High-altitude glacier permafrost earthquake chain disaster simulation method

Abstract

The present application relates to a kind of high-altitude glacier permafrost earthquake chain disaster simulation method, the method comprises: obtaining vibration simulation demand;According to vibration simulation demand, vibration simulation is carried out by pre-trained vibration simulation model control high-altitude glacier permafrost earthquake chain disaster simulation device;Wherein, vibration simulation model is obtained according to training sample training;Wherein, training sample includes vibration data sample and geological data sample;Vibration data sample is by multiple real environment data of high-altitude area typical geological structure, seismic motion data composition;Geological data sample includes the physical characteristic data of high-altitude area typical geological structure.The present application is by multiple real environment data of high-altitude area typical geological structure, seismic motion data vibration simulation model training, vibration simulation is carried out by vibration simulation model control high-altitude glacier permafrost earthquake chain disaster simulation device, realizes the vibration of high-altitude glacier permafrost zone earthquake induced chain disaster simulation.

Description

Technical Field

[0001] This invention relates to the field of simulation technology, and in particular to a method for simulating earthquake chain disasters in high-altitude glaciers and permafrost. Background Technology

[0002] With global warming, the risk of geological disasters in glacier and permafrost regions is gradually increasing, especially in the Qinghai-Tibet Plateau and other high-altitude areas, where rising temperatures, glacier melting, and deepening of permafrost layers lead to frequent occurrences of snow and ice-related geological disasters.

[0003] Climate warming has a particularly significant impact on permafrost, especially in regions like the Qinghai-Tibet Plateau. Rising temperatures accelerate glacial melting, deepen the active permafrost layer, and speed up permafrost degradation. This change not only affects the regional ecological environment but also significantly increases the risk of earthquake-induced disasters in permafrost areas. For example, under the combined effects of earthquakes and temperature changes, disaster chains such as ice avalanches, snow avalanches, and landslides exhibit dynamic evolution characteristics, posing a significant challenge to regional safety management.

[0004] Existing vibration simulation methods are all used to simulate the dynamic response of buildings to earthquakes, but these methods are almost all only applicable to thawed soil environments. They cannot simulate the superimposed effects of the special climate environment, permafrost degradation, and frequent earthquakes in high-altitude glacial permafrost regions.

[0005] Therefore, how to develop a vibration simulation method that can simulate earthquake-induced chain disasters in high-altitude glacial and permafrost regions has become an urgent problem to be solved. Summary of the Invention

[0006] (a) Technical problems to be solved

[0007] In view of the above-mentioned shortcomings and deficiencies of the prior art, the present invention provides a method for simulating earthquake chain disasters in high-altitude glaciers and permafrost.

[0008] (II) Technical Solution

[0009] To achieve the above objectives, the main technical solutions adopted by the present invention include:

[0010] A method for simulating earthquake chain disasters in high-altitude glaciers and permafrost, the method comprising:

[0011] Obtain vibration simulation requirements;

[0012] Based on the vibration simulation requirements, a high-altitude glacier permafrost earthquake chain disaster simulation device is controlled by a pre-trained vibration simulation model to perform vibration simulation.

[0013] The vibration simulation model is trained based on training samples.

[0014] The training samples include vibration data samples and geological data samples;

[0015] The vibration data sample consists of multiple real environmental data and seismic motion data of typical geological structures in high-altitude areas.

[0016] The geological data samples include physical characteristic data of typical geological structures in high-altitude areas.

[0017] Optionally, the vibration simulation model is trained through the following steps:

[0018] The training samples are clustered according to the typical geological structures of high-altitude areas, so that each typical geological structure in high-altitude areas corresponds to a class of training samples.

[0019] The vibration simulation model was trained based on various training samples to obtain the parameters of the trained vibration simulation model for each type.

[0020] All trained vibration simulation model parameters and vibration simulation models are used as trained vibration simulation models.

[0021] Optionally, the vibration data sample includes the acquisition time, acquisition location, and weight value;

[0022] The weights of vibration data samples in any class of training samples are determined through the following steps:

[0023] Vibration data samples from any type of training samples form the first set, and each element in the first set is labeled as unprocessed;

[0024] Randomly select an element marked as unprocessed from the first set as the first sample, take the collection time of the first sample as the first time, and take the collection location of the first sample as the first location;

[0025] The elements in the first set that are not the first sample are identified as the first sample location and have the smallest time difference between the sampling time and the first time location. The time difference between the sampling time of the second sample and the first time location is taken as the first time difference.

[0026] Determine the average time difference of the acquisition time of all vibration data samples acquired at the first position in any class of training samples;

[0027] The sum of the first time difference and the average time difference is determined as the second time difference;

[0028] Elements in the first set whose time difference between the collection time and the first time is no greater than the second time difference are selected as the third sample.

[0029] The third set consists of samples whose time difference between the collection time and the first time is no greater than the first time difference; the remaining third samples are then combined into a third set.

[0030] Determine the first similarity between each element in the second set and the first sample, and determine the second similarity between each element in the third set and the first sample;

[0031] The weight values ​​of the processed samples and each processed sample are determined based on the first similarity and the second similarity. The labels of the processed samples are updated to "processed". The step of randomly selecting an element labeled as unprocessed from the first set as the first sample and subsequent steps are repeated until all elements in the first set are labeled as "processed".

[0032] Optionally, determining the processed sample based on a first similarity and a second similarity includes:

[0033] Determine the first minimum, first maximum, and first mean of all first similarities;

[0034] Determine the second minimum, second maximum, and second mean of all second similarities;

[0035] The first benchmark value is determined as (first minimum value + first maximum value + 4 × first mean value) / 6;

[0036] The second benchmark value is determined as (second minimum value + second maximum value + 4 × second mean value) / 6;

[0037] The center position of the acquisition location corresponding to the first similarity score which is not greater than the first benchmark value is determined as the first position; the center position of the acquisition location corresponding to the second similarity score which is not greater than the second benchmark value is determined as the second position;

[0038] Determine whether the distance between the first position and the second position is not greater than the maximum distance between the first position and the acquisition position corresponding to the first similarity value which is not greater than the first benchmark value;

[0039] If the distance between the first position and the second position is not greater than the maximum distance between the first position and the acquisition position corresponding to the first similarity not greater than the first benchmark value, then the first sample, as well as all vibration data samples corresponding to the first similarity not less than the first benchmark value and all vibration data samples corresponding to the second similarity not less than the second benchmark value, are all determined as processing samples.

[0040] Optionally, after determining whether the distance between the first position and the second position is not greater than the maximum distance between the first position and the acquisition position corresponding to the first similarity not greater than the first benchmark value, the method further includes:

[0041] If the distance between the first position and the second position is greater than the maximum distance between the first position and the acquisition position corresponding to the first similarity value which is not greater than the first benchmark value, then the first sample and the vibration data sample with a similarity value of not greater than the first benchmark value are both determined as processing samples.

[0042] Optionally, the weight values ​​for each processed sample are determined through the following steps:

[0043] All processed samples are grouped into a fourth set;

[0044] Calculate the third similarity between any two elements in the fourth set;

[0045] The normalized value of the difference between each third similarity and the minimum value among the third similarities is determined as the weight value.

[0046] Optionally, the high-altitude glacier permafrost earthquake chain disaster simulation device includes: an environmental control chamber and a shaking table array;

[0047] The environmental control cabin includes a low-temperature control system, a rainfall unit, a snowfall unit, a strong radiation simulation unit, a humidity control unit, and a wind speed control unit.

[0048] A shaking table array consists of multiple shaking tables;

[0049] An environmental control cabin is used to simulate extreme and complex climate environments at high altitudes.

[0050] Shaking table arrays are used to simulate seismic dynamic output.

[0051] Optionally, the cryogenic control system includes: an environmental control chamber temperature control unit and a model freeze-thaw temperature control unit;

[0052] The environmental control chamber temperature control unit uses liquid nitrogen circulation cooling technology to control the temperature. The temperature control range of the environmental control chamber temperature control unit is -40 degrees Celsius to 60 degrees Celsius, and the temperature control accuracy is ±1 degree Celsius.

[0053] The model freeze-thaw temperature control unit uses a high-power cold bath box to control a unidirectional heat transfer metal cold plate. The temperature is controlled by covering the model with the cold plate. The temperature control range of the cold plate is -40℃ to 20℃, and the temperature control accuracy is ±1℃.

[0054] The rainfall unit includes water pipes and nozzles, with a rainfall capacity of 0~200 mm / h and a rainfall uniformity >90%.

[0055] The snowmaking unit includes a snowmaking machine, which operates normally in an environment of -40℃ to 20℃, with a snowmaking capacity of 0~20mm / h and a snowmaking uniformity of >90%.

[0056] The humidity control unit uses ultrasonic humidification technology and a humidity sensor detection system for real-time feedback and adjustment. The humidity control range is 10% to 90%, and the humidity control accuracy is ±3%.

[0057] The wind speed control unit includes a high-power fan, with a wind speed control range of 0 to 8 levels and a wind speed control accuracy within ±1%.

[0058] The high-intensity radiation simulation unit includes a radiation panel integrated into the inner wall of the environmental control cabin. The panel simulates intense sunlight and ultraviolet radiation at high altitudes, with a simulation range of 0–1500 watts (W / m²). 2 The accuracy of the strong radiation simulation is ±1%.

[0059] Optionally, the shaking table is a three-axis six-degree-of-freedom shaking table, which adopts an electro-hydraulic servo actuator and loads one or more of sine waves, random waves, and seismic waves;

[0060] The maximum acceleration of the vibration table is (±2.0 g, ±2.0 g, ±1.5 g), the displacement is (±500 mm, ±500 mm, ±300 mm), and the frequency range is 0.1~60 Hz.

[0061] Optionally, the shaking table array is composed of three shaking tables integrated together;

[0062] One of the vibration tables is 6m×8m and has a full load of 100 tons, and the other two vibration tables are 3m×4m and have a full load of 40 tons.

[0063] The vibration tables in the vibration table array are integrated into a single platform; the integrated vibration table array is 8m×9m and fully loaded with 180t, with maximum acceleration of (±1.8g, ±1.8g, ±1.3g), displacement of (±500mm, ±500mm, ±300mm), and frequency range of 0.1~60Hz.

[0064] or,

[0065] The shaking tables in the shaking table array move along tracks in a linear arrangement; in the linearly arranged shaking table array, the spacing between three shaking tables is 12m to 150m, and the movement range is 138m.

[0066] (III) Beneficial Effects

[0067] The beneficial effects of this invention are as follows: This invention relates to a method for simulating earthquake chain disasters in high-altitude glaciers and permafrost regions. The method includes: obtaining vibration simulation requirements; and, based on these requirements, controlling a high-altitude glacier and permafrost earthquake chain disaster simulation device to perform vibration simulation using a pre-trained vibration simulation model. The vibration simulation model is trained based on training samples. These training samples include vibration data samples and geological data samples. The vibration data samples consist of multiple real environmental data and seismic motion data of typical geological structures in high-altitude areas. The geological data samples include physical characteristic data of typical geological structures in high-altitude areas. This invention trains a vibration simulation model using multiple real environmental data and seismic motion data of typical geological structures in high-altitude areas, and then controls a high-altitude glacier and permafrost earthquake chain disaster simulation device to perform vibration simulation, thereby simulating the vibrations of earthquake-induced chain disasters in high-altitude glacier and permafrost regions. Attached Figure Description

[0068] Figure 1 A flowchart illustrating a high-altitude glacier permafrost earthquake chain disaster simulation method provided by the present invention;

[0069] Figure 2 A schematic diagram of a high-altitude glacier permafrost earthquake chain disaster simulation device provided by the present invention;

[0070] Figure 3 A schematic diagram of a vibration table array provided by the present invention, in which the vibration table is integrated into a single platform;

[0071] Figure 4 This invention provides a schematic diagram of a vibration table array formed by linearly arranging vibration tables. Detailed Implementation

[0072] To better explain and facilitate understanding of the present invention, the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0073] Existing simulation methods for earthquake-induced chain hazards in high-altitude glaciers and permafrost regions are all used to simulate the dynamic response of buildings to earthquakes. However, these methods are almost all only applicable to thawed soil environments. They cannot simulate the superimposed effects of the unique climate, permafrost degradation, and frequent earthquakes in high-altitude glacier and permafrost regions. Therefore, developing a simulation method capable of simulating earthquake-induced chain hazards in high-altitude glacier and permafrost regions has become an urgent problem to be solved.

[0074] To address this problem, this invention relates to a method for simulating earthquake chain disasters in high-altitude glaciers and permafrost regions. The method includes: obtaining vibration simulation requirements; and, based on these requirements, controlling a high-altitude glacier and permafrost earthquake chain disaster simulation device using a pre-trained vibration simulation model to perform vibration simulation. The vibration simulation model is trained using training samples, which include vibration data samples and geological data samples. The vibration data samples consist of multiple real environmental data and seismic motion data of typical geological structures in high-altitude areas. The geological data samples include physical characteristic data of typical geological structures in high-altitude areas. This invention trains a vibration simulation model using multiple real environmental data and seismic motion data of typical geological structures in high-altitude areas, and then controls a high-altitude glacier and permafrost earthquake chain disaster simulation device using the vibration simulation model to perform vibration simulation, thereby simulating the vibrations of earthquake-induced chain disasters in high-altitude glacier and permafrost regions.

[0075] See Figure 1 This embodiment provides a method for simulating earthquake chain disasters in high-altitude glaciers and permafrost. The implementation process of this method is as follows:

[0076] 101, Obtain vibration simulation requirements.

[0077] The simulated requirements are pre-inputted and can be entered by inputting a requirements document, requirements parameters, or other methods. Step 101 simply involves obtaining the input simulated requirements.

[0078] Simulation requirements include, but are not limited to: temperature, snowfall, rainfall, humidity, wind speed, and ground motion data.

[0079] 102. Based on the vibration simulation requirements, a high-altitude glacier permafrost earthquake chain disaster simulation device is controlled by a pre-trained vibration simulation model to perform vibration simulation.

[0080] The vibration simulation model is trained based on training samples.

[0081] The training samples include vibration data samples and geological data samples.

[0082] The vibration data sample consists of multiple real environmental data and seismic motion data from typical geological structures in high-altitude areas.

[0083] The geological data samples include physical characteristic data of typical geological structures in high-altitude areas.

[0084] In other words, after obtaining the vibration simulation requirements in step 101, in step 102, the vibration simulation requirements are input into the pre-trained vibration simulation model. The vibration simulation model then converts the vibration simulation requirements into control parameters, and the high-altitude glacier permafrost earthquake chain disaster simulation device is controlled to perform vibration simulation based on the control parameters. Since the simulation parameters of the high-altitude glacier permafrost earthquake chain disaster simulation device are all determined by the vibration simulation requirements, the simulated vibration is the vibration that meets the vibration simulation requirements, thus achieving accurate vibration simulation.

[0085] Furthermore, the training samples for the vibration simulation model include vibration data samples and geological data samples. The vibration data samples consist of multiple real environmental data and seismic motion data of typical geological structures in high-altitude areas. The geological data samples include physical characteristic data of typical geological structures in high-altitude areas. This enables the trained vibration simulation model to not only accurately simulate the special climatic environment of high-altitude glacial and permafrost regions, but also accurately simulate earthquakes, realizing the simulation of vibrations that induce chain reactions of earthquakes in high-altitude glacial and permafrost regions.

[0086] Specifically, the vibration simulation model is trained through the following steps:

[0087] 201. Multiple real environmental and seismic data from typical geological structures in high-altitude areas were collected to form vibration data samples, and physical characteristic data from typical geological structures in high-altitude areas were collected to form geological data samples. Training samples were formed from the vibration data samples and geological data samples.

[0088] The typical geological structures in high-altitude areas mainly include the following four types:

[0089] 1. Horizontal structure: The rock strata remain horizontal or nearly horizontal, and after being cut, they form table-shaped terraces or flat-topped mountains.

[0090] 2. Inclined structure: The rock strata form an angle with the horizontal plane, caused by folds, faults, or uneven uplift and subsidence.

[0091] 3. Fold structure: Rock strata bend under lateral compressive stress, including anticlines (convex) and synclines (concave).

[0092] 4. Fault structures: Rock fractures and significant displacement, such as normal faults (hanging wall descends), reverse faults (hanging wall rises), and strike-slip faults (horizontal displacement).

[0093] In addition, vibration data samples and geological data samples will be collected and updated regularly, and data on different geological structures, different altitudes, and different environments will be collected as comprehensively as possible so that the vibration simulation model can accurately simulate various needs.

[0094] 202. The training samples are clustered according to the typical geological structures of high-altitude areas, so that each typical geological structure in high-altitude areas corresponds to a class of training samples.

[0095] For example, typical geological structures in high-altitude areas include horizontal structures, inclined structures, fold structures, and fault structures. In step 202, training samples related to horizontal structures will be grouped into one class, training samples related to inclined structures will be grouped into another class, training samples related to fold structures will be grouped into another class, and training samples related to fault structures will be grouped into another class.

[0096] Clustering by typical geological structures in high-altitude areas allows for more targeted training on each geological structure, ensuring that the vibration simulation model can accurately simulate various geological structures.

[0097] In addition to the specific data collected (such as temperature, snowfall, rainfall, humidity, wind speed, amplitude, ground motion data, etc.), the vibration data sample also includes the collection time, collection location, and weight value.

[0098] The weight value of a vibration data sample (e.g., vibration data sample i) represents the degree of individuality of vibration data sample i within its corresponding geological structure. A larger weight value indicates a higher degree of individuality, meaning vibration data sample i is less similar to other samples within its corresponding geological structure, and is therefore considered a more unique sample with fewer similar data points, requiring more attention during training. Conversely, a smaller weight value indicates a lower degree of individuality, meaning vibration data sample i is more similar to other samples within its corresponding geological structure, and is therefore considered a more common sample with more similar data points, requiring less attention during training.

[0099] Therefore, the weights of vibration data samples are closely related to their class. The weights of vibration data samples in the training samples for any class (e.g., class j) are determined through the following steps:

[0100] 301. The vibration data samples in any type of training samples are used to form the first set, and each element in the first set is labeled as unprocessed.

[0101] For example, class j has 5 training samples: training sample 1, training sample 2, training sample 3, training sample 4, and training sample 5. The vibration data sample in training sample 1 is called vibration data sample 1, the vibration data sample in training sample 2 is called vibration data sample 2, the vibration data sample in training sample 3 is called vibration data sample 3, the vibration data sample in training sample 4 is called vibration data sample 4, and the vibration data sample in training sample 5 is called vibration data sample 5. Then the first set is {vibration data sample 1, vibration data sample 2, vibration data sample 3, vibration data sample 4, vibration data sample 5}, and at this time, the labels for vibration data sample 1, vibration data sample 2, vibration data sample 3, vibration data sample 4, and vibration data sample 5 are all unprocessed.

[0102] In other words, the label of an element in the first set indicates whether the weight value of that element has been determined. If the label is "unprocessed," it means that the weight value of that element has not been determined; if the label is "processed," it means that the weight value of that element has been determined.

[0103] 302. Randomly select an element marked as unprocessed from the first set as the first sample, take the collection time of the first sample as the first time, and take the collection location of the first sample as the first location.

[0104] For example, the first set is {vibration data sample 1, vibration data sample 2, vibration data sample 3, vibration data sample 4, vibration data sample 5}, and the labels of vibration data sample 1, vibration data sample 2, vibration data sample 3, vibration data sample 4, and vibration data sample 5 are all unprocessed. If vibration data sample 1 is selected as the first sample in step 302, then the acquisition time of the first sample (e.g., vibration data sample 1) is taken as the first time (e.g., time 1), and the acquisition position of the first sample (e.g., position 1) is taken as the first position.

[0105] Although vibration data sample 1 was selected at this point, no changes to the annotations were involved. Therefore, the annotations for vibration data sample 1, vibration data sample 2, vibration data sample 3, vibration data sample 4, and vibration data sample 5 remained unprocessed.

[0106] 303. The element that is not the first sample in the first set whose collection location is determined as the first location and whose collection time has the smallest time difference with the first time is taken as the second sample. The time difference between the collection time of the second sample and the first time is taken as the first time difference.

[0107] Because vibration data samples are collected by multiple acquisition devices located at different locations, it is possible for multiple vibration data samples to be collected at the same location but at different times (i.e., vibration data samples collected at different times by acquisition devices located at the same location during periodic acquisition), or multiple vibration data samples to be collected at different locations but at the same time (i.e., vibration data samples collected at the same time by acquisition devices located at different locations). For example, the first set is {vibration data sample 1, vibration data sample 2, vibration data sample 3, vibration data sample 4, vibration data sample 5}, where vibration data sample 1, vibration data sample 2, vibration data sample 3, and vibration data sample 4 are collected at the same location, i.e., location 1, and vibration data sample 5 is collected at location 2.

[0108] The implementation process of step 303 is as follows:

[0109] 303-1, Select all elements from the first set whose collection location is the first position and which are not the first sample as candidate samples.

[0110] For example, if the first sample is vibration data sample 1, then the candidate samples are vibration data sample 2, vibration data sample 3, and vibration data sample 4.

[0111] 303-2, determine the collection time for each candidate sample.

[0112] For example, the acquisition time of vibration data sample 2 is determined to be time 2, the acquisition time of vibration data sample 3 is determined to be time 3, and the acquisition time of vibration data sample 4 is determined to be time 4.

[0113] 303-3, determine the time difference between the collection time of each candidate sample and the first time.

[0114] It should be noted that the time difference in step 303 is a non-negative value, that is, the absolute value of the difference between the two times.

[0115] Therefore, in step 303-3, the time difference ∆t21 between vibration data sample 2 and the first time is calculated as |time2-time1|, the time difference ∆t31 between vibration data sample 3 and the first time is calculated as |time3-time1|, and the time difference ∆t41 between vibration data sample 4 and the first time is calculated as |time4-time1|.

[0116] 303-4, the candidate sample with the smallest time difference is determined as the second sample, and the time difference between the collection time of the second sample and the first time is taken as the first time difference.

[0117] For example, if ∆t21 is the smallest among ∆t21, ∆t31, and ∆t41, then vibration data sample 2 is determined as the second sample, and the first time difference is determined as ∆t21.

[0118] As can be seen from steps 303-1 to 303-4 above, the second sample was collected by the same acquisition device at the same location as the first sample, and the acquisition time of the second sample was the closest to that of the first sample. The second sample and the first sample can be considered as data collected from two consecutive acquisitions at the same location, and the first time difference can be considered as the acquisition period of periodic data acquisition at the same location.

[0119] Furthermore, the process so far has not involved any changes to the annotations, so the annotations for vibration data sample 1, vibration data sample 2, vibration data sample 3, vibration data sample 4, and vibration data sample 5 remain unprocessed.

[0120] 304. Determine the average time difference of the acquisition time of all vibration data samples acquired at the first position in any training sample class.

[0121] For example, the first set is {vibration data sample 1, vibration data sample 2, vibration data sample 3, vibration data sample 4, vibration data sample 5}, where vibration data sample 1, vibration data sample 2, vibration data sample 3, and vibration data sample 4 are collected at the same location, which is position 1, and vibration data sample 5 is collected at position 2. Then, in step 304, the average value of the acquisition time difference among vibration data sample 1, vibration data sample 2, vibration data sample 3, and vibration data sample 4 will be determined.

[0122] For example, if the time difference between vibration data sample 1 and vibration data sample 2 is determined to be ∆t21 = |time 2 - time 1|, the time difference between vibration data sample 1 and vibration data sample 3 is determined to be ∆t31 = |time 3 - time 1|, the time difference between vibration data sample 1 and vibration data sample 4 is determined to be ∆t41 = |time 4 - time 1|, the time difference between vibration data sample 2 and vibration data sample 3 is determined to be ∆t32 = |time 3 - time 2|, the time difference between vibration data sample 2 and vibration data sample 4 is determined to be ∆t42 = |time 4 - time 2|, and the time difference between vibration data sample 3 and vibration data sample 4 is determined to be ∆t43 = |time 4 - time 3|, the average time difference is (∆t21 + ∆t31 + ∆t41 + ∆t32 + ∆t42 + ∆t43) / 6.

[0123] The average time difference is the average collection period of position 1 obtained from the actual collected data.

[0124] Furthermore, the process so far has not involved any changes to the annotations, so the annotations for vibration data sample 1, vibration data sample 2, vibration data sample 3, vibration data sample 4, and vibration data sample 5 remain unprocessed.

[0125] 305. The sum of the first time difference and the average time difference is determined as the second time difference.

[0126] The first time difference is the acquisition period of position 1 obtained from two consecutive data acquisitions of position 1. The average time difference is the acquisition period of position 1 obtained from all data acquisitions of position 1. Under normal circumstances, the first time difference and the average time difference are the same. However, due to uncontrollable reasons such as recording delays, the first time difference and the average time difference may be different. In this case, it is not possible to determine which of the first time difference and the average time difference is more accurate. Therefore, when determining the second time difference, it is not to choose one of the first time difference and the average time difference for calculation, but to adopt both the first time difference and the average time difference at the same time. That is, the second time difference = the first time difference + the average time difference. Thus, the second time difference can be determined as twice the acquisition period.

[0127] Furthermore, the process so far has not involved any changes to the annotations, so the annotations for vibration data sample 1, vibration data sample 2, vibration data sample 3, vibration data sample 4, and vibration data sample 5 remain unprocessed.

[0128] 306. Elements in the first set whose time difference between the collection time and the first time is no greater than the second time difference are selected as the third sample.

[0129] For example, the first set is {vibration data sample 1, vibration data sample 2, vibration data sample 3, vibration data sample 4, vibration data sample 5}, where the acquisition time of vibration data sample 2 is time 2, the acquisition time of vibration data sample 3 is time 3, the acquisition time of vibration data sample 4 is time 4, and the acquisition time of vibration data sample 5 is time 5. In step 306, the time difference ∆t21 between time 2 and time 1 is determined to be |time 2 - time 1|, the time difference ∆t31 between time 3 and time 1 is determined to be |time 3 - time 1|, the time difference ∆t41 between time 4 and time 1 is determined to be |time 4 - time 1|, and the time difference ∆t51 between time 5 and time 1 is determined to be |time 5 - time 1|. If ∆t31 and ∆t51 are not greater than the second time difference, and ∆t21 and ∆t41 are greater than the second time difference, then vibration data sample 3 and vibration data sample 5 are determined as the third sample.

[0130] In other words, the third sample is data collected within two collection cycles at location 1, at the same collection time as the first sample. The third sample could be collected at location 1 or at other locations, and can be considered as data on the geological structures corresponding to class j collected within these two cycles. Since the time difference is non-negative, the two cycles could be two cycles forward from time 1 or two cycles backward from time 1. Therefore, the third sample is data on the geological structures corresponding to class j collected within two cycles forward and two cycles backward from time 1 (actually four cycles, with the cycle being the collection cycle at location 1).

[0131] Furthermore, the process so far has not involved any changes to the annotations, so the annotations for vibration data sample 1, vibration data sample 2, vibration data sample 3, vibration data sample 4, and vibration data sample 5 remain unprocessed.

[0132] 307. The third set is formed by collecting samples whose time difference with the first time is not greater than the first time difference, and the remaining third samples are formed into a third set.

[0133] In the third sample, those with a time difference no greater than the first time difference form the second set, and those with a time difference greater than the first time difference form the third set.

[0134] Since the third sample consists of geological structure data corresponding to class j collected within two cycles (actually four cycles, with the cycle being the acquisition cycle of position 1) starting from time 1, the third sample will be reclassified in step 307. The geological structure data corresponding to class j collected within one cycle (actually two cycles, with the cycle being the acquisition cycle of position 1) starting from time 1 will be formed into a set (i.e., the second set); the geological structure data corresponding to class j collected outside one cycle starting from time 1 but within two cycles (actually two cycles, with the cycle being the acquisition cycle of position 1) will be formed into a set (i.e., the third set).

[0135] Here, the duration of one cycle is determined by the first time difference. This is because the first time difference is obtained from only two vibration data samples, while the average time difference is obtained from multiple vibration data samples. If there are errors during data acquisition, the amount of error introduced by two vibration data samples is less than the amount of error introduced by multiple vibration data samples. Therefore, the first time difference is used as the duration of one cycle.

[0136] Thus, in step 306, all vibration data samples of the geological structure corresponding to class j (regardless of the acquisition location) are divided into two parts. The first part is acquired within two cycles before and after the first time (actually four cycles, with the cycle being the acquisition cycle of location 1), and the second part is acquired outside of two cycles before and after the first time. In step 307, the first vibration data sample is again divided into two parts. The first part (A) is acquired within one cycle before and after the first time (actually two cycles, with the cycle being the acquisition cycle of location 1), and the first part (B) is acquired outside of one acquisition cycle before and after the first time. The first part (A) forms the second set, and the first part (B) forms the third set.

[0137] Furthermore, the process so far has not involved any changes to the annotations, so the annotations for vibration data sample 1, vibration data sample 2, vibration data sample 3, vibration data sample 4, and vibration data sample 5 remain unprocessed.

[0138] Furthermore, the process so far has not involved any changes to the annotations, so the annotations for vibration data sample 1, vibration data sample 2, vibration data sample 3, vibration data sample 4, and vibration data sample 5 remain unprocessed.

[0139] 308. Determine the first similarity between each element in the second set and the first sample, and determine the second similarity between each element in the third set and the first sample.

[0140] The elements in the second set were collected closer to the first sample. For the same geological structure, the shorter the collection time, the more similar the vibration data samples are (because environmental and earthquake factors do not change much in a short period of time). Therefore, a first similarity is determined between each element in the second set and the first sample, and a second similarity is determined between each element in the third set and the first sample. The first similarity characterizes the degree of similarity between the elements in the second set and the first sample, and the second similarity characterizes the degree of similarity between the elements in the third set and the first sample.

[0141] The similarity determination scheme can adopt existing methods, such as extracting data features of each element in the second and third sets to form feature vectors, and extracting data features of the first sample to form feature vectors. The Euclidean distance between the feature vectors of each element in the second set and the feature vector of the first sample is calculated, and the Euclidean distance between the feature vectors of each element in the third set and the feature vector of the first sample is calculated. The Euclidean distance is then used to determine the similarity.

[0142] The range of Euclidean distance is [0, +∞). The smaller the Euclidean distance, the more similar the two people are, and therefore the greater the similarity. In other words, if the similarity value is Euclidean distance, then the smaller the similarity value, the more similar the two people are, and the higher the degree of similarity.

[0143] Furthermore, the process so far has not involved any changes to the annotations, so the annotations for vibration data sample 1, vibration data sample 2, vibration data sample 3, vibration data sample 4, and vibration data sample 5 remain unprocessed.

[0144] 309. Determine the weight values ​​of the processed sample and each processed sample based on the first similarity and the second similarity, and update the label of the processed sample to "processed". Repeat the step of randomly selecting an element labeled as unprocessed from the first set as the first sample and the subsequent steps (i.e., steps 302 to 309) until all elements in the first set are labeled as "processed".

[0145] The implementation process of step 309 is as follows:

[0146] 309-1, the samples to be processed are determined based on the first similarity and the second similarity.

[0147] The implementation process of step 309-1 is as follows:

[0148] 1. Determine the first minimum, first maximum, and first mean of all first similarities.

[0149] 2. Determine the second minimum, second maximum, and second mean of all second similarities.

[0150] 3. Determine the first benchmark value = (first minimum value + first maximum value + 4 × first mean value) / 6.

[0151] The first benchmark value is the similarity value between each element in the second set and the first sample. The first benchmark value can characterize the most likely similarity between each element in the second set and the first sample.

[0152] 4. Determine the second benchmark value = (second minimum value + second maximum value + 4 × second mean value) / 6.

[0153] The second benchmark value is the similarity value obtained between each element in the third set. The second benchmark value can represent the most likely similarity between each element in the third set and the first sample.

[0154] Since the elements in the second set were collected at closer times than those in the first set, if Euclidean distance is used as the similarity metric, then the first benchmark value is no greater than the second benchmark value.

[0155] 5. The center position of the acquisition location corresponding to the first similarity score not greater than the first benchmark value is determined as the first position. The center position of the acquisition location corresponding to the second similarity score not greater than the second benchmark value is determined as the second position.

[0156] For example, for the first reference value,

[0157] 1) Among all the first similarities in the second set, determine all first similarities that are not greater than the first benchmark value (for ease of explanation, the similarities determined here are named first candidate similarities).

[0158] The first candidate similarity is the first similarity that is the same as or more similar to the first benchmark value.

[0159] 2) Determine the collection location of the element corresponding to the first candidate similarity (for ease of explanation, the collection location determined here will be named the first candidate location).

[0160] The collection location here can be represented by latitude and longitude coordinates. For example, any first candidate location is (xu, yu), where u is the index of the first candidate location, xu is the longitude value of the first candidate location u, and yu is the latitude value of the first candidate location u.

[0161] 3) Determine the center position of all first candidate positions.

[0162] For example, if any first candidate position is (xu, yu), and there are a total of U first candidate positions, then the center position is (∑xu / U, ∑yu / U).

[0163] 4) Determine the first position as (∑xu / U, ∑yu / U).

[0164] The first position is the middle position of the vibration data sample that is more similar to the first sample and was acquired within one cycle before and after the first time.

[0165] For the second benchmark value,

[0166] 1) Among all the second similarities in the third set, determine all second similarities that are not greater than the second benchmark value (for ease of explanation, the similarities determined here are called second candidate similarities).

[0167] The second candidate similarity is the second similarity that is the same as or more similar to the second benchmark value.

[0168] 2) Determine the collection location of the element corresponding to the second candidate similarity (for ease of explanation, the collection location determined here will be named the second candidate location).

[0169] The collection location here can be represented by latitude and longitude coordinates. For example, any second candidate location is (xv, yv), where v is the index of the second candidate location, xv is the longitude value of the second candidate location v, and yv is the latitude value of the second candidate location v.

[0170] 3) Determine the center position of all second candidate positions.

[0171] For example, if any second candidate position is (xv, yv), and there are a total of V second candidate positions, then the center position is (∑xv / V, ∑yv / V).

[0172] 4) Determine the second position as (∑xv / V, ∑yv / V).

[0173] The second position is the middle position of the vibration data sample that was acquired one cycle before and one cycle after time 1, but within two cycles, and is more similar to the first sample.

[0174] 6. Determine whether the distance between the first position and the second position is not greater than the maximum distance between the first position and the acquisition position corresponding to the first similarity value which is not greater than the first benchmark value.

[0175] In this step:

[0176] 1) Determine the distance between the first position and the second position.

[0177] If the first position is (∑xu / U, ∑yu / U) and the second position is (∑xv / V, ∑yv / V), then the distance between the first and second positions is sqrt{(∑xv / V-∑xu / U)×(∑xv / V-∑xu / U)+(∑yv / V-∑yu / U)×(∑yv / V-∑yu / U)}.

[0178] 2) Determine the first similarity, which is no greater than the first benchmark value, i.e., the first candidate similarity.

[0179] 3) Determine the collection location of the element corresponding to each first candidate similarity, i.e., the first candidate position.

[0180] 4) Determine the distance between the first position and the collection position corresponding to the first similarity not greater than the first benchmark value, that is, determine the distance between the first position and the first candidate position. For ease of explanation, the distance determined here is named the first candidate distance.

[0181] The method for calculating the distance between the two positions is the same as the method for determining the distance between the first and second positions in 1), and will not be repeated here.

[0182] 5) Determine the maximum distance among all first candidate distances.

[0183] The maximum distance characterizes the farthest location between the acquisition locations of vibration data samples that are similar to the first sample.

[0184] 6) Determine whether the distance between the first position and the second position is not greater than the maximum distance.

[0185] The first position is the central acquisition position of similar data within the first cycle starting from the first time point. The second position is the central acquisition position of similar data outside the first cycle but within the second cycle starting from the second time point. The distance between the first and second positions is the distance between the central acquisition positions of similar data in the two cycles, with the maximum distance being the farthest position between the acquisition positions of vibration data samples that are similar to the first sample.

[0186] If the distance between the first and second positions is greater than the maximum distance, it indicates that the center acquisition positions of similar data from two periods are relatively far apart. If the distance between the first and second positions is not greater than the maximum distance, it indicates that the center acquisition positions of similar data from two periods are relatively close together.

[0187] 7. If the distance between the first position and the second position is not greater than the maximum distance between the first position and the acquisition position corresponding to the first similarity score not greater than the first benchmark value, it indicates that the acquisition positions within the two periods are relatively close. In this case, the first sample, as well as all vibration data samples corresponding to the first similarity score not less than the first benchmark value and all vibration data samples corresponding to the second similarity score not less than the second benchmark value, are all determined as processing samples. That is, the first sample and the vibration data samples that are relatively close within the two periods are all determined as processing samples.

[0188] 8. If the distance between the first position and the second position is greater than the maximum distance between the first position and the acquisition position corresponding to the first similarity value not less than the first benchmark value, it indicates that the acquisition positions within the two periods are far apart. In this case, the first sample and vibration data samples with a similarity value not greater than the first benchmark value are both determined as processing samples. That is, the first sample and vibration data samples that are relatively close within the most recent period are both determined as processing samples.

[0189] 309-2, determine the weight value of each processed sample.

[0190] The weight values ​​for each processed sample are determined through the following steps:

[0191] 1. Form a fourth set from all processed samples.

[0192] 2. Calculate the third similarity between any two elements in the fourth set.

[0193] 3. The normalized value of the difference between each third similarity and the minimum value among the third similarities is determined as the weight value.

[0194] The fourth set consists of the first sample and similar vibration data samples. Since the similarity is obtained through Euclidean distance, the smaller the similarity value, the more similar the data. Therefore, the minimum value in the third similarity is the Euclidean distance between the two most similar data in the fourth set.

[0195] For any processed sample, the difference between its third similarity and the minimum of the third similarities is the difference between the similarity between that processed sample and the first sample and the most similar sample. The normalized value of this difference is determined as the weight value of that processed sample.

[0196] Vibration data samples collected closer in distance and time are more likely to be similar. The more similar the data, the less unique the sample, and the lower its weight. In other words, the minimum value in the third similarity score represents the highest degree of similarity. The smaller the similarity between another vibration data sample and the first sample compared to the minimum value in the third similarity score, the smaller the normalized value of the difference between each third similarity score and the minimum value. This indicates a higher similarity between the other vibration data sample and the first sample, suggesting that the sample is less unique and has a lower degree of individualization, resulting in a lower weight. Therefore, the normalized value of the difference between the third similarity score and the minimum value in the third similarity score is directly proportional to the weight. Thus, the weight is determined as the normalized value of the difference between the third similarity score and the minimum value in the third similarity score.

[0197] It should be noted that the examples in this embodiment and subsequent real-time examples are only for illustrating the implementation and do not represent real data. In specific implementation, actual data can be used.

[0198] 309-3, update the label of the processed sample to "processed".

[0199] After completing step 309-2, the weight values ​​of all processed samples will be obtained, and the process of determining the weight values ​​of the processed samples is complete. Therefore, in step 309-3, the label of the processed samples is updated to "processed," indicating that the process of determining their weight values ​​is complete.

[0200] If the samples to be processed are vibration data sample 1 and vibration data sample 3, after executing step 309-3, the label of vibration data sample 1 will be updated to "processed", the label of vibration data sample 2 will remain "unprocessed", the label of vibration data sample 3 will be updated to "processed", the label of vibration data sample 4 will remain "unprocessed", and the label of vibration data sample 5 will remain "unprocessed".

[0201] 309-4. If all elements in the first set are marked as processed, it means that the weight values ​​of all elements in the first set have been determined, that is, the weight values ​​of all vibration data samples of class j have been determined, thus completing the process of determining the weight values ​​of vibration data samples in the training samples of class j. If there are elements marked as unprocessed in the first set, it means that there are elements in the first set whose weight values ​​have not been determined. Then, repeat steps 302 to 309 until all elements in the first set are marked as processed.

[0202] 203. The vibration simulation model is trained based on various training samples to obtain the parameters of the trained vibration simulation model for each type.

[0203] The training process utilizes existing methods, such as inputting various training samples into the vibration simulation model for training, and continuously adjusting the model parameters to enable accurate simulation of various seismic chain disasters. For example, based on various training samples, the vibration simulation model is repeatedly trained using a hybrid validation method combining scaled models and numerical simulations, combined with the physical characteristics of typical geological structures in high-altitude areas (such as permafrost slopes, glacial till, and glacial deposits). This allows the vibration simulation model to achieve high-fidelity scene reproduction simulations, as well as simulations of disaster chain triggering thresholds and evolution paths.

[0204] After training, vibration simulation model parameters for each geological structure will be obtained.

[0205] In addition, during training, each vibration data sample has a weight value, which represents the degree of individuality of the vibration data sample in its corresponding geological structure. By considering the weight value as a feature during training, the vibration simulation model can accurately simulate various special situations.

[0206] 204. Use the parameters of all trained vibration simulation models and the vibration simulation models as the trained vibration simulation models.

[0207] In step 102, after the vibration simulation requirements are input into the vibration simulation model, the vibration simulation model identifies the geological structure corresponding to the vibration simulation requirements, and then uses the parameters of the geological structure corresponding to the vibration simulation requirements to simulate the vibration simulation requirements.

[0208] In practical implementation, the high-altitude glacier and permafrost earthquake chain disaster simulation device can, as follows: Figure 2 The diagram shows: an environmental control chamber and a vibration table array.

[0209] The environmental control chamber and the shaking table array can be designed as an integrated unit or as separate units. For example, the environmental control chamber can be moved and installed on the upper truss platform (such as a steel frame platform) of the shaking table array, or the environmental control chamber can be moved independently to the side of the shaking table array via a track.

[0210] In practice, depending on the experimental conditions, it can be decided whether the environmental control chamber is installed on the upper truss platform (such as a steel frame platform) of the vibration table array.

[0211] The environmental control cabin can independently conduct simulations of various extreme weather events, or it can be integrated into a shaking table array to conduct simulations of the interaction between the environment and earthquakes, depending on the testing requirements.

[0212] 1. Environmental Control Chamber

[0213] The environmental control cabin is used to simulate extreme and complex climatic environments at high altitudes.

[0214] The environmental control cabin includes: a low-temperature control system, a rainfall unit, a snowfall unit, a strong radiation simulation unit, a humidity control unit, and a wind speed control unit.

[0215] The low-temperature control system, rainfall unit, snowfall unit, strong radiation simulation unit, humidity control unit, and wind speed control unit are integrated into an environmental simulation control cabin, which can be raised, lowered, or moved horizontally. The environmental simulation cabin can independently carry out engineering tests under various extreme climate events, or it can be integrated into a vibration table according to testing needs to carry out research on the interaction between the environment and earthquakes.

[0216] 1) Low-temperature control system

[0217] The cryogenic control system includes: an environmental control chamber temperature control unit and a model freeze-thaw temperature control unit.

[0218] (1) Temperature control unit of the environmental control cabin

[0219] The temperature control unit of the environmental control chamber uses liquid nitrogen circulation refrigeration technology to control the temperature.

[0220] The temperature control unit of the environmental control chamber has a temperature control range of -40℃ (Celsius) to 60℃, and a temperature control accuracy of ±1℃.

[0221] In other words, the temperature control unit of the environmental control cabin uses liquid nitrogen circulation refrigeration technology to achieve temperature control covering a wide temperature range of -40℃ to 60℃.

[0222] (2) Model freeze-thaw temperature control unit

[0223] The model freeze-thaw temperature control unit uses a high-power cold bath to control a unidirectional heat transfer metal cold plate, and the temperature is controlled by covering the model with the cold plate.

[0224] The temperature control range of the cold plate is -40℃ to 20℃, and the temperature control accuracy is ±1℃.

[0225] In other words, the model freeze-thaw control unit controls a unidirectional heat transfer metal cold plate through a high-power cold bath, and the model is covered by the cold plate to achieve freeze-thaw control of the model.

[0226] 2) Rainfall Unit

[0227] The rainfall unit consists of water pipes and sprinklers.

[0228] The rainfall capacity of each rainfall unit is 0~200 mm / h, and the rainfall uniformity is >90%.

[0229] In other words, the rainfall unit uses integrated water pipes and nozzles to generate artificial rainfall, simulating a rainfall capacity of 0~200mm / h and a uniformity of >90%.

[0230] 3) Snowfall Unit

[0231] The snowmaking unit includes a snowmaking machine.

[0232] The snowmaking machine operates normally in an environment of -40℃ to 20℃.

[0233] The snowmaking capacity of the snowmaking unit is 0~20mm / h, and the snowmaking uniformity is >90%.

[0234] In other words, the snowmaking unit is achieved by integrating a snowmaking machine into the environmental control cabin.

[0235] 4) High Radiation Simulation Unit

[0236] The high-intensity radiation simulation unit includes a radiation panel.

[0237] The radiation panel is integrated into the inner wall of the environmental control cabin and is used to simulate the strong sunlight and ultraviolet radiation at high altitudes.

[0238] The high-intensity radiation simulation range of the high-intensity radiation simulation unit is 0~1500W / m. 2 (watts per square meter), with a high-intensity radiation simulation accuracy of ±1%.

[0239] In other words, the high-intensity radiation simulation unit simulates the intense sunlight and ultraviolet radiation at high altitudes by integrating radiation panels into the cabin walls.

[0240] 5) Humidity control unit

[0241] The humidity control unit uses ultrasonic humidification technology and a humidity sensor detection system for real-time feedback and adjustment.

[0242] The humidity control unit has a humidity control range of 10% to 90% and a humidity control accuracy of ±3%.

[0243] In other words, the humidity control unit uses ultrasonic humidification technology and a humidity sensor detection system to provide real-time feedback and adjustment, thereby achieving precise control of the humidity environment.

[0244] 6) Wind speed control unit

[0245] The wind speed control unit includes a high-power fan.

[0246] The wind speed control unit has a wind speed control range of 0 to 8 levels and a wind speed control accuracy within ±1%.

[0247] In other words, the wind speed control unit achieves wind speed control through a high-power fan, which can simulate wind force from level 0 to 8 with an accuracy within ±1%.

[0248] 2. Vibration table array

[0249] Shaking table arrays are used to simulate seismic dynamic output.

[0250] The shaking table array consists of multiple shaking tables.

[0251] The vibration table is a three-axis, six-degree-of-freedom vibration table, which uses an electro-hydraulic servo actuator and loads one or more of the following: sine wave, random wave, and seismic wave.

[0252] The maximum acceleration of the vibration table is (±2.0g, ±2.0g, ±1.5g), the displacement is (±500mm, ±500mm, ±300mm), and the frequency range is 0.1~60Hz (Hertz).

[0253] Where g is the acceleration due to gravity.

[0254] For example, a shaking table array is composed of three shaking tables integrated together.

[0255] One vibration table is 6m × 8m and is fully loaded with 100t (tons), reaching full load at its maximum acceleration; the other two vibration tables are 3m × 4m and are fully loaded with 40t, and their acceleration at full load is not necessarily at its maximum value.

[0256] The vibration tables in the vibration table array are integrated into a single platform. The integrated vibration table array is 8m×9m in size and has a full load capacity of 180t. The maximum acceleration is (±1.8g, ±1.8g, ±1.3g), the displacement is (±500mm, ±500mm, ±300mm), and the frequency range is 0.1~60Hz.

[0257] or,

[0258] The shaking tables in the shaking table array move along tracks in a linear arrangement. In the linearly arranged shaking table array, the spacing between three shaking tables is 12m to 150m, and the movement range is 138m.

[0259] By integrating multiple shaking tables into a single platform or arranging them linearly via track movement, multifunctional and efficient simulations can be achieved for different research purposes, such as environmental effect studies, seismic resistance studies of engineering structures, and collaborative studies of environment and earthquake.

[0260] Taking the shaking table array as an example, which is composed of three three-axis six-degree-of-freedom shaking tables (one 6m×8m and two 3m×4m), each shaking table adopts an electro-hydraulic servo actuator, which can load sine waves, random waves and seismic waves.

[0261] A 6m × 8m shaking table, fully loaded with 100t, has the following maximum accelerations: X-axis (lateral) ±2.0g, Y-axis (longitudinal) ±2.0g, Z-axis (vertical) ±1.5g. Displacement: X-axis (lateral, horizontal displacement) ±500mm, Y-axis (longitudinal, vertical displacement) ±500mm, Z-axis (vertical, vertical displacement) ±300mm. Frequency range: 0.1~60Hz.

[0262] A 3m x 4m vibration table, fully loaded with 40t, has the following maximum accelerations: X-axis (lateral) ±2.0g, Y-axis (longitudinal) ±2.0g, Z-axis (vertical) ±1.5g. Displacement: X-axis (lateral, horizontal displacement) ±500mm, Y-axis (longitudinal, vertical displacement) ±500mm, Z-axis (vertical, vertical displacement) ±300mm. Frequency range: 0.1~60Hz.

[0263] If the vibration tables in the vibration table array are integrated into a single platform (e.g.) Figure 3 As shown in the diagram, the integrated vibration table array has a table size of 8m × 9m, and the individual table sizes are 8m × 6m + 4m × 3m + 4m × 3m. The three vibration tables together form a large-scale vibration platform with high load-bearing capacity (i.e., a vibration table array). The integrated vibration table array can bear a load of up to 180t. The maximum acceleration of the integrated vibration table array is ±1.8g in the X direction (lateral), ±1.8g in the Y direction (longitudinal), and ±1.3g in the Z direction (vertical). Displacement: ±500mm in the X direction (lateral, i.e., horizontal displacement), ±500mm in the Y direction (longitudinal, i.e., vertical displacement), and ±300mm in the Z direction (vertical, i.e., vertical displacement). The frequency range is 0.1~60Hz.

[0264] If the shaking tables in the shaking table array are arranged linearly via moving tracks, linear engineering seismic simulation can be performed (e.g.) Figure 4 As shown, the three shaking tables can be spaced at a maximum distance of 150m and a minimum distance of 12m, with a movement range of 138m, and support long-duration seismic wave input.

[0265] In addition, in practical implementation, the high-altitude glacier permafrost earthquake chain disaster simulation device may include, in addition to the aforementioned environmental control chamber and shaking table array, a data acquisition module.

[0266] The data acquisition module includes: a sensor network for low-temperature environments, high-speed camera equipment and laser scanning equipment, and a data acquisition and control host.

[0267] 1) Sensor networks in low-temperature environments

[0268] A sensor network for low-temperature environments is used to acquire static parameters during vibration simulation processes in low-temperature environments.

[0269] Sensor networks for low-temperature environments include one or more of the following: accelerometers, laser displacement sensors, pore water pressure sensors, soil pressure sensors, and fiber Bragg grating sensors.

[0270] (1) Accelerometer

[0271] The accelerometer has a range of ±50gpk (peak value of gravitational acceleration) (30 units) or ±200gpk (20 units), a sensitivity of ≥100mV / g (millivolts / gravitational acceleration) (±50gpk) or ≥10mV / g (±200gpk), a waterproof and dustproof rating of ≥IP68, a water pressure load of ≥400kPa, and an operating temperature range of -50℃ to +60℃.

[0272] (2) Laser displacement sensor

[0273] The laser displacement sensor has a measuring range of 200 mm, an operating range of ≥60 mm to 260 mm, a linearity of ≤±0.1%FS (full scale), and a linear error range of ≤±200 μm (micrometers). It exhibits temperature stability of ≤±0.02%FS / K (full scale / Kelvin), a measurement frequency of 2 kHz, and an operating temperature range of -50℃ to +60℃.

[0274] (3) Pore water pressure sensor

[0275] The pore water pressure sensor has a measurement range of -100kPa to 1.2MPa, an overload pressure of ≥150%, a breaking pressure of ≥200%, a measurement accuracy of ≤±0.45%FS, and an operating temperature range of -50℃ to +60℃.

[0276] (4) Earth pressure sensor

[0277] The earth pressure sensor has a range of 0~1.2MPa, an overload pressure of ≥120%, a breaking pressure of ≥150%, a soft contact structure, a resolution of ≤100Pa, a measurement accuracy of ≤±0.40%FS, a frequency response of 100kHz~300kHz, a response time of ≤0.2ms, and an operating temperature range of -50℃~+60℃.

[0278] (5) Fiber Bragg grating sensor

[0279] The fiber Bragg grating sensor has a light source wavelength range of 1500 nm to 1600 nm, meeting the Class 1 laser safety standard, with an output power ≥ 0.06 mW and a spectral resolution of 10 pm. The sensor's temperature measurement range is -40℃ to +80℃, with a measurement accuracy of ±0.5℃. The strain measurement range is -1500 με to 5000 με, with a resolution of ±1 με, a strain coefficient of 1.1 pm / με, a measurement accuracy of 0.3%FS, and an operating temperature range of -40℃ to 85℃.

[0280] Sensor networks for low-temperature environments can be based on fiber optic gratings (FBG) and micro-electro-mechanical systems (MEMS) technologies to achieve simultaneous acquisition and monitoring of strain, temperature, displacement, and pore water pressure at -40°C.

[0281] 2) High-speed camera equipment and laser scanning equipment

[0282] High-speed cameras and laser scanning equipment are used to acquire millisecond-level dynamic evolution parameters during vibration simulation in real time. These devices can capture millisecond-level dynamic evolution processes in vibration simulation, such as crack propagation and avalanche initiation.

[0283] Among them, the high-speed camera has ≥20 million effective pixels, an image frame rate ≥18 frames (e.g., 1000 frames), a lens focal length of 50mm, and operates normally in an environment of -50℃ to 60℃.

[0284] 3) Data Acquisition and Control Host

[0285] The data acquisition and control host is used to communicate with the host computer in real time via Ethernet, synchronize time information, location information and data exchange, and control the start-up, stop, vibration simulation and power failure control protection of the high-altitude glacier permafrost earthquake chain disaster simulation device.

[0286] The total sampling rate of the acquisition and control host is ≥400ksps (thousands of samples per second).

[0287] In addition, in specific implementations, the data acquisition module may also include: a strain and bridge measurement module and / or a vibration measurement acquisition module.

[0288] The data acquisition and control host can be equipped with an embedded industrial computer, which can control the operation of the low-temperature sensor network, high-speed camera equipment, and laser scanning equipment in the data acquisition module through the data acquisition and control host.

[0289] In practical implementation, the data acquisition and control host can communicate with the host computer in real time at high speed via Ethernet (≥1000 Mbit / s). The host features independent complex triggering functions, independent operation and self-starting (timer / absolute time), and synchronization with an external GPS (Global Positioning System) time or location information clock, enabling data exchange and phase consistency with existing laboratory equipment. The host also has an automatic power-off control protection function (ensuring data integrity in the event of a power failure), meeting the requirements for unattended, long-term, uninterrupted data acquisition.

[0290] 4) Strain and Wheatstone Bridge Measurement Module

[0291] The strain and bridge measurement module is used for automatic bridge voltage calibration.

[0292] The strain and bridge measurement module is equipped with one or more bridge modes and I / O connectors, with a maximum single-channel sampling rate of ≥128kHz, a signal bandwidth of ≥48kHz, and a signal synchronization error of ≤0.0002ms.

[0293] For example, the strain gauge and bridge measurement module is configured with full-bridge (four-wire), half-bridge (four-wire), 1 / 4-bridge (three-wire), DC excitation, and single-sense cable compensation bridge modes. The maximum sampling rate per channel is ≥128kHz, signal bandwidth is ≥ (-3dB) 48kHz, signal synchronization error is ≤0.0002ms, and it supports standard I / O (input / output) connectors such as the LEMO.1B self-locking aviation plug. It features automatic bridge voltage calibration to ensure remote bridge voltage accuracy without requiring measurement of conductor resistance or correction.

[0294] 5) Vibration measurement and acquisition module

[0295] The vibration measurement and acquisition module is used to collect vibration parameters during vibration simulation. The single-channel maximum sampling rate of the vibration measurement and acquisition module is ≥128kHz, the signal bandwidth is ≥100kHz, and the signal synchronization error is ≤0.0002ms.

[0296] For example, the vibration measurement and acquisition module is equipped with vibration modes such as DC excitation and IEPE; it supports and is compatible with EID smart conductors and TEDS smart sensors for direct read / write functions, with a maximum single-channel sampling rate of ≥128kHz, a signal bandwidth of ≥(-3dB)100kHz, and is compatible with BNC standard plugs. It has an integrated ICP signal mode and a signal synchronization error of ≤0.0002ms.

[0297] The acquisition control host and the strain and bridge measurement module and / or vibration measurement acquisition module are connected by electrical and mechanical means to form an acquisition module.

[0298] The data acquisition module operates in a temperature range of -50℃ to +60℃, has an impact / vibration protection rating of ≥100g / 3~5ms, a waterproof and dustproof rating of ≥IP65, and is made of high-strength aluminum alloy.

[0299] In other words, the data acquisition and control host can be freely combined with the strain and bridge measurement modules and the vibration measurement acquisition module. Measurement modules can be added, replaced, repositioned, and numbered to complete the electrical and mechanical connections between multiple modules. The operating temperature range is -40℃ to +50℃ (non-condensing), with an extreme operating temperature range of -50℃ to +60℃ (condensing). Shock / vibration protection level is ≥100g / 3~5ms, waterproof and dustproof rating is ≥IP65, and the exterior is made of high-strength aluminum alloy.

[0300] In addition, the high-altitude glacier permafrost earthquake chain disaster simulation device can also perform vibration control, data acquisition, and data analysis through signal acquisition and control software.

[0301] The signal acquisition and control software has functions such as managing channel configuration lists, rich classification and filtering, and sensor management lists; it automatically executes programmed measurement tasks, such as channel balancing, inputting test object data, starting measurement, saving data, and performing analysis, and can quickly and intuitively complete system configuration. In addition, the signal acquisition and control software features a human-computer interaction interface for secondary development and supports programming languages ​​such as LabVIEW, C++, C#, and Visual Basic. It also allows for free input and output of interface-based content to the system, easily switching between static and dynamic acquisition modes to meet actual experimental needs. It includes online spectrum analysis (such as FFT oscilloscope, FFT analysis, power spectrum, amplitude / phase frequency, windowing, overlap, and averaging), calculus processing, digital filtering, strain analysis, stress analysis, displacement analysis, vibration analysis, error analysis, resampling, truncation, deletion, and merging functions. Test reports can be edited according to custom templates and exported in PDF, Word, and other formats. It supports 64-bit fast data processing and covers various basic statistical operations. It has a standard third-party data software interface protocol, supporting TCP / IP and user-defined interface protocols. Real-time analysis and processing data from the signal acquisition and control software can be directly obtained through the data interface, along with Fw-S automatic calibration and correction functions and on-site Cs-i parameter calibration functions, allowing for quick and intuitive understanding of the system's current status. Input signal voltage measurement range ≥ ±20mV~±10V (volts), minimum resolution ≥ 0.01mV, voltage signal measurement error ≤ ±0.3%FS; input current signal measurement range 5~30mA (milliamperes); input strain signal measurement range ≥ ±1000με~±100000με, strain signal measurement error ≤ ±0.5%FS.

[0302] In addition, regarding servo control, the high-altitude glacier and permafrost earthquake chain disaster simulation device, based on the existing mainstream three-state servo control strategy, introduces control strategies such as μ-synthetic decoupling, robust control, and disturbance compensation to improve the response and control accuracy of the high-altitude glacier and permafrost earthquake chain disaster simulation device, suppress nonlinear changes and random perturbation disturbances, and improve the high-frequency acceleration response and control accuracy. In terms of vibration control, a zero-position linearized kinematic solution method and a non-zero-position kinematic forward and inverse solution synthesis and matrix decomposition are adopted to achieve redundant real-time high-precision collaborative control of the platform, rotation, and linkage of single / multiple units in multi-degree-of-freedom space. Regarding instrument development, a specially developed high-flow-rate, high-frequency-response servo valve achieves high-speed and high-load control, meeting the requirements of 100t vibration load and near-field large-pulse ground vibration excitation. In terms of monitoring and analysis, a high-resolution wide dynamic range acquisition instrument and a GPU (Graphics Processing Unit) big data processing and analysis module are used to achieve synchronous high-speed monitoring and analysis of changes in multiple physical environments, and to analyze the experimental model status and decide on further simulation schemes based on the monitoring results.

[0303] The high-altitude glacier and permafrost earthquake chain disaster simulation device used in this embodiment is a shaking table array designed for simulating earthquake-induced chain disasters under extreme climate conditions at high altitudes. It can closely combine the complexity of extreme environments and the dynamic evolution of multiple disasters to achieve indoor reproduction of multiple physical processes.

[0304] In practical implementation, the input waveforms for each vibration table can be regular waves, random waves, seismic waves, artificial waves, etc. The maximum overturning moment of each vibration table is not less than 400 t·m, and the maximum eccentric moment is not less than 100 t·m. The operating input signals for each vibration table are acceleration time history, velocity time history, displacement time history; response spectrum; and power spectrum. Each vibration table operates via hydraulic servo drive. The data acquisition system for each vibration table is dynamic, with 256 channels + 64 channels + 64 channels; low-frequency sampling is 1Hz; the software can perform power spectrum analysis, modal analysis, etc.

[0305] Existing shaking tables or arrays can only simulate earthquakes and cannot take into account the limitations of the interaction between earthquakes and the environment. The high-altitude glacier permafrost earthquake chain disaster simulation device used in this embodiment can dynamically simulate the disaster evolution process under the combined action of extreme climate environmental factors and earthquakes, and build a full-process experimental capability of "environmental disaster incubation - earthquake disaster - disaster chain evolution".

[0306] The vibration simulation method provided in this embodiment trains a vibration simulation model using multiple real environmental and seismic ground motion data from typical geological structures in high-altitude areas. This simulation model can precisely control the high-altitude glacier-permafrost earthquake chain disaster simulation device to perform vibrations that meet the required vibration levels. Furthermore, the environmental control chamber of the high-altitude glacier-permafrost earthquake chain disaster simulation device simulates the unique climatic environment of high-altitude glacier-permafrost regions, and the vibration table simulates earthquakes, thus simulating the vibrations that induce chain disasters in high-altitude glacier-permafrost regions.

[0307] This embodiment relates to a vibration simulation method, which involves obtaining vibration simulation requirements; based on these requirements, controlling a vibration simulation device to perform vibration simulation using a pre-trained vibration simulation model; wherein the vibration simulation model is trained based on training samples; wherein the training samples include vibration data samples and geological data samples; wherein the vibration data samples consist of multiple real environmental data and seismic ground motion data of typical geological structures in high-altitude areas; wherein the geological data samples include physical characteristic data of typical geological structures in high-altitude areas. The method of this embodiment trains a vibration simulation model using multiple real environmental data and seismic ground motion data of typical geological structures in high-altitude areas, and controls a vibration simulation device to perform vibration simulation using the vibration simulation model, thereby simulating the vibrations of earthquake-induced chain reactions in high-altitude glacial and permafrost regions.

[0308] Since the systems / devices described in the above embodiments of the present invention are systems / devices used to implement the methods of the above embodiments of the present invention, those skilled in the art can understand the specific structure and modifications of the systems / devices based on the methods described in the above embodiments of the present invention, and therefore will not be repeated here. All systems / devices used in the methods of the above embodiments of the present invention fall within the scope of protection of the present invention.

[0309] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0310] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, as well as combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions.

[0311] It should be noted that any reference numerals placed between parentheses in the claims should not be construed as limiting the claims. The word "comprising" does not exclude the presence of components or steps not listed in the claims. The word "a" or "an" preceding a component does not exclude the presence of a plurality of such components. The invention can be implemented by means of hardware comprising several different components and by means of a suitably programmed computer. In claims that enumerate several means, several of these means may be embodied by the same hardware. The use of the terms first, second, third, etc., is merely for convenience of expression and does not indicate any order. These terms can be understood as part of the component names.

[0312] Furthermore, it should be noted that in the description of this specification, the terms "one embodiment," "some embodiments," "embodiment," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Furthermore, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0313] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the claims should be interpreted to include both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0314] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, then this invention should also include these modifications and variations.

Claims

1. A method for simulating earthquake chain disasters in high-altitude glaciers and permafrost, characterized in that, The method includes: Obtain vibration simulation requirements; According to the vibration simulation requirements, a high-altitude glacier permafrost earthquake chain disaster simulation device is controlled by a pre-trained vibration simulation model to perform vibration simulation. The vibration simulation model is trained based on training samples. The training samples include vibration data samples and geological data samples; The vibration data sample consists of multiple real environmental data and seismic motion data of typical geological structures in high-altitude areas. Among them, the geological data samples include physical characteristic data of typical geological structures in high-altitude areas; The vibration simulation model is trained through the following steps: The training samples are clustered according to the typical geological structures of high-altitude areas, so that each typical geological structure in high-altitude areas corresponds to a class of training samples. The vibration simulation model was trained based on various training samples to obtain the parameters of the trained vibration simulation model for each type. All trained vibration simulation model parameters and vibration simulation models are used as trained vibration simulation models; The vibration data sample includes the acquisition time, acquisition location, and weight value; The weights of vibration data samples in any class of training samples are determined through the following steps: Vibration data samples from any type of training samples form a first set, and each element in the first set is labeled as unprocessed; Randomly select an element marked as unprocessed from the first set as the first sample, take the collection time of the first sample as the first time, and take the collection location of the first sample as the first location; The elements in the first set that are not the first sample are identified as the first sample location and have the smallest time difference between the sampling time and the first time location. The time difference between the sampling time of the second sample and the first time location is taken as the first time difference. Determine the average time difference of the acquisition time of all vibration data samples acquired at the first position in any class of training samples; The sum of the first time difference and the average time difference is determined as the second time difference; Elements in the first set whose time difference between the collection time and the first time is no greater than the second time difference are selected as the third sample. The third set consists of samples whose time difference between the collection time and the first time is no greater than the first time difference; the remaining third samples are then combined into a third set. Determine the first similarity between each element in the second set and the first sample, and determine the second similarity between each element in the third set and the first sample; Based on the first similarity and the second similarity, the weight values ​​of the processed samples and each processed sample are determined, and the label of the processed sample is updated to "processed". The step of randomly selecting an element labeled as unprocessed from the first set as the first sample and subsequent steps are repeated until all elements in the first set are labeled as "processed".

2. The method according to claim 1, characterized in that, The step of determining the processed sample based on the first similarity and the second similarity includes: Determine the first minimum, first maximum, and first mean of all first similarities; Determine the second minimum, second maximum, and second mean of all second similarities; The first benchmark value is determined as (first minimum value + first maximum value + 4 × first mean value) / 6; The second benchmark value is determined as (second minimum value + second maximum value + 4 × second mean value) / 6; The center position of the acquisition location corresponding to the first similarity score which is not greater than the first benchmark value is determined as the first position; the center position of the acquisition location corresponding to the second similarity score which is not greater than the second benchmark value is determined as the second position; Determine whether the distance between the first position and the second position is not greater than the maximum distance between the first position and the acquisition position corresponding to the first similarity value which is not greater than the first benchmark value; If the distance between the first position and the second position is not greater than the maximum distance between the first position and the acquisition position corresponding to the first similarity not greater than the first benchmark value, then the first sample, as well as all vibration data samples corresponding to the first similarity not less than the first benchmark value and all vibration data samples corresponding to the second similarity not less than the second benchmark value, are all determined as processing samples.

3. The method according to claim 2, characterized in that, After determining whether the distance between the first position and the second position is not greater than the maximum distance between the first position and the acquisition position corresponding to the first similarity value which is not greater than the first benchmark value, the method further includes: If the distance between the first position and the second position is greater than the maximum distance between the first position and the acquisition position corresponding to the first similarity value which is not less than the first benchmark value, then the first sample and the vibration data sample with a similarity value not greater than the first benchmark value are both determined as processing samples.

4. The method according to claim 1, characterized in that, The weight values ​​of each processed sample are determined through the following steps: All processed samples are grouped into a fourth set; Calculate the third similarity between any two elements in the fourth set; The normalized value of the difference between each third similarity and the minimum value among the third similarities is determined as the weight value.

5. The method according to claim 1, characterized in that, The high-altitude glacier permafrost earthquake chain disaster simulation device includes: an environmental control chamber and a shaking table array; The environmental control cabin includes a low-temperature control system, a rainfall unit, a snowfall unit, a strong radiation simulation unit, a humidity control unit, and a wind speed control unit. A shaking table array consists of multiple shaking tables; The environmental control cabin is used to simulate the extreme and complex climate environment at high altitudes; The shaking table array is used to simulate seismic dynamic output.

6. The method according to claim 5, characterized in that, The cryogenic control system includes: an environmental control chamber temperature control unit and a model freeze-thaw temperature control unit; The environmental control chamber temperature control unit uses liquid nitrogen circulation refrigeration technology to control the temperature. The temperature control range of the environmental control chamber temperature control unit is -40 degrees Celsius to 60 degrees Celsius, and the temperature control accuracy is ±1 degree Celsius. The model freeze-thaw temperature control unit uses a high-power cold bath box to control a unidirectional heat transfer metal cold plate. The model temperature is controlled by the cold plate covering it. The temperature control range of the cold plate is -40℃ to 20℃, and the temperature control accuracy is ±1℃. The rainfall unit includes water pipes and nozzles, with a rainfall capacity of 0~200 mm / h and a rainfall uniformity of >90%. The snowmaking unit includes a snowmaking machine, which operates normally in an environment of -40℃ to 20℃, with a snowmaking capacity of 0 to 20 mm / h and a snowmaking uniformity of >90%. The humidity control unit uses ultrasonic humidification technology and a humidity sensor detection system for real-time feedback and adjustment. The humidity control range is 10% to 90%, and the humidity control accuracy is ±3%. The wind speed control unit includes a high-power fan, with a wind speed control range of 0 to 8 levels and a wind speed control accuracy within ±1%. The high-intensity radiation simulation unit includes a radiation panel integrated into the inner wall of the environmental control cabin. The radiation panel is used to simulate strong sunlight and ultraviolet radiation at high altitudes, with a simulation range of 0–1500 watts / m². 2 The accuracy of the strong radiation simulation is ±1%.

7. The method according to claim 5, characterized in that, The vibration table is a three-axis, six-degree-of-freedom vibration table, which uses an electro-hydraulic servo actuator and loads one or more of the following: sine wave, random wave, and seismic wave. The maximum accelerations of the vibration table in the three directions are ±2.0 g, ±2.0 g, and ±1.5 g, respectively; the displacements in the three directions are ±500 mm, ±500 mm, and ±300 mm, respectively; and the frequency range is 0.1~60 Hz.

8. The method according to claim 5, characterized in that, The vibration table array is composed of three vibration tables integrated together; One of the vibration tables is 6m×8m and has a full load of 100 tons, and the other two vibration tables are 3m×4m and have a full load of 40 tons. The vibration tables in the vibration table array are integrated into a single platform; the integrated vibration table array is 8m×9m and fully loaded with 180t, with maximum accelerations in three directions of ±1.8g, ±1.8g, and ±1.3g, displacements in three directions of ±500mm, ±500mm, and ±300mm, and a frequency range of 0.1~60Hz. or, The vibration tables in the vibration table array move along tracks in a linear arrangement; wherein, in the linearly arranged vibration table array, the distance between three vibration tables is 12m to 150m, and the movement range is 138m.

Images