An avalanche dynamic early warning method and system based on an avalanche starting physical model

Abstract

The application provides an avalanche dynamic early warning method and system based on an avalanche starting physical model, belongs to the technical field of avalanche prediction, and acquires and preprocesses collection parameters, wherein the collection parameters include background parameters, influence parameters and trigger parameters, data correlation is performed on each parameter, when different data appears, different calibration is performed, and interference data is removed; physical model construction is performed on the preprocessed data, the constructed physical model is trained, the physical model calculation weight is optimized, and the trained physical model is obtained; the next round of collection parameters are sent to the physical model, an early warning result is output, and early warning time and an avalanche range are distributed.

Description

TECHNICAL FIELD

[0001] The present specification relates to the technical field of avalanche prediction, in particular to an avalanche dynamic early warning method and system based on an avalanche initiation physical model. BACKGROUND

[0002] As one of the most active and destructive natural disasters in mountainous areas, avalanches, similar to landslides and debris flows, have the characteristics of potentiality and suddenness, and are often difficult to accurately predict. The occurrence of avalanche disasters not only directly threatens local traffic safety, but also can cause serious consequences such as damage to infrastructure, casualties and livestock losses. According to statistics, the direct economic losses caused by avalanches in the world in the past decade have an average annual growth rate of more than 15%, and the misjudgment rate of avalanche early warning in some high-altitude areas is as high as 30% or more, which seriously restricts the efficiency of disaster prevention and mitigation. The accuracy and timeliness of avalanche monitoring are directly related to the reliability of the early warning system, and existing monitoring technologies rely on meteorological data and multi-element time series data generated by physical sensor networks. Such data has the characteristics of high dimension, strong spatio-temporal dependence, and nonlinear dynamic evolution of abnormal patterns, and traditional analysis methods often fail to effectively capture key risk signals, resulting in delayed or misjudged early warnings, which is a weak link in the avalanche disaster prevention and control system.

[0003] The existing Chinese patent, with the name of an avalanche monitoring and early warning method based on deep learning, CN119782772A, includes: obtaining environmental monitoring data in the past period of time and preprocessing, constructing a training data set; constructing an avalanche monitoring and early warning model and training; inputting the forecast data of environmental monitoring into the trained avalanche monitoring and early warning model for data reconstruction, obtaining the corresponding reconstructed data; calculating the data reconstruction error; determining the data anomaly according to the error and the preset threshold, completing the avalanche monitoring and early warning. The present invention solves the problem that the anomaly detection model is difficult to capture key features under high-dimensional and complex dependent data, improves the accuracy and robustness of the anomaly detection model, and is especially suitable for processing multi-element time series data such as avalanche monitoring data, capturing data anomalies in avalanche monitoring data, and realizing avalanche monitoring and early warning.

[0004] The above technology aims to calculate the early warning time of the avalanche by obtaining the avalanche monitoring data, capture key features, and optimize the accuracy of anomaly detection. However, due to the wide sampling data of avalanches and too many interference factors, the subsequent calculation accuracy is easily reduced.

[0005] Therefore, it is necessary to provide an avalanche dynamic early warning method and system based on an avalanche initiation physical model, which can clean the collected data in advance and eliminate non-associated interference data. SUMMARY

[0006] To solve the above technical problems, the embodiment of the present specification provides an avalanche dynamic early warning method and system based on an avalanche starting physical model.

[0007] In some embodiments, an avalanche dynamic early warning method based on an avalanche starting physical model, pre-processes collected parameters, the collected parameters including background parameters, influence parameters and trigger parameters, data correlation is performed on each parameter, when there is differential data, differential calibration is performed, and interference data is removed;

[0008] The pre-processed data is subjected to physical model construction, the constructed physical model is trained, the physical model calculation weight is optimized, and the trained physical model is obtained;

[0009] The next round of collected parameters is sent to the physical model, and the early warning result is output.

[0010] Further, the background parameters include rock-soil slope gradient, shear stress and ecological environment, the influence parameters include ice-snow slope surface shear strength, snowfall and temperature, and the trigger parameters include snowfall, wind, biological activity and accidental factor, wherein the rock-soil slope gradient, shear stress, temperature, snowfall and wind can be directly obtained through a single data source, the ecological environment, ice-snow slope surface shear strength, biological activity and accidental factor are calculated through at least one data source, and the background parameters, influence parameters and trigger parameters are all based on the data output by the data source arranged in the to-be-tested snowfield.

[0011] Further, the background parameters, influence parameters and trigger parameters are applied in different levels in the physical model, multiple levels successively transmit data, and the weight of the background parameters, influence parameters and trigger parameters in the corresponding level successively increases.

[0012] Further, the differential data is marked based on three types of parameters of the collected parameters, when one parameter is different from the other two types of parameters, differential line calibration is performed, including manually selecting or removing the differential parameter, when the differential parameter is selected, the other two types of parameters are removed.

[0013] Further, the interference data includes weather trends that do not conform to most collected data, the collected data is obtained based on various types of weather sensors, the collected data is classified based on the obtained weather types, and the most collected data refers to at least three quarters of all source data under the same weather type.

[0014] Further, the physical model includes an analysis layer, a cross-learning layer and an output layer,

[0015] The analysis layer includes substituting the pre-processed data into a time domain analyzer, the time domain analyzer calculates the change trend of the parameters in the last time period as the predicted change trend in the next time period.

[0016] The cross-learning layer superimposes multiple types of predicted change trends on each other, labels the same change trend and the opposite change trend for deep learning training, and records the type weight corresponding to the two change trends;

[0017] According to the cross-learning results of multiple types, the avalanche early warning level and early warning time of the monitored area are output.

[0018] Also includes an avalanche dynamic early warning system based on an avalanche starting physical model, including a collection unit, a transmission unit, a calculation unit and an early warning unit,

[0019] The collection unit includes sensing devices for obtaining each weather type, each weather type is associated with at least one collection parameter, and each collection parameter corresponds to at least one sensing device;

[0020] The transmission unit includes sending the collected data on the sensing device to the calculation unit through a wireless transmission channel;

[0021] The calculation unit is a calculation server arranged in the monitored snowfield, and the calculation server calculates the avalanche early warning level and early warning time by calling historical collection parameters and real-time collection data;

[0022] The early warning unit sends a short message to the user's mobile phone bound to the snowfield signal tower with the early warning level and early warning time.

[0023] Further, the calculation server is provided with a calculation software, the input end of the calculation software is the data sent by the transmission unit, and the output end of the calculation software is the early warning unit.

[0024] The beneficial effects of the present application are:

[0025] 1. The data collected from the monitored snowfield is optimized, the interference data is eliminated, and the accuracy of the avalanche occurrence time and collapse range is improved;

[0026] 2. The data sources for predicting avalanches are divided into multiple categories, the weight of the data sources is given according to the influence law of historical data on the climate, and multiple physical models are constructed, which is convenient for directly predicting the time and range of the avalanche in the future. BRIEF DESCRIPTION OF DRAWINGS

[0027] This specification will be further illustrated in the form of exemplary embodiments, which will be described in detail through the drawings. These embodiments are not limiting, and in these embodiments, the same numbers represent the same structures, wherein:

[0028] Figure 1 is a schematic diagram of the working principle of the collection parameter according to the embodiment of the present specification;

[0029] Figure 2is a working schematic diagram of a sensor according to some embodiments of the present specification. DETAILED DESCRIPTION

[0030] To more clearly illustrate the technical solutions of the embodiments of the present specification, a brief introduction will be given to the drawings needed in the embodiment description. Obviously, the drawings in the following description are only some examples or embodiments of the present specification, and for those skilled in the art, the present specification can be applied to other similar scenarios without creative labor on the basis of these drawings. Unless it is clear from the language environment or otherwise stated, the same reference numbers in the drawings represent the same structures or operations.

[0031] It should be understood that the "system", "device", "unit" and / or "module" used herein is a method for distinguishing different components, elements, parts, sections or assemblies at different levels. However, if other words can achieve the same purpose, the words can be replaced by other expressions.

[0032] As shown in the specification and claims, unless the context clearly indicates otherwise, the words "one", "a", "an", and / or "the" do not refer to the singular, but can also include the plural. Generally speaking, the terms "comprise" and "include" only indicate the inclusion of the steps and elements explicitly identified, and these steps and elements do not constitute an exclusive list, and the method or device can also include other steps or elements.

[0033] Flowcharts are used in the present specification to illustrate the operations performed by the system according to the embodiments of the present specification. It should be understood that the preceding or subsequent operations are not necessarily performed in sequence. On the contrary, each step can be processed in reverse order or simultaneously. At the same time, other operations can be added to these processes, or one or more steps of the operation can be removed from these processes.

[0034] Embodiments:

[0035] Please refer to Figure 1 The background parameters include the slope of the rock-soil slope, shear stress, ecological environment, the influence parameters include the shear strength of the ice and snow slope, snowfall, temperature, and the trigger parameters include snowfall, wind, biological activity and accidental factors, wherein the slope of the rock-soil slope, shear stress, temperature, snowfall and wind can be directly obtained through a single data source, the ecological environment, the shear strength of the ice and snow slope, biological activity and accidental factors are calculated through at least one data source, and the background parameters, influence parameters and trigger parameters are all based on the data output by the data source arranged in the snowfield to be measured.

[0036] Slope is the angle between the slope and the horizontal plane, which directly affects the gravity component (downward force). The steeper the slope, the higher the avalanche risk. Slope values ​​for slope areas are extracted using GIS software using digital terrain models (DTMs) or LiDAR data. Risk levels are categorized by slope gradient: <25° is low risk (snowpack tends to stabilize); 25° to 45° is medium-to-high risk (typical avalanche-prone slopes); and >45° is extremely high risk (slopes prone to avalanches).

[0037] Shear stress is the tangential stress in the rock mass of a slope along a potential sliding surface. When shear stress exceeds the shear strength, it can trigger a landslide or avalanche (by triggering snowpack instability). Embedding stress gauges (such as vibrating wire sensors) on potential sliding surfaces directly measures shear stress changes in real time.

[0038] The ecological environment indirectly affects avalanche risk through vegetation coverage, root anchorage capacity, terrain roughness, etc.

[0039] Vegetation snow-fixing capacity: Calculate vegetation cover (VFC) using UAV multispectral imagery or ground surveys, using pixel dichotomy to estimate:

[0040]

[0041] Impact of vegetation type: Coniferous forests (such as pine trees) have better branches and snow-retention capabilities than grasslands or shrubs.

[0042] Terrain roughness: Use DEM to calculate surface roughness index (such as standard deviation, coefficient of variation). The higher the roughness, the more stable the snowpack.

[0043] Shear strength is the ultimate stress at which snowpack resists shear failure and is closely related to the snow layer's structure, temperature, and humidity. Regional snowfall is estimated using microwave radiometers (such as AMSR-E) or radar (such as spaceborne SAR). Embedded temperature sensors (such as thermocouples) measure temperatures at different depths in the snow layer. A greater gradient (e.g., surface temperature > deep temperature) increases the likelihood of a weak layer (such as molasses snow) forming, reducing shear strength.

[0044] Biological activity includes field surveys to measure the density of animal activity traces, assess disturbances to the snow structure (e.g., rodent burrows may become avalanche triggers), and calculate root volume density using ground penetrating radar (GPR) or root excavation methods. The denser the root system, the greater the snow-fixing capacity.

[0045] Random factors include human activities, which are used to record the frequency of skiers, climbers, etc. entering high-risk areas, statistics through GPS tracks or surveillance cameras, and earthquake magnitudes obtained using earthquake monitoring networks ( M ) and epicentral distance ( R ), the probability of triggering an avalanche is estimated by the empirical formula: ,( a , b , c For area fitting parameters).

[0046] The physical model in this embodiment is constructed as follows: ARI =∑( wi × xi ),( wi is the weight, xi is the normalized index value). For example, slope accounts for 30%, shear strength accounts for 25%, snowfall accounts for 20%, etc. A multi-level model is obtained by stacking random forests, neural networks, etc.

[0047] The specific monitoring devices used in this embodiment are shown in Table 1,

[0048] Table 1 Partial monitoring device types

[0049]

[0050] It is worth noting that please refer to Figure 2 , due to the existence of multiple same weather data, repeated application in different parameters, interference data including not in line with the majority of the collected data (such as Figure 2 CG1-CG6) calculated weather trend, the collected data is based on various types of weather sensor to obtain data, based on the type of weather obtained, the collected data is classified by source, the majority of the collected data refers to at least three quarters of all source data under the same weather type. The difference data is marked based on three types of parameters of the collection parameters, when one parameter is different from the other two types of parameters, the difference row calibration is performed, including manually selecting or excluding the difference parameter, when the difference parameter is selected, the other two types of parameters are excluded.

[0051] In summary, the various types of collected data of the snowfield to be monitored are classified and divided into three major parameters in this embodiment, which are used as training parameters for constructing a physical model, and the various levels of the physical model are optimized to obtain a time and range of snow avalanche that can be accurately output.

[0052] The computer readable storage medium includes, for example, a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various media that can store program codes.

[0053] Based on the same inventive concept, the embodiment of the present application provides an engineering cost progress management system, comprising a memory and a processor, and the memory stores a program capable of realizing any method on the processor.

[0054] Those skilled in the art can clearly understand that, for the convenience and brevity, only the division of the above functional modules is taken as an example for description, and in actual application, the above functions can be completed by different functional modules according to needs, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. The specific working process of the above-described system, device and unit can refer to the corresponding process in the foregoing method embodiments, which will not be repeated here.

[0055] In several embodiments provided in the present application, it should be understood that the disclosed system, device and method can be implemented in other ways. For example, the above-described device embodiments are only schematic, for example, the division of the modules or units is only a logical function division, and actual implementation can have another division manner, for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the units shown or discussed can be indirect coupling or communication connection through some interface, device or unit, and can be electrical, mechanical or other forms.

[0056] The units described as separate components can or can not be physically separate, and the components shown as units can or can not be physical units, that is, they can be located in one place, or can be distributed on a plurality of network units. Part or all of the units can be selected according to actual needs to achieve the purpose of the embodiment scheme.

[0057] In addition, each functional unit in each embodiment of the present application can be integrated in one processing unit, or each unit can be physically present separately, or two or more units can be integrated in one unit. The above integrated unit can be realized in the form of hardware or in the form of software functional unit.

[0058] The integrated unit, if implemented in the form of a software function unit and sold or used as an independent product, can be stored in a computer-readable storage medium. Based on such an understanding, the technical solutions of the present application, essentially or in other words, the part that contributes to the prior art or the whole or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium, and includes a number of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) or a processor to execute all or part of the steps of the methods described in the various embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory, a random access memory, a magnetic disk or an optical disk, and various media that can store program codes.

[0059] The embodiments of the specific implementation are the preferred embodiments of the present application, not limited to the protection scope of the present application, so: any equivalent changes made according to the structure, shape, principle of the present application should be covered within the protection scope of the present application.

Claims

1. A dynamic avalanche warning method based on avalanche initiation physical model, characterized in that: Acquisition parameters are obtained for preprocessing. The acquisition parameters include background parameters, influencing parameters, and trigger parameters. Data association is performed on each parameter. When discrepant data appears, discrepancy calibration is performed to eliminate interference data. Interference data includes weather trends that do not conform to the calculated majority of collected data. The collected data is acquired based on various types of weather sensors. The collected data is classified based on the acquired weather type. The majority of collected data refers to data that accounts for at least three-quarters of all source data under the same weather type. Construct a physical model for the preprocessed data, train the constructed physical model, optimize the physical model calculation weights, and obtain a trained physical model; Send the new round of collected parameters to the physical model and output the early warning results.

2. The avalanche dynamic warning method based on the avalanche initiation physical model according to claim 1, characterized in that: Background parameters include the slope of the rock and soil slope, shear stress, and ecological environment; influencing parameters include the shear strength of the ice and snow slope, snowfall, and temperature; triggering parameters include snowfall, wind, biological activity, and accidental factors. Among them, the slope of the rock and soil slope, shear stress, temperature, snowfall, and wind can be directly obtained through a single data source, and the ecological environment, shear strength of the ice and snow slope, biological activity, and accidental factors are calculated through at least one data source. The background parameters, influencing parameters, and triggering parameters are all based on data output by the data source set in the snow area to be tested.

3. The avalanche dynamic warning method based on the avalanche initiation physical model according to claim 2, characterized in that: Background parameters, impact parameters and trigger parameters are applied at different levels in the physical model. Multiple levels transmit data in sequence, and the weights of background parameters, impact parameters and trigger parameters in the corresponding levels increase in sequence.

4. The avalanche dynamic warning method based on the avalanche initiation physical model according to claim 3, characterized in that: The difference data is marked based on the three categories of parameters of the acquisition parameters. When one of the parameters differs from the parameters of the other two categories, difference calibration is performed, including manually selecting or eliminating the difference parameter. When the difference parameter is selected, the parameters of the other two categories are eliminated.

5. A method for dynamic avalanche warning based on a physical model of avalanche initiation as claimed in claim 4, characterized in that: The physical model includes analysis layer, cross-learning layer and output layer. The analysis layer includes substituting the pre-processed data into the time domain analyzer, which calculates the change trend of the parameters in the previous time period as the expected change trend in the next time period; The cross-learning layer superimposes multiple types of expected change trends, marks the same change trends and opposite change trends for deep learning training, and records the type weights corresponding to the two change trends; Based on the cross-learning results of multiple types, the avalanche warning level and warning time of the area to be monitored are output.

Images