference between adjacent maxima (peaks) and minima (troughs) in the waveform is extracted and divided by the time interval between the two points to obtain the rate of change of the waveform during that period. For example, if a sensor records that the strain value increases from 0.001 to 0.004 in 0.05 seconds, the rate of change is (0.004 - 0.001) / 0.05 = 0.06.

[0072] Subsequently, based on the physical laws of blast vibration wave propagation, the theoretical time for the vibration wave to travel from the blast source to each sensor grid position is calculated using 1022. This time is then aligned with the time axis of the waveform irregularity sequence to establish a mapping relationship between "sensor position → waveform time point". For example, for a sensor 20 meters from the blast source, with a vibration wave speed of 2000 m / s, the corresponding time point is 20 / 2000 = 0.01 seconds.

[0073] Finally, the waveform rise and fall rate values ​​of each sensor are input to a proportional converter via a 1023 converter. When the rate of change exceeds a reference value of 0.03, it is converted into a gain coefficient according to a linear rule. This coefficient is multiplied by the original waveform irregularity value at the corresponding time point to generate a local gain factor. The gain factors of all time points are accumulated sequentially along the time axis to form a continuous feature enhancement channel. For example, the original waveform irregularity value at time point 0.01 is 1.0 × gain coefficient 1.2 = 1.2.

[0074] In practical applications, assuming a scenario of detecting blasting damage on an iron ore slope, 16 sensors are deployed on the slope surface in a 2m × 2m grid, as per step 1021. After the blast, sensor C3 records the strain waveform: it rises from a trough of 0.0005 to a peak of 0.0035 within 0.02 seconds, with a rate of change of (0.0035 - 0.0005) / 0.02 = 0.15.

[0075] In step 1022, sensor C3 is 35 meters from the blast source, the rock wave velocity is 2500 m / s, and the corresponding time point is 35 / 2500 = 0.014 seconds. Establish a mapping from "C3 position → waveform sequence 14 milliseconds".

[0076] Using 1023, the rate of change is 0.15 > the baseline value of 0.03, and the gain coefficient = 1.0 + 0.15 × 1.5 = 1.225, which is the linear transformation rule; the irregularity value of the original waveform at the 14th millisecond is 0.9 × 1.225 = 1.1025, which is the local gain factor; the gain factors at all time points are superimposed to form a channel, and the channel value suddenly increases to 1.1025 at the 14th millisecond. This sudden increase point matches the spatial location of the abnormal area detected in subsequent steps, accurately locating the crack development zone inside the slope.

[0077] The overall scheme of step 102 above, through spatial gridded strain monitoring and precise time matching of vibration waveforms, transforms the physical deformation characteristics of the structural surface into waveform analysis modulation parameters, significantly enhancing the waveform characteristic response intensity of the hidden damage area, and providing a high-sensitivity input for subsequent spatial positioning.

[0078] 103. Using the encoding and decoding structure, based on the waveform irregularity sequence data processed by the feature enhancement channel, learn the characteristic distribution pattern of the blasting vibration waveform under non-damage conditions, and perform reconstruction operation on the vibration waveform data generated by the current blasting based on the characteristic distribution pattern;

[0079] Optionally, step 103 may specifically include the following steps:

[0080] 1031. The waveform irregularity sequence data processed by the feature enhancement channel is divided into multiple undamaged state sample blocks according to the time sequence of historical blasting events;

[0081] 1032. The undamaged sample block is processed by the compression feature extraction layer of the encoding and decoding structure to generate a feature distribution pattern library that represents the essential laws of the waveform;

[0082] 1033. Input the current blasting vibration waveform data into the recovery feature reconstruction layer of the encoding and decoding structure for reconstruction operation, and then output the reconstructed waveform sequence based on the feature distribution pattern library.

[0083] In the above scheme, the undamaged state sample block is waveform data segments corresponding to periods when the structure was undamaged, selected from historical blasting events. The compressed feature extraction layer is the processor in the encoding structure that removes random noise and extracts the core change features of the waveform. The feature distribution pattern library is a database that stores the core change patterns of the waveform, including the fluctuation intensity range and periodic features. The restored feature reconstruction layer is the processor in the decoding structure that reconstructs the theoretical waveform based on the core features.

[0084] In this embodiment, step 1031 first filters events from the historical blasting database that have been manually inspected and confirmed to be structurally undamaged, and extracts waveform data for a specific time period within each event. The continuous waveform is then cut into equal-length data blocks of fixed duration, each data block containing a complete fluctuation cycle. For example, after filtering 50 undamaged blasts, 0.8 seconds of waveform data are extracted for each blast, and then cut into 4 data blocks of 0.2 seconds each.

[0085] Next, in step 1032, the data block is input into the compression layer of the encoding structure. This layer eliminates random fluctuations by merging adjacent data points: In the first stage, every 10 original data points are merged into 1 feature point, retaining the central trend value of the data set; in the second stage, the fluctuation pattern of the feature point sequence is analyzed, the positional relationship of peaks and troughs and the range of change intensity are recorded, and feature entries are formed and stored in the pattern library. For example, 200 original points are compressed into 20 feature points, recording "peak interval 0.05 seconds ± 0.01 seconds, intensity 0.7-1.2".

[0086] Finally, in step 1033, the current burst waveform is divided into data blocks of equal duration and input into the reconstruction layer of the decoding structure. This layer searches the feature distribution pattern library for the feature entry that best matches the current block and reconstructs the theoretical waveform sequence based on the fluctuation pattern recorded in the entry. For example, if the current block finds the entry "peak interval 0.05 seconds", a reconstructed sequence of equally spaced peaks is generated according to this pattern.

[0087] In practical applications, during deep-hole blasting rock mass damage detection in gold mines, vibration monitoring data from 20 historical blasts are retrieved via step 1031. A time interval of 0.5-1.3 seconds following each blast is selected, and borehole video is used to confirm the integrity of the rock mass. Each 0.8-second waveform segment is cut at 0.1-second intervals to generate 160 sample blocks, i.e., 20 blasts × 8 blocks.

[0088] Step 1032 inputs the sample block into the compression layer for processing: every 10 milliseconds of raw data (10 points) is merged into 1 feature point, and the median value is taken; the feature point sequence is analyzed: the peak position interval is identified as 0.12 seconds ± 0.02, and the average intensity of the trough is 0.3; 120 feature entries such as "interval 0.12 seconds, intensity 0.3 ± 0.1" are stored in the pattern library.

[0089] In step 1033, the data block at 0.7 seconds of the current blast {original sequence: [0.4, 1.2, 0.3, 1.5...]} is input into the reconstruction layer: matching the entry "interval 0.12 seconds, intensity 0.3-0.5" in the pattern library; the sequence is reconstructed according to the 0.12-second peak interval: [0.35, 0.38, 0.85, 0.32, 0.82...]; anomaly development: the original sequence suddenly increases to 1.5 at 0.72 seconds, the corresponding point of the reconstructed sequence is 0.82, the difference is 0.68, and the historical average difference is <0.2.

[0090] The overall scheme of step 103 above establishes a benchmark library by extracting the essential laws of historical undamaged waveforms, and removes random interference components in the reconstruction operation, so that abnormal fluctuations are significantly highlighted in the comparison, thereby improving the waveform feature recognition capability of hidden rock mass damage.

[0091] 104. By comparing the actual waveform irregularity sequence of the blasting vibration waveform with the corresponding sequence obtained by reconstruction operation, the characteristic differences of the abnormal vibration area are identified and amplified, and a reconstruction error peak distribution reflecting the degree of waveform abnormality is generated.

[0092] Optionally, step 104 may specifically include the following steps:

[0093] 1041. Calculate the original difference between the actual waveform irregularity sequence and the reconstructed waveform sequence of the blasting vibration waveform at each time node;

[0094] 1042. Within a preset continuous time interval, perform directional cumulative calculation on the original difference values ​​to generate an interval difference cumulative amount;

[0095] 1043. Select the prominent points in the cumulative difference of the interval that exceed the average level of the adjacent area, and construct the peak distribution of reconstruction error that reflects the degree of waveform abnormality.

[0096] In the above scheme, the original difference value is the difference between the actual waveform value and the reconstructed waveform value at the same moment. Directional cumulative calculation involves summarizing positive and negative differences separately over consecutive time periods and then merging them. The cumulative difference over an interval is a composite value reflecting the net intensity of abnormal waveform fluctuations within that time period. The average level of adjacent regions is the arithmetic mean of the cumulative values ​​over equally long periods before and after the target time period. The peak distribution of the reconstruction error is a set of coordinates marking the significant spatiotemporal locations of abnormal fluctuations.

[0097] In this embodiment, step 1041 first strictly aligns the actual waveform sequence and the reconstructed sequence using millisecond-level timestamps. Subtraction is then performed on the values ​​at each identical time point, preserving the positive or negative sign of the calculation result to generate a signed original difference value sequence. For example, at millisecond 50, the actual value is 1.5, minus the reconstructed value 0.8, resulting in a difference value of +0.7.

[0098] Then, in step 1042, a fixed-duration analysis window is set, and three steps are performed within the window: separating positive difference values ​​(>0) and negative difference values ​​(<0); summing the positive and negative values ​​separately; and adding the positive and negative sums (algebraic sum) to generate the cumulative difference for the interval. For example, in the window, the positive value is +0.7 + 0.3 = 1.0, the negative value is -0.5, and the cumulative difference is 1.0 - 0.5 = 0.5.

[0099] Finally, step 1043 calculates the ratio of the cumulative amount of the target window to the average cumulative amount of its two adjacent windows (of the same length): take the average cumulative amount A of the front window and the average cumulative amount B of the rear window; calculate the reference baseline value = (A+B) / 2; divide the cumulative amount of the target window by the reference baseline value to obtain the ratio. When the ratio exceeds a preset threshold, mark the center time point of the window as a highlight point. For example, if the cumulative amount of the target window is 0.5, the average of the front window is 0.2, and the average of the rear window is 0.1, the ratio = 0.5 / ((0.2+0.1) / 2) = 3.33.

[0100] In practical applications, during the detection of rock mass damage from blasting on iron ore slopes, the difference between the actual waveform sequence (0.95, 1.82, 0.33, 2.05, 0.58) and the reconstructed sequence (0.85, 0.92, 0.75, 1.12, 0.73) collected at the slope monitoring point during the 0.35-0.40 second period after blasting is calculated by aligning them point by point: 0.95-0.58 at the 0.35 second. 85 = +0.10, 0.36 seconds 1.82 - 0.92 = +0.90, 0.37 seconds 0.33 - 0.75 = -0.42, 0.38 seconds 2.05 - 1.12 = +0.93, 0.39 seconds 0.58 - 0.73 = -0.15, generating the signed original difference sequence [+0.10, +0.90, -0.42, +0.93, -0.15].

[0101] In step 1042, the analysis window is set to 0.35-0.40 seconds. The positive difference values ​​(+0.10, +0.90, +0.93) and negative difference values ​​(-0.42, -0.15) are separated. The positive cumulative sum = 0.10 + 0.90 + 0.93 = 1.93, and the negative cumulative sum = -0.42 - 0.15 = -0.57. The algebraic sum is calculated as the cumulative difference between the intervals = 1.93 + (-0.57) = 1.36.

[0102] In step 1043, the average cumulative amount of the adjacent front window (0.30-0.35 seconds) is taken as 0.28, and the average cumulative amount of the rear window (0.40-0.45 seconds) is taken as 0.31. The reference baseline value is calculated as (0.28+0.31) / 2=0.295. The target window cumulative amount of 1.36 is divided by the baseline value to obtain the ratio value = 1.36 / 0.295≈4.61. When the preset threshold is 3.0, 4.61>3.0 triggers the flag, and the window center time point of 0.375 seconds is identified as the highlight point.

[0103] The overall scheme of step 104 above enhances the significance of continuous abnormal fluctuations through directional accumulation, effectively suppresses random interference by using the adjacent region comparison mechanism, and accurately constructs a feature map representing the spatiotemporal distribution of hidden damage.

[0104] 105. When the location of the abrupt change point of the strain gradient on the structural surface detected by the fiber Bragg grating sensor network coincides spatially with the location of the significant peak point in the peak distribution of the reconstruction error, the association rule engine is triggered to generate a thermal map of the blast damage area based on the spatial density of the coincident point. The thermal map of the blast damage area defines the spatial range of damage caused by the blast to the target structure.

[0105] Optionally, step 105 may specifically include the following steps:

[0106] 1051. Obtain the set of coordinates of the locations of abrupt changes in strain gradient on the structural surface detected by the fiber Bragg grating sensor network;

[0107] 1052. Establish a spatial proximity matching rule between the coordinate set and the significant peak point positions in the reconstruction error peak distribution, and mark the two points as a matching point pair when the distance between them is less than a preset proximity distance threshold;

[0108] Specifically, step 1052 may include the following process: reading the first set of spatial location points of the coordinate set, and simultaneously reading the second set of spatial location points of significant peak points in the peak distribution of the reconstruction error; traversing each location point in the first set of spatial location points and calculating its planar straight-line distance to all location points in the second set of spatial location points; when the planar straight-line distance is less than a preset neighbor distance value, marking the corresponding strain gradient abrupt change point and the reconstruction error peak point as a set of matching point pairs.

[0109] 1053. Statistically calculate the distribution density values ​​of all matching point pairs on the surface of the blasting target area, convert the distribution density values ​​into heat map color level values ​​and fill them into the corresponding spatial positions to generate a heat map of the blasting damage area. The heat map of the blasting damage area defines the spatial range of damage caused by the blast to the target structure.

[0110] In the above scheme, the strain gradient abrupt change point is the sensor location coordinate where the strain value on the structural surface changes drastically exceeding a preset threshold per unit time. The significant peak point is the spatial coordinate of an anomaly point in the reconstruction error peak distribution where the value exceeds a preset multiple of the average level of adjacent areas. The spatial proximity matching rule is a spatial location association logic based on straight-line distance. The distribution density value is the statistical number of matching point pairs per unit surface area. The heatmap color level value is a numerical encoding rule that uses color depth to represent the probability of damage.

[0111] In this embodiment, step 1051 first scans the strain rate data recorded at each monitoring point in the fiber Bragg grating sensor network, filters out sensor locations where the strain increase or decrease per unit time exceeds a preset abrupt change threshold, and extracts their three-dimensional spatial coordinates to form a set of abrupt change points. For example, sensor A's strain value increases from 0.001 to 0.005 in 0.08 seconds, with a change rate of 0.004 / second > the threshold of 0.003, and its recorded coordinates are (5.2, 3.1, 0).

[0112] Subsequently, in step 1052, a spatial matching distance threshold is set. Each coordinate point in the strain mutation point set is traversed, and its three-dimensional straight-line distance to all points in the reconstruction error peak point set is calculated. When any peak point's distance is less than the threshold, the strain mutation point and the peak point are bound as a matching pair, and their coordinates are recorded. For example, the distance between strain point (5.2, 3.1, 0) and peak point (5.3, 3.0, 0) = [(5.2-5.3)²+(3.1-3.0)²]^(1 / 2) = 0.14 meters < 1.2 meters → mark the matching pair.

[0113] Finally, in step 1053, the surface of the target area to be blasted is divided into equal-area grids, and the number of matching point pairs contained in each grid is counted as the distribution density value. According to a preset color-gradient mapping rule, the density value is converted into a color code: density 0 → white (RGB 255,255,255), density ≥ 5 → dark red (RGB 255,0,0), and intermediate values ​​are linearly interpolated to generate transition colors. The color-gradient values ​​are then filled into the corresponding grids to generate a spatial heatmap. For example: grid (5-6,3-4) contains 3 sets of matching points → density 3 → corresponding orange-red (RGB 255,165,0).

[0114] In practical applications, during the damage calibration of blasted rock walls in uranium mine chambers, 48 ​​fiber optic sensors were deployed on a 50m long rock wall surface in step 1051. After blasting, the strain change rate of 12 sensors exceeded the threshold of 0.004 / second. The coordinates of the abrupt change points were recorded, including: sensor 3 (12.3, 5.7, -0.2), sensor 7 (18.6, 4.9, -0.3), ... (10 other points), forming a set of abrupt change points containing 12 three-dimensional coordinates.

[0115] Step 1052 reconstructs the set of error peak points containing 18 coordinates, setting a proximity distance threshold of 1.0 meter; the distance between sensor 3 (12.3, 5.7, -0.2) and P5 (12.4, 5.8, -0.2) is calculated as [(0.1)² + (0.1)²]^(1 / 2) = 0.14 meters < 1.0 meter → mark as a matching pair; the distance between sensor 7 (18.6, 4.9, -0.3) and P9 (18.5, 4.8, -0.3) is calculated as √[(0.1)² + (0.1)²] = 0.14 meters < 1.0 meter → mark as a matching pair; finally, 8 sets of matching point pairs are generated.

[0116] In step 1053, the rock wall surface is divided into 1m×1m grids. Grid [12-13,5-6] contains 2 sets of matching points → density 2.0; grid [18-19,4-5] contains 3 sets of matching points → density 3.0. According to the color gradation rules: density 2.0 → yellow RGB255,255,0, density 3.0 → orange RGB 255,165,0; in the generated heat map, grid (18-19,4-5) displays orange, and grid (12-13,5-6) displays yellow.

[0117] The overall scheme in step 105 above eliminates interference from misjudgments of single-source data through a dual spatial verification mechanism of strain mutation and waveform anomaly; the heat map generated based on the matching point density intuitively presents the boundary of the damaged area, providing spatial positioning basis for the precise treatment of hidden damage.

[0118] The following is a complete embodiment for steps 101-105:

[0119] like Figure 2 As shown, in the monitoring of rock fracture propagation during deep-hole blasting in iron mines, step 101 involves setting up 8 vibration monitoring points on the sidewall of the roadway to collect the original vibration waveform 0-2 seconds after blasting. The waveform is segmented in 20-millisecond units, and the average value of the peak-to-trough height difference is calculated for each unit. For example, in the 0.05-0.07 second unit: peak 0.8mV, trough 0.2mV → height difference 0.6mV, generating a waveform irregularity sequence [0.6, 1.2, 0.9, ...] that dynamically reflects the energy release fluctuations.

[0120] In step 102, a 5×5 fiber optic sensor grid with a spacing of 1m is arranged on the top plate of the blasting area. Sensor B3 records that the strain value suddenly increases from 0.003 to 0.007 within 0.1 seconds, with a change rate of 0.04 / second. Based on the rock mass wave velocity of 2800m / s and the distance of B3 from the blast source of 22.4m, the corresponding time point of 8 milliseconds is determined. This change rate is converted into a gain coefficient of 1.3, and the original value of the modulated waveform sequence of 0.9 in the 8th millisecond is enhanced to 1.17.

[0121] Step 103 calls 30 historical non-destructive blasting data and extracts the enhanced waveform sequence of the first 0.5 seconds; the compression layer merges the data of every 100 milliseconds into 10 feature points and retains the maximum fluctuation value; the current blasting 0.3-second measured sequence [0.2, 1.5, 0.3] matches the "medium intensity fluctuation" feature of the pattern library and outputs the reconstructed sequence [0.65, 0.72, 0.68].

[0122] In step 104, the difference between the measured value of 1.5 and the reconstructed value of 0.72 in the 0.3-second time interval is +0.78; take an 80-millisecond window from 0.28 to 0.36 seconds, accumulate the positive difference of 1.92 and the negative difference of -0.57, with a cumulative amount of 1.35; the mean of adjacent windows is 0.25, the ratio is 5.4 > the threshold of 3.0, and 0.32 seconds is marked as an abnormal peak point.

[0123] In step 105, the fiber optic sensor D2 (15.3, 8.7, -0.5) detects a sudden change in strain with a rate of change of 0.05 / second; the distance between it and the reconstructed peak point P6 [15.4, 8.6, -0.5] is 0.14 meters, which is less than the 1.2-meter threshold, so a matching point pair is marked; a 2m×2m grid (15-17, 8-10) contains 3 sets of matching points → density 1.5 sets / m², which is mapped to a dark orange thermal block.

[0124] This scheme integrates the complexity characteristics of vibration waveforms with the strain gradient response of structural surfaces to construct a dual-source collaborative damage calibration mechanism: the strain change rate is converted into waveform feature gain coefficients in real time to enhance the signal sensitivity of the crack propagation area; the theoretical sequence is reconstructed based on a non-damaging waveform feature library, and abnormal fluctuations are amplified through directional accumulation; the proximity matching rule between strain abrupt change points and abnormal waveform peaks effectively suppresses the interference of geological reflection pseudo-signals; the finally generated heat map calibrates the distribution range of internal cracks in the rock mass with millimeter-level spatial resolution, guides precise grouting treatment, and significantly improves the efficiency of hidden damage prevention and control.

[0125] Figure 3 This application provides a schematic diagram of the structure of an intelligent calibration system for extracting blasting vibration waveform features and determining damage range, as shown in the embodiment of the present application. Figure 3 As shown, the system includes:

[0126] Calculation module 31 is used to calculate the waveform irregularity sequence, which characterizes the irregular changes in waveform morphology, based on the original vibration waveform data collected from the blasting vibration monitoring point during the blasting engineering process.

[0127] The acquisition module 32 is used to acquire the strain vibration wave of the structural surface caused by the blast in real time through a fiber optic grating sensor network deployed on the surface of the blasting target area, and map the spatial distribution information of the strain change rate in the strain vibration wave into a feature enhancement channel for information enhancement of the waveform irregularity sequence.

[0128] Learning module 33 is used to learn the characteristic distribution pattern of blasting vibration waveform under non-damage conditions based on the waveform irregularity sequence data processed by the feature enhancement channel using the encoding and decoding structure, and to reconstruct the vibration waveform data generated by the current blasting based on the characteristic distribution pattern.

[0129] The identification module 34 is used to identify and amplify the characteristic differences of the abnormal vibration area by comparing the actual waveform irregularity sequence of the blasting vibration waveform with the corresponding sequence obtained by reconstruction operation, and to generate a reconstruction error peak distribution that reflects the degree of waveform abnormality.

[0130] The generation module 35 is used to trigger the association rule engine when the location of the abrupt change point of the strain gradient on the structural surface detected by the fiber Bragg grating sensor network coincides spatially with the location of the significant peak point in the peak distribution of the reconstruction error. Based on the spatial density of the coincident point, the module generates a thermal map of the blast damage area, which defines the spatial range of damage caused by the blast to the target structure.

[0131] Figure 3 The aforementioned intelligent calibration system for extracting blasting vibration waveform features and determining damage range can perform... Figure 1 The implementation principle and technical effects of the blasting vibration waveform feature extraction and damage range intelligent calibration method described in the illustrated embodiment will not be repeated here. The specific operation methods of each module and unit in the blasting vibration waveform feature extraction and damage range intelligent calibration system in the above embodiments have been described in detail in the embodiments related to this method, and will not be elaborated upon here.

[0132] In one possible design, Figure 3 The blasting vibration waveform feature extraction and damage range intelligent calibration system of the embodiment shown can be implemented as a computing device, such as... Figure 4 As shown, the computing device may include a storage component 41 and a processing component 42;

[0133] The storage component 41 stores one or more computer instructions, wherein the one or more computer instructions are invoked and executed by the processing component 42.

[0134] The processing component 42 is used for the above Figure 1 The embodiment describes a method for extracting blasting vibration waveform features and intelligently calibrating damage range.

[0135] The processing component 42 may include one or more processors to execute computer instructions to complete all or part of the steps in the above-described method. Alternatively, the processing component may be implemented as one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the above-described method.

[0136] Storage component 41 is configured to store various types of data to support operations at the terminal. The storage component can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0137] Of course, computing devices may also include other components, such as input / output interfaces, display components, communication components, etc.

[0138] Input / output interfaces provide interfaces between processing components and peripheral interface modules, which can be output devices, input devices, etc.

[0139] The communication components are configured to facilitate wired or wireless communication between computing devices and other devices.

[0140] The computing device can be a physical device or an elastic computing host provided by a cloud computing platform. In this case, the computing device can refer to a cloud server, and the aforementioned processing components, storage components, etc., can be basic server resources rented or purchased from the cloud computing platform.

[0141] This application also provides a computer storage medium storing a computer program, which, when executed by a computer, can perform the above-described functions. Figure 1 The embodiment shown illustrates a method for extracting blasting vibration waveform features and intelligently calibrating damage range.

[0142] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0143] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0144] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0145] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for extracting blasting vibration waveform features and intelligently calibrating damage range, characterized in that, include: During the blasting operation, based on the original vibration waveform data collected from the blasting vibration monitoring points, a waveform irregularity sequence characterizing the irregular changes in waveform morphology is calculated. By deploying a fiber optic grating sensor network on the surface of the blasting target area, strain vibration waves of the structural surface caused by the blast are collected in real time, and the spatial distribution information of the strain change rate in the strain vibration waves is mapped into a feature enhancement channel to enhance the information of the waveform irregularity sequence. Using the encoding and decoding structure, based on the waveform irregularity sequence data processed by the feature enhancement channel, the characteristic distribution pattern of the blasting vibration waveform under non-damage conditions is learned, and the vibration waveform data generated by the current blasting is reconstructed based on the characteristic distribution pattern. By comparing the actual waveform irregularity sequence of the blasting vibration waveform with the corresponding sequence obtained by reconstruction operation, the characteristic differences of the abnormal vibration area are identified and amplified, and a reconstruction error peak distribution reflecting the degree of waveform abnormality is generated. When the location of the abrupt change point of the strain gradient on the structural surface detected by the fiber Bragg grating sensor network coincides spatially with the location of the significant peak point in the peak distribution of the reconstruction error, the association rule engine is triggered to generate a thermal map of the blast damage area based on the spatial density of the coincident point. The thermal map of the blast damage area defines the spatial range of damage caused by the blast to the target structure.

2. The method according to claim 1, characterized in that, The method involves using a fiber optic grating sensor network deployed on the surface of the blasting target area to collect strain vibration waves on the structural surface in real time caused by the blast, and mapping the spatial distribution information of the strain change rate in the strain vibration waves into a feature enhancement channel for enhancing the information of the waveform irregularity sequence, including: The spatial grid cells of the fiber optic grating sensor network are used to collect the strain vibration waves on the surface of the structure caused by the blast in real time, and the rise and fall rate of the waveform of the strain vibration waves on the surface of the structure is extracted. The waveform rise and fall rate of change is time-aligned with the corresponding time node of the waveform irregularity sequence according to the position information of the spatial grid cell to establish a mapping relationship; Based on the mapping relationship, the rise and fall rate of the waveform is converted into a local gain factor of the waveform irregularity sequence, and the feature enhancement channel covering the entire time axis is formed by superimposing the local gain factor.

3. The method according to claim 1, characterized in that, When the location of the abrupt change in strain gradient on the structural surface detected by the fiber Bragg grating sensor network spatially coincides with the location of a significant peak point in the peak distribution of the reconstruction error, an association rule engine is triggered to generate a thermal map of the blast damage region based on the spatial density of the overlapping points. This thermal map defines the spatial extent of the damage caused by the blast to the target structure, including: Obtain the set of coordinates of the locations of abrupt changes in strain gradient on the structural surface detected by the fiber Bragg grating sensor network; Establish a spatial proximity matching rule between the coordinate set and the significant peak point positions in the reconstruction error peak distribution, and mark two points as a matching point pair when the distance between them is less than a preset proximity distance threshold. The distribution density values ​​of all matching point pairs on the surface of the blasting target area are statistically analyzed. The distribution density values ​​are converted into heat map color level values ​​and filled into the corresponding spatial positions to generate a heat map of the blasting damage area. The heat map of the blasting damage area defines the spatial range of damage caused by the blast to the target structure.

4. The method according to claim 1, characterized in that, The method utilizes an encoding / decoding structure to learn the characteristic distribution pattern of blasting vibration waveforms under non-damaging conditions based on the waveform irregularity sequence data processed by the feature enhancement channel. Based on this characteristic distribution pattern, a reconstruction operation is performed on the vibration waveform data generated by the current blast, including: The waveform irregularity sequence data after the feature enhancement channel is processed is divided into multiple undamaged state sample blocks according to the time sequence of historical blasting events. The undamaged sample block is processed by the compression feature extraction layer of the encoding and decoding structure to generate a feature distribution pattern library that represents the essential laws of the waveform. The current blasting vibration waveform data is input into the recovery feature reconstruction layer of the encoding and decoding structure for reconstruction operation, and then the reconstructed waveform sequence is output based on the feature distribution pattern library.

5. The method according to claim 4, characterized in that, The step of identifying and amplifying the characteristic differences of abnormal vibration regions by comparing the actual waveform irregularity sequence of the blasting vibration waveform with the corresponding sequence obtained from the reconstruction operation, and generating a reconstruction error peak distribution reflecting the degree of waveform abnormality, includes: The original difference between the actual waveform irregularity sequence and the reconstructed waveform sequence of the blasting vibration waveform is calculated at each time node. Within a preset continuous time interval, the original difference values ​​are calculated in a directional manner to generate an interval difference accumulation amount; By selecting prominent points in the cumulative difference between the intervals that exceed the average level of adjacent regions, a peak distribution of reconstruction error reflecting the degree of waveform anomaly is constructed.

6. The method according to claim 2, characterized in that, Based on the mapping relationship, the process of converting the waveform's rise and fall rates of change into local gain factors for the waveform's irregularity sequence, and then superimposing these local gain factors to form a feature enhancement channel covering the entire time axis, includes: According to the mapping relationship, the waveform rise and fall rate of change of each spatial grid cell is input into the proportional converter, and the waveform gain scaling factor corresponding to the time node is output. The waveform gain scaling factor is applied to the corresponding time node of the waveform irregularity sequence, and the waveform irregularity value at the corresponding time node is amplified by multiplication to generate a local gain factor. For all time points, the local gain factor is accumulated along the time axis to form a time-continuous characteristic enhancement channel covering the entire blasting vibration process.

7. The method according to claim 3, characterized in that, The step of establishing a spatial proximity matching rule between the coordinate set and the significant peak points in the reconstruction error peak distribution, where two points are marked as a matching point pair when the distance between them is less than a preset proximity distance threshold, includes: Read the first set of spatial location points of the coordinate set, and at the same time read the second set of spatial location points of significant peak points in the peak distribution of the reconstruction error; Traverse each point in the first set of spatial points and calculate its planar straight-line distance from all points in the second set of spatial points; When the straight-line distance in the plane is less than the preset neighbor distance value, the corresponding strain gradient abrupt change point and the reconstruction error peak point are marked as a set of matching point pairs.

8. A system for extracting blasting vibration waveform features and intelligently calibrating damage range, characterized in that, include: The calculation module is used to calculate the waveform irregularity sequence, which characterizes the irregular changes in waveform morphology, based on the original vibration waveform data collected from the blasting vibration monitoring points during the blasting engineering process. The acquisition module is used to acquire strain vibration waves on the surface of the structure caused by the blast in real time through a fiber optic grating sensor network deployed on the surface of the blasting target area, and to map the spatial distribution information of the strain change rate in the strain vibration waves into a feature enhancement channel for information enhancement of the waveform irregularity sequence. The learning module is used to learn the characteristic distribution pattern of the blasting vibration waveform under non-damage conditions based on the waveform irregularity sequence data processed by the feature enhancement channel using the encoding and decoding structure, and to perform reconstruction operation on the vibration waveform data generated by the current blasting based on the characteristic distribution pattern. The identification module is used to identify and amplify the characteristic differences of the abnormal vibration area by comparing the actual waveform irregularity sequence of the blasting vibration waveform with the corresponding sequence obtained by reconstruction operation, and to generate a reconstruction error peak distribution that reflects the degree of waveform abnormality. The generation module is used to trigger the association rule engine when the location of the abrupt change point of the strain gradient on the structural surface detected by the fiber Bragg grating sensor network coincides spatially with the location of the significant peak point in the peak distribution of the reconstruction error. Based on the spatial density of the coincident point, the module generates a thermal map of the blast damage area, which defines the spatial range of damage caused by the blast to the target structure.

9. A computing device, characterized in that, It includes a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are invoked and executed by the processing component to implement the intelligent calibration method for blasting vibration waveform feature extraction and damage range as described in any one of claims 1 to 7.

10. A computer storage medium, characterized in that, The device contains a computer program that, when executed by a computer, implements a method for extracting blasting vibration waveform features and intelligently calibrating damage range as described in any one of claims 1 to 7.

Images