A photovoltaic slope safety evaluation method considering spatio-temporal variability of multi-source parameters

Abstract

The application provides a photovoltaic slope safety evaluation method considering spatiotemporal variability of multi-source parameters, comprising: collecting geological exploration data and multi-source monitoring data of a photovoltaic slope, determining a random distribution type of initial multi-source parameters, and selecting key monitoring data; performing data matching of multi-source parameters and monitoring data, and screening multi-source parameters and monitoring data by using a MIC algorithm; based on field geological exploration data, a three-dimensional simulation model considering spatiotemporal variability of multi-source parameters of the photovoltaic slope is constructed, and a data set is obtained; based on a 1D-CNN-MCMC deep learning model framework, the data set is trained, and a posterior probability distribution of a safety factor FS of the photovoltaic slope is dynamically updated in combination with field monitoring data; a failure probability P of the photovoltaic slope is calculated f , and safety evaluation and early warning are performed. The application comprehensively considers uncertain factors such as slope geological parameter variability, environment, and photovoltaic structure, uses field multi-source monitoring data, obtains a photovoltaic slope failure probability more in line with actual conditions, and performs safety evaluation and early warning.

Description

Technical Field

[0001] This invention relates to the field of geotechnical engineering safety monitoring technology, specifically to a method for evaluating the safety of photovoltaic slopes that integrates the spatiotemporal variability of multi-source monitoring data. Background Technology

[0002] Photovoltaics, as a clean and renewable energy source, requires a significant amount of land for power generation. However, a considerable amount of vacant land and space resources exist along highway slopes. Therefore, highway slopes can be utilized as a new application scenario for photovoltaic power generation, reducing carbon emissions in the transportation sector. Currently, the photovoltaic slope industry is in its nascent stage, and existing slope stability assessment methods do not incorporate the load effects of photovoltaic structures. Directly using traditional slope safety methods will result in inaccurate and unreliable evaluation results, posing significant safety risks.

[0003] Numerous parameters influence the safety of photovoltaic (PV) slopes, including soil and rock parameters, environmental factors, and PV structure parameters. Effectively selecting these parameters is crucial for PV slope safety assessment. Furthermore, previous slope safety assessment methods typically only considered the spatial variability of soil and rock parameters, rarely addressing the impact of time effects. Moreover, these methods primarily relied on theoretical equations or numerical simulations to account for time effects, lacking practical guidance and making it difficult to guarantee long-term prediction accuracy. Summary of the Invention

[0004] This invention aims to solve the problem of selecting multi-source parameters and monitoring data for photovoltaic slopes, and to consider the spatiotemporal variability of multi-source parameters for safety assessment and early warning of photovoltaic slopes, so as to ensure the stable operation of photovoltaic slopes.

[0005] To achieve the above objectives, this invention proposes a photovoltaic slope safety evaluation method considering the spatiotemporal variability of multi-source parameters, characterized by the following steps:

[0006] Step 1: Collect geological survey data and multi-source monitoring data of photovoltaic slopes, determine the random distribution type of initial multi-source parameters, and select key monitoring data;

[0007] Step 2: Perform data matching between multi-source parameters and monitoring data, and use the MIC algorithm to filter multi-source parameters and monitoring data;

[0008] Step 3: Based on the field geological survey data, construct a three-dimensional simulation model that considers the spatiotemporal variability of multi-source parameters of the photovoltaic slope, and obtain the dataset;

[0009] Step 4: Based on the 1D-CNN-MCMC deep learning model framework, train the dataset and dynamically update the posterior probability distribution of the photovoltaic slope safety factor FS by combining it with on-site monitoring data;

[0010] Step 5: Calculate the failure probability P of the photovoltaic slope. f Safety assessments and early warnings are conducted.

[0011] Preferably, step 1 further includes the following steps:

[0012] Step 1.1: The collected photovoltaic slope geological survey data includes, but is not limited to, slope topography, slope geology, drilling results, and photovoltaic structure design drawings. The initial multi-source parameter information that can be collected includes: the mean and standard values ​​of the physical and mechanical parameters of the soil and rock mass (cohesion C, internal friction angle α, elastic modulus E, Poisson's ratio v); the mean and standard deviation of rainfall RF; the mean and standard deviation of wind speed AV; and the mean and standard deviation of the location P of the photovoltaic structure on the slope. The random distribution type of the initial multi-source parameters, i.e., the prior distribution of the multi-source parameters, can be determined using the following formula.

[0013] θ prio =C,α,E,v,…,RF,AV,P

[0014] Where C is cohesion, α is the internal friction angle, E is the elastic modulus, v is Poisson's ratio, RF is rainfall, AV is wind speed, and P is the photovoltaic structure location parameter.

[0015] Step 1.2: The collected multi-source monitoring data for the photovoltaic slope includes, but is not limited to, slope surface displacement (SD), deep displacement (DD), soil moisture content (SM), relative displacement (PD) of the photovoltaic structure columns, tilt of the photovoltaic structure columns (PA), and strain of the photovoltaic structure braces and columns (PS). The multi-source monitoring data is divided into key monitoring data and other monitoring data. Based on the slope failure characteristics, key monitoring data affecting the safety of the photovoltaic slope are selected for use in the subsequent MIC calculation in Step 2.2.

[0016] Preferably, step 2 further includes the following steps:

[0017] Step 2.1: The multi-source parameters are spatial variables G = {g1, g2, ..., g...} n (A random distribution determined by the mean and standard deviation), while multi-source monitoring data is a time series D. t ={d1,d2,…,d n (Varies with time t). A matching strategy is used to calculate the correlation between the two, that is, at each time point t, a multi-source parameter value G is associated. t Multi-source parameter variables were matched with key monitoring data. To avoid randomness, the Monte Carlo method was used to randomly sample the multi-source parameters N times.

[0018] Step 2.2: Perform MIC calculations on the multi-source parameters sampled N times and the key monitoring data respectively. Take the average of the calculated MICs; a higher average MIC indicates a stronger correlation. Set a threshold to filter out multi-source parameters with high correlation to the key monitoring data. The MIC calculation can be expressed by the following formula:

[0019]

[0020] In the formula, integers a and b represent the number of segments in the x and y directions, I(D|G) represents the mutual information (MI) based on grids G and D, Ω(a,b) represents the set of two-dimensional grids of size a×b, L(n) is used to limit the size of the grid division region, and n is the length of the input data sequence. In this invention, L(n) is taken as n. 0.6 .

[0021] Step 2.3: The multi-source monitoring data includes key monitoring data and other monitoring data. Based on the multi-source monitoring data of the photovoltaic slope collected in Step 1.2, calculate the MIC values ​​of the other monitoring data and the key monitoring data, and set a threshold to filter out the remaining key monitoring data to be used in the subsequent Step 4.2.

[0022] Preferably, step 3 further includes the following steps:

[0023] Step 3.1: Based on the on-site geological survey data of the photovoltaic slope, establish a three-dimensional photovoltaic slope simulation model. Optimal Latin hypercube sampling is used to uniformly extract M random sample combinations. Based on random field theory, the spatiotemporal variability of multi-source parameters is set. These multi-source parameters are used as the model input for the three-dimensional simulation model to perform simulation calculations and obtain the corresponding associated response value g(θ). This can be expressed by the following formula;

[0024] g(θ)=g(C,α,E,v,…,RF,AV,P)

[0025] Step 3.2: Obtain the required correlation response values. Except for the safety factor FS, the obtained correlation response values ​​must correspond to existing monitoring data. Construct a dataset from the multi-source parameter inputs and correlation response value outputs.

[0026] Preferably, step 4 further includes the following steps:

[0027] Step 4.1: Build a multi-input multi-output 1D-CNN deep learning framework model to predict associated response values. Divide the dataset obtained in Step 3 into training set, validation set and test set, adjust the hyperparameters to obtain a surrogate model with high prediction accuracy and strong generalization ability, which can be expressed by the following formula;

[0028] 1D-CNN = {θ, g(θ)}

[0029] Step 4.2: Combine the remaining key monitoring data D selected by MIC in Step 2.3 Mon Based on Bayes' theorem and Markov chain sampling algorithm, the posterior distribution of multi-source parameters is obtained using the Markov chain Monte Carlo (MCMC) method. This process dynamically updates the posterior distribution with real-time input of monitoring data, thereby obtaining the posterior distribution θ of the multi-source parameters considering the time effect. post As shown in the following formula:

[0030]

[0031] In the formula, l is the standardization factor, which ensures the consistency of the cumulative probability throughout the entire θ range.

[0032] Step 4.3: The posterior distribution of the multi-source parameters obtained through dynamic updating in Step 4.2 is predicted using a 1D-CNN surrogate model. This allows for real-time and rapid acquisition of the posterior probability distribution of the associated response value (safety factor FS), as shown in the following formula:

[0033] 1D-CNN = {θ post ,g(θ post )}

[0034] Preferably, step 5 further includes the following steps:

[0035] Step 5.1: Based on reliability theory, establish the function Z of the photovoltaic slope, as shown in the following equation:

[0036] Z = FS-1

[0037] Step 5.2: Based on the posterior distribution of the safety factor FS obtained in Step 4.3 and the function function Z of the photovoltaic slope established in Step 5.1, the real-time failure probability P of the slope is calculated using the Monte Carlo method. f As shown in the following formula:

[0038]

[0039] In the formula, P f Let P(FS<1) be the probability of instability of the photovoltaic-slope system, and P(FS<1) be the instability probability determined by the safety factor; N (k) For the k-th sample generated by sampling, I{·} is an indicator function; if FS < 1, I{·} = 1, otherwise I{·} = 0; n FS This represents the total number of FS<1.

[0040] The real-time failure probability P of the slope is calculated using the safety factor FS. f Safety assessments and early warning classifications are conducted, as shown in Table 1:

[0041]

[0042] Table 1 shows that the safety levels are divided into five levels: Level I, Level II, Level III, Level IV, and Level V. The warning levels are also divided into five levels: green, blue, yellow, orange, and red. Green represents safety, blue represents entering the warning state, and yellow, orange, and red represent the alarm state.

[0043] Compared with the prior art, the present invention has the following beneficial effects:

[0044] 1. Comprehensive consideration of multi-source parameter selection and photovoltaic structure

[0045] Breaking through the limitations of traditional slope evaluation that ignores the load effect of photovoltaic structures, this paper proposes an evaluation method that integrates soil and rock parameters, environmental factors, and photovoltaic structure parameters.

[0046] A matching strategy for multi-source parameters and key monitoring data based on the MIC algorithm is proposed. The Monte Carlo method is used to eliminate the influence of chance and screen multi-source parameters and monitoring data.

[0047] 2. Innovative Application of Spatiotemporal Variability Simulation Models

[0048] Based on random field theory, the spatiotemporal variability of multi-source parameters is set, and the multi-source parameters are used as model inputs of a three-dimensional simulation model to obtain the corresponding associated response values.

[0049] We construct a multi-input multi-output 1D-CNN-MCMC deep learning framework to predict associated response values ​​and train datasets. This breaks through the limitations of static analysis in traditional numerical simulations and enables real-time updating of the posterior probability distribution of the safety factor FS.

[0050] 3. Data-driven dynamic assessment of time effects

[0051] Based on Bayes' theorem and Markov chain Monte Carlo (MCMC) method for updating the posterior distribution of multi-source parameters, the posterior probability distribution of the associated response value (safety factor FS) can be obtained in real time and rapidly through surrogate model prediction. A dynamic update mechanism for the posterior distribution of the safety factor FS is proposed, enabling the evaluation results to evolve in real time with monitoring data.

[0052] The real-time failure probability P of the slope is calculated using the safety factor FS. f Conduct safety assessments and early warning classifications, establish a five-level early warning system, and improve the engineering practicality of long-term forecasts. Attached Figure Description

[0053] Figure 1 This is a flowchart of a photovoltaic slope safety evaluation method that considers the spatiotemporal variability of multi-source parameters according to the present invention;

[0054] Figure 2A deep learning model framework for dynamically updating the posterior distribution of the safety factor FS of photovoltaic slopes is provided as an embodiment of the present invention. Detailed Implementation

[0055] To better understand the above-mentioned objectives, features and advantages of the present invention, the following description, in conjunction with the accompanying drawings and embodiments, further illustrates a photovoltaic slope safety evaluation method considering the spatiotemporal variability of multi-source parameters.

[0056] The aforementioned method for safety evaluation of photovoltaic slopes considering the spatiotemporal variability of multi-source parameters includes the following steps:

[0057] Step 1: Collect geological survey data and multi-source monitoring data of photovoltaic slopes, determine the random distribution type of initial multi-source parameters, and select key monitoring data;

[0058] Specifically, step 1 includes the following steps:

[0059] Step 1.1 specifically involves collecting on-site geological data of the photovoltaic slope, including but not limited to slope topographic maps, slope geological conditions, and photovoltaic structure design drawings. The collected multi-source parameter information includes: the mean and standard values ​​of the physical and mechanical parameters of the soil and rock mass (cohesion C, internal friction angle α, elastic modulus E, Poisson's ratio v); the mean and standard deviation of rainfall RF; the mean and standard deviation of wind speed AV; and the mean and standard deviation of the location P of the photovoltaic structure on the slope. After determining the initial multi-source parameters, the random distribution type of these corresponding initial multi-source parameters is found according to the relevant specifications or references. Cohesion C and internal friction angle α follow a log-normal distribution; elastic modulus E, Poisson's ratio v, the location P of the photovoltaic structure on the slope, and rainfall RF follow a normal distribution; and wind speed AV follows an extreme value type I distribution.

[0060] The specific implementation described in 1.2 involves collecting multi-source monitoring data for the photovoltaic slope, including but not limited to: slope surface displacement (SD), deep displacement (DD), soil moisture content (SM), relative displacement (PD) of the photovoltaic structure columns, tilt of the photovoltaic structure columns (PA), and strain of the photovoltaic structure braces and columns (PS). Considering that this embodiment is a highway photovoltaic slope, the project site experiences low annual rainfall and has few steep slopes along the route. Therefore, the landslide is typically characterized by deep sliding. Based on this, deep displacement (DD) is selected as the key monitoring data for the MIC calculation in subsequent step 2.2.

[0061] Step 2: Perform data matching between multi-source parameters and monitoring data, and use the MIC algorithm to filter multi-source parameters and monitoring data;

[0062] The correlation is calculated using the MIC algorithm, and the MIC calculation is shown in the following formula:

[0063]

[0064] In the formula, integers a and b represent the number of segments in the x and y directions, I(D|G) represents the mutual information (MI) based on grids G and D, Ω(a,b) represents the set of two-dimensional grids of size a×b, L(n) is used to limit the size of the grid division region, and n is the length of the input data sequence. Usually, L(n) = n 0.6 In this embodiment, L(n) takes the value n. 0.6 .

[0065] The specific implementation of step 2.1: The multi-source parameters are spatial variables G = {g1, g2, ..., g...} n (A random distribution determined by the mean and standard deviation), while the key monitoring data, deep displacement DD, is a time series D. t ={d1,d2,…,d n (Varies with time t). A matching strategy is used to calculate the correlation between the two, that is, at each time point t, a multi-source parameter value G is associated. t Multi-source parameter variables were matched with key monitoring data on deep displacement (DD). To avoid randomness, the Monte Carlo method was used to randomly sample the multi-source parameters N times.

[0066] Step 2.2 is specifically implemented by performing MIC calculations on the multi-source parameters sampled N times and the deep displacement DD of the key monitoring data, and taking the average value of the MIC as shown in the following formula:

[0067]

[0068] In the formula, MIC j,i It is the MIC value of the j-th multi-source parameter at the i-th sampling point and the key data, and N is the number of samplings.

[0069] A higher MIC value indicates a stronger correlation. A threshold is set to filter out multi-source parameters that are highly correlated with the key monitoring data of deep displacement (DD). In this embodiment, the threshold is set to 0.3, and multi-source parameters below 0.3 will be discarded.

[0070] Step 2.3 is implemented in detail, and the multi-source monitoring data includes key monitoring data and other monitoring data. Based on the multi-source monitoring data of the photovoltaic slope collected in step 1.2, the MIC value of the other monitoring data and the deep displacement DD of the key detection data is calculated. A threshold is set to filter out the other key monitoring data that are highly correlated with the deep displacement DD of the key detection data, which will be used for updating the Markov chain Monte Carlo method MCMC in the subsequent step 4.2. In this embodiment, the MIC threshold is set to 0.3, and other monitoring data below 0.3 will be discarded.

[0071] Step 3: Based on the field geological survey data, construct a three-dimensional simulation model that considers the spatiotemporal variability of multi-source parameters of the photovoltaic slope, and obtain the dataset;

[0072] Specifically, step 3 includes the following steps:

[0073] Step 3.1 is implemented by establishing a three-dimensional photovoltaic slope simulation model based on on-site geological survey data. Based on random field theory, multi-source parameters are linked to space through correlation distance. A simple spatial random field distribution is adopted, and the spatiotemporal variability of the multi-source parameters is set. Optimal Latin hypercube sampling is used to uniformly extract a random sample combination of M multi-source parameters. This random sample combination of M multi-source parameters is used as the model input for simulation calculation to obtain the corresponding implicit correlation response value g(θ), which can be expressed by the following formula:

[0074] g(θ)=g(C,α,E,v,…,RF,AV,P)

[0075] Step 3.2 involves numerical simulation to obtain the required associated response values, including slope surface displacement SD, deep displacement DD, soil moisture content SM, relative displacement PD of the photovoltaic structure column, and tilt of the photovoltaic structure column PA. Except for the safety factor FS, the obtained associated response values ​​must correspond to the remaining monitoring data selected in step 2.3. The multi-source parameter inputs and associated response value outputs are then used to construct a dataset.

[0076] Step 4: Based on the 1D-CNN-MCMC deep learning model framework, train the dataset and dynamically update the posterior probability distribution of the photovoltaic slope safety factor FS by combining it with on-site monitoring data;

[0077] Specifically, step 4 includes the following steps:

[0078] Step 4.1 is specifically implemented by building a multi-input multi-output 1D-CNN deep learning framework model to predict associated response values, such as... Figure 2 The constructed 1D-CNN consists of 5 convolutional layers, 1 residual layer, an activation function layer, 1 unfolded layer, 1 dropout layer, and 2 fully connected layers. The loss function is the mean squared error (MSE), and the evaluation metric is the coefficient of determination (R-squared, R0). 2 A 300-round early stopping strategy is implemented. The dataset obtained in step 3 is divided into a training set, a validation set, and a test set. The training set is used to train the model, the validation set is used to adjust the model's hyperparameters, and the test set is used to test the model's accuracy. The resulting surrogate model, with high prediction accuracy and strong generalization ability, can be expressed by the following formula:

[0079] 1D-CNN = {θ, g(θ)}

[0080] Step 4.2 is implemented in conjunction with the remaining key monitoring data D selected by MIC in step 2.3. Mon Based on Bayes' theorem, a Markov algorithm is used to extract Markov chains, and the posterior distribution θ of the multi-source parameters is obtained by updating using the Markov chain Monte Carlo method (MCMC). post This process dynamically updates the posterior distribution as other key monitoring data are input in real time, thereby obtaining the parametric posterior distribution that takes into account the time effect, as shown in the following formula:

[0081]

[0082] In the formula, l is the standardization factor, which ensures the consistency of the cumulative probability throughout the entire θ range.

[0083] In step 4.3, the posterior distribution of the multi-source parameters obtained through the dynamic update in step 4.2 is predicted by the 1D-CNN surrogate model, as shown in the following formula, which can obtain the posterior distribution of the associated response value (safety factor FS) in real time and quickly.

[0084] 1D-CNN = {θ post ,g(θ post )}

[0085] Step 5: Calculate the failure probability P of the photovoltaic slope. f Safety assessments and early warnings are conducted.

[0086] Specifically, step 5 includes the following steps:

[0087] Step 5.1 is implemented as follows: Based on reliability theory, the function Z of the photovoltaic slope is established, as shown in the following formula:

[0088] Z = FS-1

[0089] Step 5.2 is specifically implemented as follows: Based on the posterior distribution of the safety factor FS obtained in step 4.3, and the function Z of the photovoltaic slope established in step 5.1, the real-time failure probability P of the slope is calculated using the Monte Carlo method. f As shown in the following formula:

[0090]

[0091] In the formula, P f Let P(FS<1) be the probability of instability of the photovoltaic slope, and N be the probability of instability determined by the safety factor. (k) Let I{·} be the k-th sample generated by sampling, and let I{·} be an indicator function. If FS < 1, I{·} = 1; otherwise, I{·} = 0. FS This represents the total number of FS<1.

[0092] The real-time failure probability P of the slope is calculated using the safety factor FS. f Safety assessments and early warning classifications are conducted, as shown in Table 1:

[0093]

[0094] Table 1 shows that the safety levels are divided into five levels: Level I, Level II, Level III, Level IV, and Level V. The warning levels are also divided into five levels: green, blue, yellow, orange, and red. Green represents safety, blue represents entering the warning state, and yellow, orange, and red represent the alarm state.

[0095] Safety assessment level I indicates the highest safety, with the photovoltaic slope in a stable state and a failure probability between 0% and 5%, belonging to a low-probability event. The warning level is green, and no alarm is triggered. Similarly, safety assessment level V indicates the lowest safety, with the photovoltaic slope in an unstable state and a failure probability between 90% and 100%, belonging to a high-probability event. The warning level is red, and an alarm is required.

[0096] This invention solves the problem of early warning lag caused by static assumptions of geological parameters in traditional methods by integrating multi-source monitoring data with deep learning models, and realizes dynamic probabilistic assessment of slope safety status.

[0097] Although the present invention has been described in detail in this specification with general description and specific embodiments, some modifications or improvements can be made to it, which will be obvious to those skilled in the art. Therefore, all such modifications or improvements made without departing from the spirit of the present invention are within the scope of protection claimed by the present invention.

Claims

1. A method for safety evaluation of photovoltaic slopes considering the spatiotemporal variability of multi-source parameters, characterized in that, Includes the following steps: Step 1: Collect geological survey data and multi-source monitoring data of photovoltaic slopes, determine the random distribution type of initial multi-source parameters, and select key monitoring data; Step 2: Perform data matching between multi-source parameters and monitoring data, and use the MIC algorithm to filter multi-source parameters and monitoring data; Step 2.1: Multi-source parameters are spatial variables (A random distribution determined by the mean and standard deviation), while multi-source monitoring data is a time series. (Changes with time t); A matching strategy is used to calculate the correlation between the two, that is, a multi-source parameter value is associated at each time point t. To match multi-source parameter variables with key monitoring data, the Monte Carlo method was used to randomly sample multi-source parameters in order to avoid randomness. Second-rate; Step 2.2: Sample The multi-source parameters were used to calculate the MIC (Minimum Indicator) with the key monitoring data. The average value of the calculated MIC was taken. A threshold was set to filter out the multi-source parameters that were highly correlated with the key monitoring data. The MIC calculation can be expressed by the following formula: Integer Indicates in Number of segments in the direction, This represents the mutual information (MI) based on grids G and D. Indicates a value of A collection of two-dimensional grids Used to limit the size of the grid division region. The length of the input data sequence. ; Step 2.3: The multi-source monitoring data includes: key monitoring data and other monitoring data; Calculate the MIC values ​​of the remaining monitoring data and the key monitoring data, and set a threshold to filter out the remaining key monitoring data that need to be used in the subsequent step 4. Step 3: Based on the field geological survey data, construct a three-dimensional simulation model that considers the spatiotemporal variability of multi-source parameters of the photovoltaic slope, and obtain the dataset; Step 4: Based on the 1D-CNN-MCMC deep learning model framework, train the dataset and dynamically update the safety factor of the photovoltaic slope in combination with on-site monitoring data. FS The posterior probability distribution; Step 4.1: Build a multi-input multi-output 1D-CNN deep learning framework model to predict associated response values. Divide the dataset obtained in Step 3 into training set, validation set and test set, adjust the hyperparameters to obtain a surrogate model with high prediction accuracy and strong generalization ability, which can be expressed by the following formula; Step 4.2: Combine the remaining key monitoring data filtered out by MIC in Step 2.

3. Based on Bayes' theorem and Markov chain sampling algorithm, the posterior distribution of multi-source parameters is obtained using the Markov chain Monte Carlo (MCMC) method. This process dynamically updates the posterior distribution with real-time input of monitoring data, thereby obtaining the posterior distribution of multi-source parameters considering time effects. As shown in the following formula: In the formula, As a standardization factor, it guarantees the entire Consistency of cumulative probability within the range; Step 4.3: The posterior distribution of the multi-source parameters obtained through dynamic updating in Step 4.2 is predicted using a 1D-CNN surrogate model to obtain the associated response value (safety factor) in real time and quickly. The posterior probability distribution of is shown in the following equation: ; Step 5: Calculate the failure probability of the photovoltaic slope Safety assessments and early warnings are conducted.

2. The photovoltaic slope safety evaluation method considering the spatiotemporal variability of multi-source parameters according to claim 1, characterized in that, Step 1 also includes the following steps: Step 1.1: Collect on-site geological survey data for the photovoltaic slope, including but not limited to slope topographic maps, slope geological conditions, and photovoltaic structure design drawings; the collected initial multi-source parameters include the physical and mechanical parameters of the soil and rock mass (cohesion). internal friction angle Elastic modulus Poisson's ratio The mean and standard values ​​of rainfall; Mean and standard deviation; wind speed Mean and standard deviation; location of the photovoltaic structure on the slope. The mean and standard deviation of the multi-source parameters; once the initial multi-source parameters are determined, the prior distribution of the multi-source parameters is obtained, which can be expressed by the following formula; in, For cohesion, It is the internal friction angle. For elastic modulus, Poisson's ratio, For rainfall, For wind speed, These are the location parameters of the photovoltaic structure; The random distribution types of these parameters can be found by referring to the relevant standards or references; Step 1.2: Collect multi-source monitoring data of the photovoltaic slope, including but not limited to slope surface displacement. Deep displacement Soil moisture content Relative displacement of photovoltaic structure columns and the tilt of the photovoltaic structure column Photovoltaic structure diagonal bracing and column strain The multi-source monitoring data is divided into key monitoring data and other monitoring data. Based on the slope failure characteristics, key monitoring data that affect the safety of photovoltaic slopes are selected for MIC calculation in subsequent step 2.

3. The photovoltaic slope safety evaluation method considering the spatiotemporal variability of multi-source parameters according to claim 1, characterized in that, Step 3 also includes the following steps: Step 3.1: Based on the on-site geological survey data of the photovoltaic slope, establish a three-dimensional photovoltaic slope simulation model; use optimal Latin hypercube sampling for uniform extraction. Using a combination of random samples, and based on random field theory, the spatiotemporal variability of multi-source parameters is set. These multi-source parameters are then used as inputs to a three-dimensional simulation model to perform simulation calculations and obtain the corresponding correlation response values. It can be expressed by the following formula; Step 3.2: Obtain the required associated response values; excluding the safety factor. In addition, the obtained correlation response values ​​must correspond to existing monitoring data; the multi-source parameter inputs and correlation response value outputs are used to construct a dataset.

4. The photovoltaic slope safety evaluation method considering the spatiotemporal variability of multi-source parameters according to claim 1, characterized in that, Step 5 also includes the following steps: Step 5.1: Based on reliability theory, establish the function of the photovoltaic slope. As shown in the following formula: Step 5.2: Based on the safety factor obtained in Step 4.3 The posterior distribution, and the function of the photovoltaic slope established in step 5.

1. The real-time failure probability of the slope was calculated using the Monte Carlo method. As shown in the following formula: In the formula, The probability of photovoltaic-slope instability. The probability of instability is determined by the safety factor; For the k-th sample generated by sampling, For an indicator function, if , ,otherwise ; for The total number; Through safety factor Calculated real-time failure probability of the slope Safety assessments and early warning classifications are conducted. Safety assessment levels are divided into five levels: Level I, Level II, Level III, Level IV, and Level V. Early warning levels are divided into five levels: green, blue, yellow, orange, and red, corresponding to different failure probability ranges.

5. The method according to claim 1, characterized in that, The threshold of the MIC algorithm is set to 0.3 to filter multi-source parameters and monitoring data with a correlation higher than 0.

3.

6. The method according to claim 2, characterized in that, The photovoltaic slope along the highway experiences low annual rainfall and has few steep slopes along the route, allowing for the selection of deep displacement. This is key monitoring data.

7. The method according to claim 3, characterized in that, In step 3.2, the required associated response values, including the slope surface displacement, are obtained through numerical simulation calculation. Deep displacement Soil moisture content Relative displacement of photovoltaic structure columns and the tilt of the photovoltaic structure column .

8. The method according to claim 1, characterized in that, The constructed 1D-CNN consists of 5 convolutional layers, 1 residual layer, 1 activation function layer, 1 unfolded layer, 1 dropout layer, and 2 fully connected layers. The loss function is the mean squared error (MSE), and the evaluation metric is the coefficient of determination (R-squared). Set up an early stop strategy for 300 rounds.

Images