Silo internal state inversion early warning method, device, equipment, medium and product

Abstract

The application discloses a silo internal state inversion early warning method, device, equipment, medium and product, relates to the technical field of data processing, and the method comprises the following steps: dividing various sensor data into a conventional environment data set, a conventional rainfall environment data set and a rainfall-temperature difference coupling data set; processing the conventional environment data set, the conventional rainfall environment data set and the rainfall-temperature difference coupling data set to obtain a fusion data set; inputting the fusion data set into a pre-constructed internal state inversion model to obtain internal state parameters; determining the internal state risk coefficient of the silo based on the internal state parameters; and determining the internal state risk level of the silo and the early warning scheme corresponding to the risk level based on the internal state risk coefficient and a preset safety threshold. The application enables the silo management to anticipate potential risks in advance, timely take effective measures, effectively avoid the occurrence of safety accidents, and improves the safety of the silo management.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and in particular to a method, apparatus, equipment, medium and product for inverting and early warning of the internal state of a silo. Background Technology

[0002] Against the backdrop of the deep integration of the digital economy and the real economy, intelligent operation and maintenance of industrial warehousing facilities has become a core direction for improving production efficiency and safety levels. Among them, steel structure silos, as key storage equipment for bulk materials such as grain and building materials, directly affect the smooth operation of the industrial chain and supply chain due to their structural safety and operational stability. Currently, steel structure silos are in a complex environment for a long time and are affected by multiple factors such as material load, natural rainfall, and temperature and humidity changes. Their surface condition and internal structure are prone to hidden deterioration. If timely monitoring and early warning are not provided, it may lead to major accidents such as collapse and material leakage. Therefore, the demand for accurate perception of silo status and risk early warning is becoming increasingly urgent.

[0003] Currently, the main existing technologies used in the field of silo condition monitoring and early warning include traditional manual inspection and single-sensor monitoring methods. Traditional manual inspection usually involves recording the surface condition of the silo through periodic on-site inspections and handheld device testing, relying on the experience of maintenance personnel to judge potential risks. Single-sensor monitoring involves deploying strain gauges, pressure sensors, and other devices in local areas of the silo to collect data such as stress and vibration at single points, and triggering early warnings through simple threshold comparisons. In addition, some scenarios adopt an offline data analysis mode, where the collected data is uploaded to the cloud for batch processing and then a status report is generated.

[0004] However, in practical use, it still has shortcomings: traditional manual inspection is limited by long cycles and strong subjectivity, making it difficult to capture real-time changes in the silo's condition, and there are blind spots in the monitoring of hidden areas, which can easily lead to missed risk assessments; single sensor monitoring can only obtain local data and lacks multi-dimensional information correlation. For example, it only monitors stress values ​​while ignoring the impact of rainfall, which cannot fully reflect the overall condition of the silo. Simple threshold comparison is also difficult to adapt to dynamic risk changes in complex environments, resulting in low early warning accuracy. When the silo ages or the material characteristics change, the matching degree between the early warning threshold and the actual risk decreases, resulting in delayed early warnings or false alarms, which cannot guarantee the safety of silo management. Summary of the Invention

[0005] The purpose of this application is to provide a method, device, equipment, medium, and product for inverting and early warning of the internal state of a silo, which can improve the security of silo management.

[0006] To achieve the above objectives, this application provides the following solution:

[0007] Firstly, this application provides a method for early warning based on the internal state of a silo, including:

[0008] The collected sensor data is divided into a conventional environmental dataset, a conventional rainfall environmental dataset, and a rainfall-temperature difference coupled dataset; wherein, the various sensor data are collected by various types of sensors pre-installed on the silo;

[0009] The conventional environment dataset, the conventional rainfall environment dataset, and the rainfall-temperature difference coupled dataset are processed to obtain a fused dataset;

[0010] The fused dataset is input into a pre-built internal state inversion model to obtain the internal state parameters output by the internal state inversion model;

[0011] The internal state risk coefficient of the silo is determined based on the internal state parameters.

[0012] Based on the internal state risk coefficient and the preset safety threshold, the internal state risk level of the silo and the corresponding early warning scheme are determined.

[0013] Optionally, the step of dividing the collected data from various sensors into a conventional environmental dataset, a conventional rainfall environmental dataset, and a rainfall-temperature difference coupled dataset specifically includes:

[0014] The core surface parameters and the corresponding environmental parameters are obtained from the collected data from various sensors; wherein, the environmental parameters include rainfall, temperature, humidity and temperature fluctuation values.

[0015] Based on the environmental parameters, the core surface parameters are divided into a regular environment dataset, a regular rainfall environment dataset, and a rainfall-temperature difference coupled dataset. In the regular environment dataset, the rainfall corresponding to the core surface parameters is less than a preset rainfall amount; in the regular rainfall environment dataset, the rainfall corresponding to the core surface parameters is greater than or equal to the preset rainfall amount, and the temperature fluctuation value corresponding to the core surface parameters is less than a preset fluctuation threshold; in the rainfall-temperature difference coupled dataset, the rainfall corresponding to the core surface parameters is greater than or equal to the preset rainfall amount, and the temperature fluctuation value corresponding to the core surface parameters is greater than or equal to the preset fluctuation threshold.

[0016] Optionally, the process of processing the conventional environmental dataset, the conventional rainfall environmental dataset, and the rainfall-temperature difference coupled dataset to obtain a fused dataset specifically includes:

[0017] Determine the normal rainfall deviation value between the normal rainfall environment dataset and the normal environment dataset;

[0018] Determine the rainfall-temperature difference coupling deviation value between the rainfall-temperature difference coupling dataset and the normal rainfall environment dataset;

[0019] Based on the conventional environmental dataset, the conventional rainfall environmental dataset, and the rainfall-temperature difference coupled dataset, characteristic difference indicators are determined;

[0020] Based on the aforementioned feature difference index, the conventional rainfall environment dataset and the rainfall-temperature difference coupled dataset are processed respectively to obtain the rainfall influence coefficient and the rainfall-temperature fluctuation joint influence coefficient.

[0021] The conventional environmental dataset, the conventional rainfall deviation value, the rainfall-temperature difference coupling deviation value, the rainfall influence coefficient, and the rainfall-temperature fluctuation joint influence coefficient are fused to obtain a fused dataset.

[0022] Optionally, the step of processing the conventional rainfall environment dataset and the rainfall-temperature difference coupled dataset based on the feature difference index to obtain the rainfall influence coefficient and the rainfall-temperature fluctuation joint influence coefficient specifically includes:

[0023] Based on the aforementioned feature difference index, a univariate linear fit is performed on the conventional rainfall environment dataset to obtain the first regression equation;

[0024] Based on the aforementioned feature difference index, a binary linear fit is performed on the rainfall-temperature difference coupled dataset to obtain the second regression equation;

[0025] Determine the rainfall impact coefficient from the first regression equation;

[0026] The combined influence coefficient of rainfall and temperature fluctuations is determined from the second regression equation.

[0027] Optionally, the internal state parameters include material moisture content, stacking skew angle, weld health, rainwater infiltration rate, and environmental erosion rate. Determining the internal state risk coefficient of the silo based on these internal state parameters specifically includes:

[0028] The material moisture content, the stacking skew angle, the weld health, the rainwater infiltration rate, and the environmental erosion rate are normalized respectively to obtain the target material moisture content, the target stacking skew angle, the target weld health, the target rainwater infiltration rate, and the target environmental erosion rate.

[0029] The first weight for determining the material moisture content, the second weight for the stacking skew angle, the third weight for the weld health, the fourth weight for the rainwater infiltration rate, and the fifth weight for the environmental erosion rate are determined.

[0030] The internal state risk coefficient of the silo is calculated based on the target material moisture content, the target stacking skew angle, the target weld health, the target rainwater infiltration rate, the target environmental erosion rate, the first weight, the second weight, the third weight, the fourth weight, and the fifth weight.

[0031] Optionally, the preset safety threshold includes a first safety threshold and a second safety threshold, wherein the first safety threshold is less than the second safety threshold. The step of determining the internal state risk level of the silo and the corresponding early warning scheme based on the internal state risk coefficient and the preset safety threshold specifically includes:

[0032] If the internal state risk coefficient is less than the first safety threshold, the internal state risk level of the silo is determined to be low risk level, and a low risk warning scheme corresponding to the low risk level is obtained.

[0033] If the internal state risk coefficient is greater than or equal to the first safety threshold and less than the second risk level, then the internal state risk level of the silo is determined to be a medium risk level, and a medium risk early warning scheme corresponding to the medium risk level is obtained.

[0034] If the internal state risk coefficient is greater than or equal to the second risk level, then the internal state risk level of the silo is determined to be a high-risk level, and a high-risk early warning scheme corresponding to the high-risk level is obtained.

[0035] Secondly, this application provides a silo internal state inversion early warning device, comprising:

[0036] A partitioning unit is used to divide the collected data from various sensors into a conventional environmental dataset, a conventional rainfall environmental dataset, and a rainfall-temperature difference coupled dataset; wherein, the various sensor data are collected by various types of sensors pre-installed on the silo;

[0037] The processing unit is used to process the conventional environment dataset, the conventional rainfall environment dataset, and the rainfall-temperature difference coupled dataset to obtain a fused dataset;

[0038] The input unit is used to input the fused dataset into a pre-built internal state inversion model to obtain the internal state parameters output by the internal state inversion model.

[0039] The first determining unit is used to determine the internal state risk coefficient of the silo based on the internal state parameters.

[0040] The second determining unit is used to determine the internal state risk level of the silo and the early warning scheme corresponding to the risk level based on the internal state risk coefficient and the preset safety threshold.

[0041] Thirdly, this application provides a computer device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the silo internal state inversion early warning method described in any one of the above.

[0042] Fourthly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the silo internal state inversion early warning method described above.

[0043] Fifthly, this application provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the silo internal state inversion early warning method described above.

[0044] In a sixth aspect, this application provides a chip, the chip including a processor and a communication interface, the communication interface being coupled to the processor, the processor being used to run a program or instructions, and the processor executing the program or instructions implementing the steps of the silo internal state inversion early warning method described above.

[0045] According to the specific embodiments provided in this application, the following technical effects are disclosed:

[0046] This application provides a method, device, equipment, medium, and product for silo internal state inversion and early warning. By meticulously dividing and fusing data collected from multiple sensors into different environmental datasets, it provides a comprehensive and accurate data foundation for internal state analysis. Using a pre-built internal state inversion model, internal state parameters can be accurately derived from the input fused dataset, thereby determining a reasonable internal state risk coefficient. Based on a comparison of this risk coefficient with a preset safety threshold, the risk level of the silo's internal state can be accurately determined, and a corresponding early warning scheme can be quickly matched. Therefore, this approach enables silo management to anticipate potential risks and take timely and effective measures to effectively prevent safety accidents, thus improving the safety of silo management. Attached Figure Description

[0047] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0048] Figure 1 A flowchart illustrating a method for inverting and warning about the internal state of a silo, provided in an embodiment of this application;

[0049] Figure 2 A schematic diagram of the functional modules of a silo internal state inversion and early warning device provided in an embodiment of this application;

[0050] Figure 3 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0051] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0052] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0053] In one exemplary embodiment, such as Figure 1 As shown, a method for inverting and warning about the internal state of a silo is provided. This method is executed by a computer device, specifically by a terminal or server alone, or by both a terminal and a server. In this embodiment, it includes the following steps 101 to 105. Wherein:

[0054] Step 101: Divide the collected data from various sensors into a regular environmental dataset, a regular rainfall environmental dataset, and a rainfall-temperature difference coupled dataset.

[0055] In this embodiment, the various sensor data are collected by multiple types of sensors pre-installed on the silo. Furthermore, the data can be collected under a fixed temperature environment, defined as a temperature range of 20-25°C. The various types of sensors can be installed in key areas of the silo surface; these key areas include the top load-bearing area, side wall welds, and the bottom material contact area. The specific data collection scheme is as follows:

[0056] In the top load-bearing area, surface-mounted strain sensors are used to measure the stress state of the area by attaching them to the surface of the steel plate. The sampling frequency is set to 10Hz to ensure that instantaneous stress changes can be captured. At the same time, force sensors are arranged at key load-bearing parts of the silo top structure. By deploying multiple points, a monitoring network is formed to monitor the dynamic changes of force at each load-bearing point in real time.

[0057] In addition, for monitoring the side wall welds, surface-mounted distributed fiber optic sensors with a fiber diameter of 0.9mm are used to achieve full-process monitoring of the welds through optical time-domain reflectometry. Combined with surface-mounted ultrasonic sensors, non-destructive testing can be performed directly on the weld surface without damaging the steel structure, allowing for in-depth detection of the weld's internal integrity and timely detection of internal defects or micro-cracks. The temperature sensor uses a platinum resistance sensor with a sampling interval of 1 minute, simultaneously collecting the surface temperature of the silo wall and the ambient temperature around the silo. Temperature data during rainfall is recorded, and the temperature fluctuation value is calculated every two consecutive hours starting from the start time of rainfall, i.e., the maximum temperature difference between adjacent two hours.

[0058] A surface array pressure sensor is installed in the bottom material contact area (i.e., the bottom of the silo). A heat insulation buffer layer is set between the sensor and the steel plate to prevent the sensor from being damaged by high temperature and friction of the coal. The sensor spacing is set to 50cm×50cm to obtain the pressure distribution at the bottom in real time. Since the movement and compaction of the material in the bottom contact area may cause vibration, a bonded triaxial vibration sensor is equipped with a sampling frequency of 1kHz to detect the dynamic changes in vibration caused by material movement and compaction in real time, thereby determining whether the material is over-compacted or unevenly piled.

[0059] In this embodiment of the application, two sets of controls can be set up: a normal environment and a rainfall environment. The rainfall environment can further include a normal rainfall environment and a rainfall-temperature difference coupled environment. Specifically:

[0060] "Routine environment" specifically refers to various monitoring data collected on key areas of the silo surface under fixed temperature conditions of 20-25℃ and rainfall of <10mm. Rainfall data is collected over a short period (e.g., within 24 hours).

[0061] Normal rainfall environment specifically refers to a rainfall of ≥10mm under a fixed temperature of 20-25℃ and an ambient temperature fluctuation of <8℃ within 2 hours during the rainfall;

[0062] Rainfall-temperature difference coupled environment: refers to a scenario where the rainfall is ≥10mm and the ambient temperature fluctuates by ≥8℃ within 2 hours during the rainfall.

[0063] Among them, the rainfall-temperature difference coupled environment mostly corresponds to rainstorm weather. Because rainstorms are characterized by concentrated short-term rainfall and significant evaporative heat absorption effects, they can quickly remove heat from the surface of steel structure silos, causing the silo wall temperature to drop significantly in a short period of time.

[0064] During actual data collection in a rainfall environment, when the rainfall is ≥10mm, the temperature and humidity sensors around the silo and the original platinum resistance temperature sensor are activated simultaneously: the temperature and humidity sensors record the changes in air temperature and humidity at 10-minute intervals to provide data support for subsequent analysis of the impact of temperature and humidity on the silo state under rainfall conditions; the platinum resistance temperature sensor maintains a sampling interval of 1 minute to simultaneously collect the surface temperature of the silo wall and the ambient temperature around the silo, and record the temperature data during the rainfall process.

[0065] Starting from the start time of rainfall, calculate the temperature fluctuation value for every consecutive 2 hours. Temperature fluctuation refers to the difference between the highest and lowest temperatures within any consecutive 2 hours during the rainfall process. If the temperature fluctuation is <8℃, the status data of key areas on the silo surface collected during this rainfall process will be marked as a regular rainfall environment dataset; if the temperature fluctuation is ≥8℃, it will be marked as a rainfall-temperature difference coupled dataset.

[0066] Specifically, the calculation method for temperature fluctuation values ​​is as follows;

[0067] Starting from the time when rainfall begins, calculate the maximum difference in surface temperature of the silo wall and the maximum difference in ambient temperature around the silo within each consecutive 2-hour period.

[0068] The maximum difference between the maximum temperature difference on the silo wall surface and the maximum temperature difference in the surrounding environment is selected as the current temperature fluctuation value for the current two consecutive hours.

[0069] Two types of rainfall environment datasets and a regular environment dataset collected under normal conditions are uploaded to the database via a synchronous transmission protocol and stored in a labeled manner as a regular environment dataset, a regular rainfall environment dataset, and a rainfall-temperature difference coupled dataset.

[0070] It is important to explain that the surface temperature of the silo wall is significantly reduced due to the heavy rain. A temperature difference of ≥8°C within 2 hours during the rainfall will cause significant thermal deformation of the steel structure. The steel plates of the silo wall and the welded joints shrink due to the sudden drop in temperature. The shrinkage rate varies in different parts. This non-uniform shrinkage will generate additional stress at the welded joints, which will increase the propagation rate of existing microcracks in the welded joints and increase the structural safety risk.

[0071] As an optional implementation, step 101, which divides the collected data from various sensors into a conventional environmental dataset, a conventional rainfall environmental dataset, and a rainfall-temperature difference coupled dataset, may include:

[0072] The core surface parameters and the corresponding environmental parameters are obtained from the collected data from various sensors; wherein, the environmental parameters include rainfall, temperature, humidity and temperature fluctuation values.

[0073] Based on the environmental parameters, the core surface parameters are divided into a regular environment dataset, a regular rainfall environment dataset, and a rainfall-temperature difference coupled dataset. In the regular environment dataset, the rainfall corresponding to the core surface parameters is less than a preset rainfall amount; in the regular rainfall environment dataset, the rainfall corresponding to the core surface parameters is greater than or equal to the preset rainfall amount, and the temperature fluctuation value corresponding to the core surface parameters is less than a preset fluctuation threshold; in the rainfall-temperature difference coupled dataset, the rainfall corresponding to the core surface parameters is greater than or equal to the preset rainfall amount, and the temperature fluctuation value corresponding to the core surface parameters is greater than or equal to the preset fluctuation threshold.

[0074] This implementation method precisely acquires core surface and environmental parameters from multiple sensor data points, providing a comprehensive basis for data segmentation. Based on key indicators such as rainfall and temperature fluctuations in the environmental parameters, the data is scientifically divided into three datasets. This refined segmentation more accurately reflects the silo status under different environmental conditions, making subsequent analysis more targeted. It helps to promptly identify potential risks to the silos under different environmental combinations, allowing for proactive preventative measures and ensuring the safe and stable operation of the silos.

[0075] In this embodiment, the core surface parameters may include parameters such as the stress variation amplitude of the top load-bearing area, the peak force of the force sensor, the crack propagation rate of the side wall weld, the uniformity coefficient of the bottom pressure distribution, and the vibration spectrum characteristics.

[0076] Step 102: Process the regular environment dataset, the regular rainfall environment dataset, and the rainfall-temperature difference coupled dataset to obtain a fused dataset.

[0077] As an optional implementation, step 102 may involve processing the conventional environmental dataset, the conventional rainfall environmental dataset, and the rainfall-temperature difference coupled dataset to obtain a fused dataset, including:

[0078] Determine the normal rainfall deviation value between the normal rainfall environment dataset and the normal environment dataset;

[0079] Determine the rainfall-temperature difference coupling deviation value between the rainfall-temperature difference coupling dataset and the normal rainfall environment dataset;

[0080] Based on the conventional environmental dataset, the conventional rainfall environmental dataset, and the rainfall-temperature difference coupled dataset, characteristic difference indicators are determined;

[0081] Based on the aforementioned feature difference index, the conventional rainfall environment dataset and the rainfall-temperature difference coupled dataset are processed respectively to obtain the rainfall influence coefficient and the rainfall-temperature fluctuation joint influence coefficient.

[0082] The conventional environmental dataset, the conventional rainfall deviation value, the rainfall-temperature difference coupling deviation value, the rainfall influence coefficient, and the rainfall-temperature fluctuation joint influence coefficient are fused to obtain a fused dataset.

[0083] This implementation method accurately captures data differences under various environmental conditions by determining the deviation values ​​between different datasets. Processing the dataset based on characteristic difference indicators yields rainfall and joint impact coefficients, enabling in-depth analysis of the degree of environmental factors' influence on silos. Integrating this key information with conventional environmental datasets into a fused dataset makes the data more comprehensive and representative, fully reflecting the actual state of silos under different environmental combinations. This facilitates early detection of potential risks, timely implementation of countermeasures, and ensures the safe and stable operation of silos.

[0084] In this embodiment of the application, the formula for calculating the conventional rainfall deviation value can be:

[0085]

[0086]

[0087] In this embodiment of the application, the formula for calculating the rainfall-temperature difference coupling deviation value can be:

[0088]

[0089]

[0090] Where W represents the rainfall-temperature difference coupling deviation value, This represents the absolute deviation value of the rainfall-temperature difference coupling. This represents the mean of the rainfall-temperature difference coupled dataset.

[0091] In this embodiment of the application, the specific calculation process of the feature difference index is as follows:

[0092] First, calculate the statistic:

[0093]

[0094] in, The sample size of the target dataset. For the sample size of the reference dataset, It is the sample mean of the target dataset. It is the sample mean of the reference dataset. It is the sample variance of the target dataset. It is the sample variance of the reference dataset; the target dataset includes a regular environmental dataset, a regular rainfall environmental dataset, and a rainfall-temperature difference coupled dataset.

[0095] Calculate the degrees of freedom again:

[0096]

[0097] After calculating the degrees of freedom, look up the critical value for that degree of freedom in a t-distribution table. Compare the calculated statistic t with the critical value. If the statistic t is greater than the critical value, then there is a statistically significant difference between the two sets of data. The t-distribution table lists the critical values ​​of the statistic t at different significance levels and degrees of freedom. The critical value is a threshold used to determine whether the calculated statistic t is large enough to conclude that the difference in the means of the two sets of data is not caused by random error, but is statistically significant.

[0098] Optionally, the methods for processing the conventional rainfall environment dataset and the rainfall-temperature difference coupled dataset based on the feature difference index to obtain the rainfall influence coefficient and the rainfall-temperature fluctuation joint influence coefficient may include:

[0099] Based on the aforementioned feature difference index, a univariate linear fit is performed on the conventional rainfall environment dataset to obtain the first regression equation;

[0100] Based on the aforementioned feature difference index, a binary linear fit is performed on the rainfall-temperature difference coupled dataset to obtain the second regression equation;

[0101] Determine the rainfall impact coefficient from the first regression equation;

[0102] The combined influence coefficient of rainfall and temperature fluctuations is determined from the second regression equation.

[0103] This implementation method involves performing a univariate linear fit on a typical rainfall dataset to accurately derive a first regression equation reflecting the single factor of rainfall, thereby determining the rainfall influence coefficient. A bivariate linear fit is then performed on a rainfall-temperature difference coupled dataset to obtain a second regression equation considering the combined effects of rainfall and temperature fluctuations, thus determining the joint influence coefficient. This approach scientifically quantifies the impact of different environmental factors on silos, making data fusion and analysis more targeted.

[0104] In this embodiment of the application, a quantitative relationship model of surface state parameters changing with rainfall is established:

[0105] For a typical rainfall dataset, first define a uniformity coefficient k, then use univariate linear fitting. Let the rainfall be x, and the value of k be y. The regression equation is:

[0106]

[0107] in, This represents the uniformity coefficient k value when the theoretical rainfall is 0. The average change in the uniformity coefficient k for every 1 mm increase in rainfall was obtained by statistical fitting of historical samples under normal rainfall conditions.

[0108] By calculating the specific regression equation for k, we can understand the specific relationship between surface state parameters and changes in rainfall.

[0109] A binary linear regression was performed on the rainfall-temperature difference coupled dataset, with rainfall x and temperature fluctuation z as variables and uniformity coefficient k as the dependent variable. The regression equation is as follows:

[0110]

[0111] The joint influence coefficient of rainfall and temperature fluctuation was obtained. This is the baseline deviation value when rainfall x = 0 mm and temperature fluctuation z = 0 °C. This refers to the individual influence coefficient of rainfall. This is the coefficient for the individual influence of temperature fluctuations, which is the sum of the coefficient for the individual influence of rainfall and the coefficient for the individual influence of temperature fluctuations.

[0112] Step 103: Input the fused dataset into the pre-built internal state inversion model to obtain the internal state parameters output by the internal state inversion model.

[0113] In this embodiment, the internal state parameters include material state parameters, structural safety parameters, and environmental interaction parameters; the material state parameters may include material moisture content and stacking skew angle; the structural safety parameters may include weld health; and the environmental interaction parameters may include rainwater infiltration rate and environmental erosion rate.

[0114] In this embodiment, the internal state inversion model takes a fused dataset as input and trains a lightweight improved random forest algorithm using historical labeled data. The historical labeled data contains corresponding samples of surface state parameters and internal states under different environments. The model learns the association rules between "abnormal surface parameters" and "deterioration of internal state" in real time at the edge. The model relies on incremental data for iterative optimization, that is, a fine-tuning is triggered every 100 sets of data. The output results are appended with confidence to achieve inversion analysis and output internal state parameters.

[0115] It should be noted that the lightweight improved random forest algorithm has the following improvement directions:

[0116] Feature dimensional pruning: Based on the five dimensions of the fused dataset, redundant features are removed by calculating mutual information entropy. When the mutual information entropy is >0.8, it is determined to be a redundant feature (e.g., when the rainfall is <5mm, the mutual information entropy between the rainfall influence coefficient and the conventional benchmark value is >0.8, it is determined to be a redundant feature and will not be included in the model input for the time being). The input feature dimension is dynamically compressed to reduce the computational load of edge nodes.

[0117] Optimization of the number of decision trees: Cross-validation is used to determine the optimal number of decision trees. While ensuring model accuracy (prediction accuracy ≥ 92%), the number of decision trees is controlled at 20-30. At the same time, a "leaf node sample size threshold" (minimum sample size ≥ 5) is introduced to avoid overfitting and computational redundancy.

[0118] Incremental learning adaptation: For real-time data update scenarios at the edge, a "local tree update mechanism" is designed. When the model is fine-tuned every 100 sets of incremental data, only the decision tree with a prediction error > 5% is replaced, instead of the entire model is reconstructed. This reduces the fine-tuning time to 1 / 5 of that of traditional random forests, adapting to the real-time processing needs of edge nodes.

[0119] The lightweight improved random forest algorithm is based on the existing random forest algorithm, and performs feature pruning, structure optimization and incremental learning adaptation for the edge computing needs of silo monitoring scenarios.

[0120] Specifically, the moisture content distribution is calculated by the inversion model using a trained mapping relationship to determine the moisture content of each region:

[0121]

[0122] For example, if the standard baseline moisture content is 10%, the relative pressure deviation in a certain area is 20%, and the vibration attenuation rate is 15%, then the moisture content of that area = 10% + (0.2 × 0.3 + 0.15 × 0.7) × 10 = 12.25%;

[0123] The packing morphology is calculated by the inversion model using the centroid offset of the pressure distribution to determine the packing skew angle.

[0124]

[0125] Where k is the calibration coefficient, which is obtained by fitting historical data, and its usual value is 6;

[0126] Weld health is quantified by the integrity rate. The base score is based on the crack length under normal conditions. In rainy conditions, if the crack propagation rate is detected by optical fiber at 50% faster than in normal conditions, and the internal defect size is detected by ultrasonic waves.

[0127] If the inversion model uses the normal rainfall deviation value and rainfall influence coefficient relative to the normal environment from the fused dataset, then:

[0128]

[0129] If the "rainfall-temperature difference coupling deviation value" and "rainfall amount + temperature fluctuation joint influence coefficient" are used, the additional influence of temperature fluctuation needs to be added:

[0130]

[0131]

[0132] For example, if the temperature fluctuation value is 12℃, under the rain-temperature difference coupled environment, if the fiber optic monitoring shows that the crack propagation rate is 50% faster than in the normal environment, and the ultrasonic detection shows that the internal defect size increase is 0.2mm, then the additional deviation of temperature fluctuation is calculated as 0.5, and the health score is 46, which can intuitively reflect the additional negative impact of temperature difference on the health score of the weld.

[0133] The rainwater infiltration path is determined based on the fact that welded seams at rainwater infiltration points will undergo localized deformation due to increased humidity, and that temperature sensor data in this area is positively correlated with external humidity. The model locates the infiltration path through clustering of deformation anomalies and temperature-humidity correlation analysis.

[0134]

[0135] Wherein, V0 is the baseline penetration rate, which is obtained by fitting historical data and is usually taken as 0.1 mm / h. The specific index of deformation is obtained based on the strain value change of the weld seam.

[0136] In this embodiment, the deformation is first established using conventional environmental data to determine the reference strain value of the weld seam, specifically:

[0137] In a typical scenario with a fixed temperature environment [(20-25℃) + rainfall] < 10mm, continuous strain data of each monitoring segment of the weld seam were collected; and the average strain value of each monitoring segment under this typical scenario was defined as the reference strain, which was used as a reference for subsequent judgment of whether deformation exists.

[0138] Then, the monitored strain changes are converted into physical deformations, specifically:

[0139] For a specific monitoring segment, the real-time strain under actual conditions is first acquired, and then compared with the reference strain to obtain the strain change. Based on the geometric relationship, the actual deformation of the segment is obtained by using strain = deformation / original length.

[0140] Environmental erosion rate is calculated based on the environmental impact coefficients of the fused dataset combined with the total annual rainfall.

[0141]

[0142] The rainfall impact coefficient is calculated based on the correlation model between surface state parameters and rainfall, and the basic corrosion rate is the corrosion value under normal dry conditions, which is usually taken as 0.3 mm / year.

[0143] Step 104: Determine the internal state risk coefficient of the silo based on the internal state parameters.

[0144] As an optional implementation, step 104, which determines the internal state risk coefficient of the silo based on the internal state parameters, may include:

[0145] The material moisture content, the stacking skew angle, the weld health, the rainwater infiltration rate, and the environmental erosion rate are normalized respectively to obtain the target material moisture content, the target stacking skew angle, the target weld health, the target rainwater infiltration rate, and the target environmental erosion rate.

[0146] The first weight for determining the material moisture content, the second weight for the stacking skew angle, the third weight for the weld health, the fourth weight for the rainwater infiltration rate, and the fifth weight for the environmental erosion rate are determined.

[0147] The internal state risk coefficient of the silo is calculated based on the target material moisture content, the target stacking skew angle, the target weld health, the target rainwater infiltration rate, the target environmental erosion rate, the first weight, the second weight, the third weight, the fourth weight, and the fifth weight.

[0148] This implementation method normalizes various internal state parameters, unifying data dimensions and making different parameters comparable. Determining the weight of each parameter reasonably reflects its impact on silo safety. Calculating the internal state risk coefficient based on the normalized parameters and corresponding weights allows for a comprehensive and accurate assessment of the silo's internal state risk, taking into account multiple factors. This helps in the timely detection of potential safety hazards in silos, enabling proactive and effective preventative measures, thereby ensuring the safe and stable operation of the silos and reducing the likelihood of accidents.

[0149] In this embodiment of the application, the method for normalizing the material moisture content, the stacking skew angle, the weld health, the rainwater infiltration rate, and the environmental erosion rate can be as follows:

[0150] Target material moisture content X1:

[0151]

[0152] Target stacking skew angle X2:

[0153]

[0154] Target weld health level X3:

[0155]

[0156] The target rainwater infiltration rate X4 is normalized:

[0157]

[0158] When the actual infiltration rate of rainwater exceeds the safe infiltration rate threshold, X4 takes the value of 1;

[0159] The target environment erosion rate X5 was also normalized:

[0160]

[0161] Similarly, when the actual environmental erosion rate exceeds the safe threshold for erosion rate, X5 is set to 1.

[0162] In this embodiment, the internal state risk coefficient is calculated using a weighted fusion method, and the formula is as follows:

[0163]

[0164] in, The weights of each internal state parameter are obtained by training a machine learning regression model based on historical risk data and accident samples, and can be dynamically adjusted by the closed-loop optimization module based on the early warning effect. X1 is the target material moisture content, X2 is the target stacking skew angle, X3 is the target weld health, X4 is the target rainwater infiltration rate, and X5 is the target environmental erosion rate.

[0165] Step 105: Based on the internal state risk coefficient and the preset safety threshold, determine the internal state risk level of the silo and the early warning scheme corresponding to the risk level.

[0166] In this embodiment of the application, the preset security threshold includes a first security threshold and a second security threshold, and the first security threshold is less than the second security threshold.

[0167] For example, the first safety threshold can be 30, and the second safety threshold can be 60.

[0168] As an optional implementation, step 105, based on the internal state risk coefficient and a preset safety threshold, determines the internal state risk level of the silo and the corresponding early warning scheme, which may include:

[0169] If the internal state risk coefficient is less than the first safety threshold, the internal state risk level of the silo is determined to be low risk level, and a low risk warning scheme corresponding to the low risk level is obtained.

[0170] If the internal state risk coefficient is greater than or equal to the first safety threshold and less than the second risk level, then the internal state risk level of the silo is determined to be a medium risk level, and a medium risk early warning scheme corresponding to the medium risk level is obtained.

[0171] If the internal state risk coefficient is greater than or equal to the second risk level, then the internal state risk level of the silo is determined to be a high-risk level, and a high-risk early warning scheme corresponding to the high-risk level is obtained.

[0172] This implementation method precisely categorizes risks into low, medium, and high levels by comparing internal risk coefficients with preset safety thresholds. Corresponding early warning plans are then matched to each risk level, making responses more targeted and hierarchical. A mild early warning plan is implemented for low-risk situations to avoid overreaction; for medium and high-risk situations, appropriate early warnings are activated promptly, and effective measures are taken quickly. This tiered early warning mechanism helps silo managers quickly identify the degree of risk, respond in a timely manner, eliminate potential safety hazards at their inception, and ensure the safe and stable operation of the silos.

[0173] For example, for low-risk situations where R < 30, the warning information is only pushed to the mobile terminal operation and maintenance backend and presented in the form of a message reminder, such as "Current structural parameters are normal". The corresponding warning solution is the routine maintenance plan.

[0174] For medium-risk cases (30≤R<60), early warning information is simultaneously pushed to mobile terminals and managers' mobile phones, presented in the form of an orange prompt box + vibration reminder. The content clearly marks the main risk parameters such as "welding degree health score 45". The supporting early warning plan includes a special investigation checklist.

[0175] For high-risk scenarios with an R≥60, an early warning message triggers an audio push notification, simultaneously sent to mobile terminals and managers' phones. This is accompanied by a flashing red light and a voice broadcast, providing a mandatory reminder. The content also includes a real-time risk data dashboard, such as a penetration rate change curve. If the accompanying early warning plan is in place, the emergency response process will be activated.

[0176] It should be further explained that all early warning information and plan execution records are written to the log in real time, including data such as push time, recipient, and operation feedback, providing a traceable basis for subsequent closed-loop optimization.

[0177] Optionally, based on the execution results of the supporting early warning scheme, the deviation rate between the actual and expected effects can be calculated, the weights of the internal state parameters can be dynamically adjusted, and the optimized weights can be fed back into the calculation formula of the internal state risk coefficient through weighted fusion to improve the accuracy of the early warning.

[0178] The formula for calculating the deviation rate between actual and expected results is:

[0179]

[0180] The actual risk index decrease is the risk index R before the implementation of the early warning plan minus the risk index R after implementation. The expected risk index decrease is the risk reduction value that the early warning plan of this level should achieve based on historical data.

[0181] It needs further explanation that the parameter weight adjustment must cover all five-dimensional fusion data dimensions associated with internal state parameters. In particular, for the temperature fluctuation additional deviation specific to the rainfall-temperature difference coupled environment and its association with weld health, and the joint influence coefficient of rainfall-temperature fluctuation and its association with environmental erosion rate, if the deviation rate corresponding to such parameters exceeds 15% for two consecutive times, the weight of the associated internal state parameters, such as the X3 weld health score, should be increased by 5%-8% to ensure that the early warning accuracy of high-risk scenarios is optimized first. The adjusted weights need to be verified, and only after the accuracy is confirmed to be improved can they be fed back into the calculation formula of the internal state risk coefficient through weighted fusion to ensure the reliability of the results.

[0182] Implementing steps 101 to 105 as described above enables silo management to proactively identify potential risks, take timely and effective measures, effectively prevent safety accidents, and improve the safety of silo management. Furthermore, this application makes the data more comprehensive and representative, fully reflecting the actual state of the silos under different environmental combinations, facilitating the early detection of potential risks. In addition, this application can scientifically quantify the impact of different environmental factors on silos, making data fusion and analysis more targeted. Moreover, this application can promptly identify potential safety hazards in silos, allowing for proactive and effective preventative measures, thereby ensuring the safe and stable operation of silos and reducing the likelihood of accidents. Furthermore, this application enables silo managers to quickly identify the degree of risk and respond promptly.

[0183] Based on the same inventive concept, this application also provides a silo internal state inversion early warning device for implementing the aforementioned silo internal state inversion early warning method. The solution provided by this device is similar to the solution described in the above method; therefore, the specific limitations in one or more embodiments of the silo internal state inversion early warning device provided below can be found in the limitations of the silo internal state inversion early warning method described above, and will not be repeated here.

[0184] In one exemplary embodiment, such as Figure 2 As shown, a silo internal state inversion early warning device is provided, comprising:

[0185] The partitioning unit 201 is used to partition the collected data from various sensors into a regular environmental dataset, a regular rainfall environmental dataset, and a rainfall-temperature difference coupled dataset; wherein, the data from various sensors are collected by various types of sensors pre-installed on the silo;

[0186] Processing unit 202 is used to process the conventional environment dataset, the conventional rainfall environment dataset, and the rainfall-temperature difference coupled dataset to obtain a fused dataset;

[0187] Input unit 203 is used to input the fused dataset into a pre-built internal state inversion model to obtain the internal state parameters output by the internal state inversion model;

[0188] The first determining unit 204 is used to determine the internal state risk coefficient of the silo based on the internal state parameters.

[0189] The second determining unit 205 is used to determine the internal state risk level of the silo and the early warning scheme corresponding to the risk level based on the internal state risk coefficient and the preset safety threshold.

[0190] Implementing the above-described methods enables silo management to proactively identify potential risks and take timely and effective measures to prevent safety accidents, thereby improving the safety of silo management. Furthermore, this application makes the data more comprehensive and representative, fully reflecting the actual state of the silos under different environmental combinations, facilitating the early detection of potential risks. In addition, this application can scientifically quantify the impact of different environmental factors on silos, making data fusion and analysis more targeted. Moreover, this application can promptly identify potential safety hazards in silos, allowing for proactive and effective preventative measures, thus ensuring the safe and stable operation of silos and reducing the likelihood of accidents. Furthermore, this application enables silo managers to quickly identify the degree of risk and respond promptly.

[0191] In one exemplary embodiment, a computer device is provided, which may be a server or a terminal, and its internal structure diagram may be as follows. Figure 3 As shown, the computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs in the non-volatile storage media to run. The database stores silo internal state inversion and early warning data. The I / O interfaces are used for information exchange between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a silo internal state inversion and early warning method.

[0192] Those skilled in the art will understand that Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0193] In one exemplary embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0194] In one exemplary embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0195] In one exemplary embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0196] In one exemplary embodiment, a chip is provided, the chip including a processor and a communication interface, the communication interface being coupled to the processor, the processor being used to run programs or instructions to implement the steps in the above method embodiments and achieve the same technical effect, and will not be described again here to avoid repetition.

[0197] It should be understood that the chip mentioned in the embodiments of this application may also be referred to as a system-on-a-chip, system chip, chip system, or system-on-a-chip, etc.

[0198] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0199] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0200] The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0201] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0202] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A silo internal state inversion early warning method, characterized in that, The silo internal state inversion early warning method comprises: dividing the collected multiple sensor data into a normal environment data set, a normal rainfall environment data set, and a rainfall-temperature difference coupling data set, wherein the multiple sensor data is collected by multiple types of sensors pre-set on the silo; processing the normal environment data set, the normal rainfall environment data set, and the rainfall-temperature difference coupling data set to obtain a fusion data set, wherein the rainfall amount corresponding to the core surface parameter in the normal environment data set is less than a preset rainfall amount, the rainfall amount corresponding to the core surface parameter in the normal rainfall environment data set is greater than or equal to the preset rainfall amount, and the temperature fluctuation value corresponding to the core surface parameter in the normal rainfall environment data set is less than a preset fluctuation threshold, the rainfall amount corresponding to the core surface parameter in the rainfall-temperature difference coupling data set is greater than or equal to the preset rainfall amount, and the temperature fluctuation value corresponding to the core surface parameter in the rainfall-temperature difference coupling data set is greater than or equal to the preset fluctuation threshold; inputting the fusion data set into a pre-constructed internal state inversion model to obtain internal state parameters output by the internal state inversion model; determining an internal state risk coefficient of the silo based on the internal state parameters; determining an internal state risk level of the silo and a corresponding early warning scheme based on the internal state risk coefficient and a preset safety threshold; wherein the processing of the normal environment data set, the normal rainfall environment data set, and the rainfall-temperature difference coupling data set to obtain a fusion data set specifically comprises: determining a normal rainfall deviation value between the normal rainfall environment data set and the normal environment data set; determining a rainfall-temperature difference coupling deviation value between the rainfall-temperature difference coupling data set and the normal rainfall environment data set; determining a characteristic difference index based on the normal environment data set, the normal rainfall environment data set, and the rainfall-temperature difference coupling data set, wherein the characteristic difference index comprises a statistical quantity and a degree of freedom; processing the normal rainfall environment data set and the rainfall-temperature difference coupling data set based on the characteristic difference index to obtain a rainfall amount influence coefficient and a rainfall amount-temperature fluctuation joint influence coefficient; fusing the normal environment data set, the normal rainfall deviation value, the rainfall-temperature difference coupling deviation value, the rainfall amount influence coefficient, and the rainfall amount-temperature fluctuation joint influence coefficient to obtain a fusion data set; wherein the calculation formula of the normal rainfall deviation value is: ； ； the calculation formula of the rainfall-temperature difference coupling deviation value is: ； ； wherein W represents the rainfall-temperature difference coupling bias value, represents the rainfall-temperature difference coupling absolute bias value, represents the rainfall-temperature difference coupling dataset mean value; the calculation formula of the statistical quantity t is: ； wherein, is a sample size of the target dataset, is a sample size of the reference dataset, is a sample mean of the target dataset, is a sample mean of the reference dataset, is a sample variance of the target dataset, is a sample variance of the reference dataset; the target dataset comprises a regular environmental dataset, a regular rainfall environmental dataset, and a rainfall-temperature difference coupled dataset; The degrees of freedom The calculation formula is: ； and the processing of the normal rainfall environment data set and the rainfall-temperature difference coupling data set based on the characteristic difference index to obtain a rainfall amount influence coefficient and a rainfall amount-temperature fluctuation joint influence coefficient specifically comprises: performing one-dimensional linear fitting on the normal rainfall environment data set based on the characteristic difference index to obtain a first regression equation; performing binary linear fitting on the rainfall-temperature difference coupling dataset based on the characteristic difference index, to obtain a second regression equation; determining a rainfall influence coefficient from the first regression equation; determining a rainfall-temperature fluctuation joint influence coefficient from the second regression equation; wherein the first regression equation is: ； wherein the independent variable x in the first regression equation represents rainfall, the dependent variable y in the first regression equation represents the uniformity coefficient of the pressure distribution at the bottom of the silo, a1 is the uniformity coefficient when the rainfall is theoretically 0, and b1 is the average change in the uniformity coefficient when the rainfall increases by 1 mm, both a1 and b1 being obtained by statistically fitting historical samples under a conventional rainfall environment; the second regression equation is: ； wherein the independent variable x in the second regression equation represents rainfall, the independent variable z in the second regression equation represents the temperature fluctuation value, the dependent variable y in the second regression equation represents the uniformity coefficient of the pressure distribution at the bottom of the silo, a2 represents the basic deviation value when the rainfall is 0 and the temperature fluctuation value is 0℃, b2 represents the rainfall-only influence coefficient, and c2 represents the temperature fluctuation-only influence coefficient.

2. The silo internal state inversion early warning method according to claim 1, characterized in that, The dividing of the collected multiple sensor data into the conventional environment dataset, the conventional rainfall environment dataset, and the rainfall-temperature difference coupling dataset specifically includes: obtaining core surface parameters and environment parameters corresponding to the core surface parameters from the collected multiple sensor data; wherein the environment parameters include rainfall, temperature, humidity, and temperature fluctuation value; dividing the core surface parameters based on the environment parameters to obtain the conventional environment dataset, the conventional rainfall environment dataset, and the rainfall-temperature difference coupling dataset.

3. The silo internal state inversion early warning method according to claim 1, characterized in that, The internal state parameters include material moisture content, accumulation skew angle, weld joint health degree, rainwater penetration rate, and environmental erosion rate, and the determination of the internal state risk coefficient of the silo based on the internal state parameters specifically includes: normalizing the material moisture content, the accumulation skew angle, the weld joint health degree, the rainwater penetration rate, and the environmental erosion rate respectively to obtain target material moisture content, target accumulation skew angle, target weld joint health degree, target rainwater penetration rate, and target environmental erosion rate; determining a first weight of the material moisture content, a second weight of the accumulation skew angle, a third weight of the weld joint health degree, a fourth weight of the rainwater penetration rate, and a fifth weight of the environmental erosion rate; calculating the internal state risk coefficient of the silo based on the target material moisture content, the target accumulation skew angle, the target weld joint health degree, the target rainwater penetration rate, the target environmental erosion rate, the first weight, the second weight, the third weight, the fourth weight, and the fifth weight.

4. The silo internal state inversion early warning method according to claim 1, characterized in that, The preset safety threshold includes a first safety threshold and a second safety threshold, and the first safety threshold is less than the second safety threshold, and the determination of the internal state risk level of the silo and the early warning scheme corresponding to the risk level based on the internal state risk coefficient and the preset safety threshold specifically includes: If the internal state risk coefficient is less than the first safety threshold, it is determined that the internal state risk level of the silo is a low risk level, and a low risk early warning scheme corresponding to the low risk level is obtained; If the internal state risk coefficient is greater than or equal to the first safety threshold and less than a second safety threshold, it is determined that the internal state risk level of the silo is a medium risk level, and a medium risk early warning scheme corresponding to the medium risk level is obtained; If the internal state risk coefficient is greater than or equal to the second safety threshold, it is determined that the internal state risk level of the silo is a high risk level, and a high risk early warning scheme corresponding to the high risk level is obtained.

5. A silo internal state inversion early warning device, characterized in that, The silo internal state inversion early warning device comprises: A division unit is configured to divide the collected multiple sensor data into a regular environment data set, a regular rainfall environment data set, and a rainfall-temperature difference coupling data set, wherein the multiple sensor data is collected by multiple types of sensors pre-set on the silo; A processing unit is configured to process the regular environment data set, the regular rainfall environment data set, and the rainfall-temperature difference coupling data set to obtain a fusion data set, wherein the rainfall amount corresponding to the core surface parameter included in the regular environment data set is less than a preset rainfall amount, the rainfall amount corresponding to the core surface parameter included in the regular rainfall environment data set is greater than or equal to the preset rainfall amount, and the temperature fluctuation value corresponding to the core surface parameter included in the regular rainfall environment data set is less than a preset fluctuation threshold, the rainfall amount corresponding to the core surface parameter included in the rainfall-temperature difference coupling data set is greater than or equal to the preset rainfall amount, and the temperature fluctuation value corresponding to the core surface parameter included in the rainfall-temperature difference coupling data set is greater than or equal to the preset fluctuation threshold; An input unit is configured to input the fusion data set into a pre-constructed internal state inversion model to obtain internal state parameters output by the internal state inversion model; A first determination unit is configured to determine an internal state risk coefficient of the silo based on the internal state parameters; A second determination unit is configured to determine an internal state risk level of the silo and an early warning scheme corresponding to the risk level based on the internal state risk coefficient and a preset safety threshold; The processing unit processes the regular environment data set, the regular rainfall environment data set, and the rainfall-temperature difference coupling data set to obtain a fusion data set in the following manner: Determine a regular rainfall deviation value between the regular rainfall environment data set and the regular environment data set; Determine a rainfall-temperature difference coupling deviation value between the rainfall-temperature difference coupling data set and the regular rainfall environment data set; Determine a feature difference index based on the regular environment data set, the regular rainfall environment data set, and the rainfall-temperature difference coupling data set, wherein the feature difference index includes a statistical quantity and a degree of freedom; Process the regular rainfall environment data set and the rainfall-temperature difference coupling data set based on the feature difference index to obtain a rainfall amount influence coefficient and a rainfall amount-temperature fluctuation joint influence coefficient; The conventional environment dataset, the conventional rainfall deviation value, the rainfall-temperature difference coupling deviation value, the rainfall influence coefficient and the rainfall-temperature fluctuation combined influence coefficient are fused to obtain a fused dataset; The calculation formula of the conventional rainfall deviation value is: ； ； The calculation formula of the rainfall-temperature difference coupling deviation value is: ； ； wherein W represents the rainfall-temperature difference coupling bias value, represents the rainfall-temperature difference coupling absolute bias value, represents the rainfall-temperature difference coupling dataset mean value; The calculation formula of the statistical quantity t is: ； wherein, is a sample size of the target dataset, is a sample size of the reference dataset, is a sample mean of the target dataset, is a sample mean of the reference dataset, is a sample variance of the target dataset, is a sample variance of the reference dataset; the target dataset comprises a regular environmental dataset, a regular rainfall environmental dataset, and a rainfall-temperature difference coupled dataset; The degrees of freedom The calculation formula is: ； The processing unit processes the conventional rainfall environment dataset and the rainfall-temperature difference coupling dataset based on the feature difference index to obtain a rainfall influence coefficient and a rainfall-temperature fluctuation combined influence coefficient in the following manner: The conventional rainfall environment dataset is linearly fitted based on the feature difference index to obtain a first regression equation; The rainfall-temperature difference coupling dataset is linearly fitted based on the feature difference index to obtain a second regression equation; A rainfall influence coefficient is determined from the first regression equation; A rainfall-temperature fluctuation combined influence coefficient is determined from the second regression equation; The first regression equation is: ； In the first regression equation, the independent variable x represents the rainfall, the dependent variable y represents the uniformity coefficient of the silo bottom pressure distribution, a1 is the uniformity coefficient when the rainfall is 0 in theory, and b1 is the average change amount of the uniformity coefficient when the rainfall increases by 1 mm, both a1 and b1 being obtained by statistically fitting historical samples under a conventional rainfall environment; The second regression equation is: ； In the second regression equation, the independent variable x represents the rainfall, the independent variable z represents the temperature fluctuation value, and the dependent variable y represents the uniformity coefficient of the silo bottom pressure distribution, a2 represents the basic deviation value when the rainfall is 0 and the temperature fluctuation value is 0℃, b2 represents the rainfall independent influence coefficient, and c2 represents the temperature fluctuation independent influence coefficient.

6. A computer device comprising: A memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the computer program to implement the steps of the silo internal state inversion early warning method of any one of claims 1-4.

7. A computer readable storage medium having stored thereon a computer program, characterized in that The computer program is executed by the processor to implement the steps of the silo internal state inversion early warning method of any one of claims 1-4.

8. A computer program product comprising a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the silo internal state inversion early warning method of any one of claims 1-4.

Images