A wind-resistant early warning method and system for a flexible photovoltaic support

Abstract

The application relates to the technical field of photovoltaic power generation, in particular to a wind resistance early warning method and system for a flexible photovoltaic support. The method comprises the following steps: determining the stability of each connecting part of the photovoltaic support according to vibration data of the connecting parts; obtaining a predicted wind speed and a predicted wind direction at the next moment; taking the orientation of the photovoltaic support and the inclination angle of the cell panel at the current moment as attitude information, calculating the attention degree of each connecting part according to the attitude information and the predicted wind direction; calculating the predicted wind load at the next moment according to the predicted wind speed, the predicted wind direction and the attitude information, and calculating the local risk value of each connecting part; weighting and summing the risk values of the connecting parts by using the normalized attention degrees to obtain the overall risk value; and realizing wind resistance early warning according to the overall risk value. Through the technical scheme, the accuracy of wind resistance early warning of the flexible photovoltaic support can be improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of photovoltaic power generation, in particular to a wind resistance early warning method and system for a flexible photovoltaic support. BACKGROUND

[0002] As an important component of solar photovoltaic systems, flexible photovoltaic supports gradually occupy an important position in the photovoltaic field because they can adapt to different environmental conditions and installation requirements. In order to maintain maximum power generation, flexible photovoltaic supports often adjust their angles according to the position of the sun to maximize the collection efficiency of solar energy during operation. However, when facing adverse weather conditions such as strong winds, how to ensure the safety and stability of flexible photovoltaic supports is a problem that needs to be solved urgently.

[0003] At present, the patent application file with the publication number CN118609325A discloses a photovoltaic tracking support wind resistance early warning protection method and system, wherein the method includes: obtaining historical wind speed data with a length of an integer multiple of the prediction period according to the configured prediction period; dividing the obtained wind speed data according to the prediction period to form multiple data segments; then calculating the mean data corresponding to each data segment to obtain the mean data set M; taking the data in the mean data set M as the input of the Kalman filter model, calculating the wind speed mean M2 of the prediction period, based on the obtained wind speed mean M2 and the power plant turbulence level parameter Ir and the turbulence standard deviation std, calculating the prediction mean M3 through the calculation formula M3=M2+Ir×(0.75×M2+5.6)×2, and calculating the prediction limit wind speed MS according to the calculation formula MS=M2+3×M3, if the prediction limit wind speed MS is greater than the wind speed warning threshold, then the strong wind early warning process is executed.

[0004] The above method uses the Kalman filter model to obtain the wind speed mean of the prediction period, and then determines the prediction limit wind speed, and realizes the strong wind early warning by comparing the prediction limit wind speed with the wind speed warning threshold. However, the above method ignores the influence of the stability of the photovoltaic support structure on the wind resistance, which may cause false alarm and result in inaccurate wind resistance early warning. SUMMARY

[0005] In order to solve the technical problem of inaccurate wind resistance early warning, the present application provides a wind resistance early warning method and system for a flexible photovoltaic support, which can improve the accuracy of wind resistance early warning for a flexible photovoltaic support.

[0006] In a first aspect, the application provides a wind resistance warning method for a flexible photovoltaic support, the warning method comprising: determining the stability of each connection part of the photovoltaic support according to vibration data of vibration sensors at each connection part; obtaining a predicted wind speed and a predicted wind direction at the next moment using a time series prediction model; taking the orientation of the photovoltaic support and the inclination angle of the cell panel at the current moment as attitude information, and calculating the attention degree of each connection part according to the attitude information and the predicted wind direction, comprising: inputting the attitude information and the predicted wind direction into a ranking model to output a ranking result of the force size of each connection part, and the attention degree of each connection part is negatively correlated with the ranking in the ranking result; calculating the predicted wind load at the next moment according to the predicted wind speed, the predicted wind direction and the attitude information, and calculating the local risk value of each connection part, the local risk value is positively correlated with the predicted wind load and negatively correlated with the stability; weighting and summing the risk values of each connection part using the normalized attention degree to obtain the overall risk value; and issuing a warning in response to the overall risk value being greater than a risk threshold.

[0007] Preferably, determining the stability of each connection part of the photovoltaic support according to the vibration data of the vibration sensors at each connection part comprises: performing Fourier transform on the vibration data to obtain energy values of each connection part at different vibration frequencies; and taking the normalized energy values as weights to calculate the central vibration frequency of each connection part, and the stability of each connection part is negatively correlated with the central vibration frequency.

[0008] When the stability of the connection part is large, the energy values are concentrated in the low frequency region, and the central vibration frequency is a small value. When the connection part is loose, cracked or has a change in stiffness, the energy values spread to the medium and high frequency region, and the central vibration frequency increases. According to the central vibration frequency of each connection part, the stability of each connection part can be accurately measured, providing a data basis for subsequent wind resistance warning.

[0009] Preferably, the stability of the connection part is . ; The central vibration frequency of the connection part , is the maximum vibration frequency.

[0010] Preferably, the training method of the time series prediction model comprises: collecting wind speed sequences and wind direction sequences in a historical time period, and taking the wind speed and direction at the next adjacent moment in the historical time period as numerical labels; inputting the wind speed sequences and wind direction sequences in the historical time period into the time series prediction model to obtain an output result; and performing multiple iterations on the time series prediction model based on the mean square error loss between the output result and the numerical labels until the number of iterations is greater than a preset number or the mean square error loss is less than a preset loss, and the training is completed.

[0011] Preferably, the stability of the connection part The attention degree of each connection part is satisfying the relationship: ; is the connection part The ranking in the ranking result, is the number of each connection part.

[0012] The higher the ranking of a connection part in the ranking result, the greater the force of the wind on the connection part, and the more likely the connection part is to be at risk of falling due to the large wind force. Therefore, a higher attention degree is assigned to the connection part with a higher ranking.

[0013] Preferably, the training method of the ranking model comprises: collecting a plurality of training samples, the training samples comprising attitude information and a predicted wind direction, and force values of each connection part under the attitude information and the predicted wind direction, arranging each connection part in descending order of force values to obtain a ranking label; inputting the training samples into the ranking model to obtain a ranking result, and iteratively training the ranking model according to the ranking loss between the ranking result and the ranking label until the ranking loss is less than a loss threshold or the number of iterations is greater than a number threshold, to obtain a trained ranking model.

[0014] The ranking model is trained so that the ranking model can learn the ranking result under different combinations of attitude information and predicted wind direction, ensuring the accuracy of the attention degree of each connection part.

[0015] Preferably, the predicted wind load of the next moment satisfies the relationship:

[0016] ; wherein, is the air density, is the windward area, is the drag coefficient of the battery panel, is the predicted wind speed, the windward area is , is the area of the battery panel, and are the normal vector of the battery panel and the predicted wind direction, respectively, and the normal vector of the battery panel is related to the attitude information.

[0017] Preferably, the local risk value of the connection part is: ; is the predicted wind load, is the stability of the connection part . Preferably, the overall risk value

[0018] is: ​; is the total number of connection parts, is the connection part of the attention, is the sum of the attention of each connection part, is the local risk value of the connection part .

[0019] Since the predicted wind load of the battery panel is taken as the force of the wind on each connection part in the process of calculating the local risk value, the difference between the force of each connection part under the attitude information and the predicted wind direction is not considered, and the risk value of each connection part is weighted and summed by using the normalized attention, to obtain the overall risk value, so that the overall risk value can truly and accurately reflect the influence of the wind on the photovoltaic support.

[0020] The second aspect of the application also provides an anti-wind early warning system for a flexible photovoltaic support, comprising a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the anti-wind early warning method for a flexible photovoltaic support according to the first aspect of the application is realized.

[0021] The technical solution of the application has the following beneficial technical effects:

[0022] The photovoltaic support comprises a plurality of connection parts, and as the use time of the photovoltaic support increases, the connection parts will be loose to different degrees, and the different degrees of looseness will cause differences in the wind resistance of the connection parts. Therefore, the stability of each connection part is determined according to the vibration data of each connection part, so as to obtain accurate early warning results subsequently; further, the attitude information of the photovoltaic support at the current moment and the predicted wind direction at the next moment are input into a sorting model to determine the size relationship of the forces between the connection parts, and a larger attention is allocated to the connection part with a larger force; the local risk value of each connection part is calculated according to the predicted wind load at the next moment and the stability of each connection part, and the local risk value is used to represent the influence degree of the predicted wind load at the next moment on each connection part; the risk values of the connection parts are weighted and summed by using the normalized attention, to obtain the overall risk value; the anti-wind early warning is realized according to the overall risk value, the wind resistance of each connection part is comprehensively considered, and the overall risk value is accurately quantified according to the force of each connection part at the next moment, so as to improve the accuracy of the anti-wind early warning of the flexible photovoltaic support. BRIEF DESCRIPTION OF DRAWINGS

[0023] Figure 1 is a flowchart of an anti-wind early warning method for a flexible photovoltaic support according to an embodiment of the application.

[0024] Figure 2 is a structural block diagram of an anti-wind early warning system for a flexible photovoltaic support according to an embodiment of the application. DETAILED DESCRIPTION

[0025] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0026] According to a first aspect of the present application, the present application provides a wind resistance early warning method of a flexible photovoltaic support. Figure 1 is a flow chart of a wind resistance early warning method of a flexible photovoltaic support according to an embodiment of the present application. As shown in Figure 1 , the wind resistance early warning method of the flexible photovoltaic support includes steps S101 to S106, which are described in detail below.

[0027] S101, determining the stability of each connecting part according to the vibration data of the vibration sensor at each connecting part of the photovoltaic support.

[0028] In one embodiment, the photovoltaic support is formed as a whole through connecting parts such as bearings, hinges and screws, and vibration sensors are arranged at each connecting part to collect vibration data of each connecting part within a preset time period. According to the vibration data, the loosening degree of each connecting part can be determined. If a vibration data shows a large vibration, it indicates that the connecting part is loose, and the stability of the connecting part is worse.

[0029] Specifically, determining the stability of each connecting part according to the vibration data of the vibration sensor at each connecting part of the photovoltaic support includes: performing Fourier transform on the vibration data to obtain energy values of each connecting part at different vibration frequencies; taking the normalized energy values as weights to calculate the center vibration frequency of each connecting part, and the stability of each connecting part is negatively correlated with the center vibration frequency.

[0030] Among them, the center vibration frequency of the connecting part satisfies the relationship:

[0031] ; is the maximum value of the vibration frequency, is the energy value of the vibration frequency , and is the sum of energy values of all vibration frequencies. When the stability of the connecting part is large, the energy value is concentrated in the low frequency region, and the center vibration frequency is a small value. When the connecting part appears loose, cracks or stiffness changes, the energy value spreads to the medium and high frequency region, and the center vibration frequency increases. Therefore, the higher the center vibration frequency, the more unstable the corresponding connecting part. That is to say, the stability of each connecting part is negatively correlated with the center vibration frequency. Specifically, the stability of the connecting part

[0032] ; is the center vibration frequency of the connecting part , and is the maximum vibration frequency.

[0033] In this way, the stability of each connecting part is accurately measured according to the vibration data of the vibration sensor of each connecting part. The worse the stability, the greater the possibility of loose, cracks or stiffness changes in the corresponding connecting part, and the weaker the wind resistance of the corresponding connecting part. This provides a data basis for subsequent wind resistance warning.

[0034] S102, obtaining the predicted wind speed and direction of the next moment by using a time series prediction model.

[0035] In an embodiment, the wind speed sequence and direction sequence in a preset time period are collected, and the preset time period is the current moment and a plurality of historical moments before the current moment. The wind speed sequence and direction sequence are input into the time series prediction model to obtain the predicted wind speed and direction of the next moment. In the embodiment of the present application, the number of moments contained in the preset time period is 15.

[0036] The time series prediction model can be a recurrent neural network such as LSTM or GRU.

[0037] It should be noted that before obtaining the predicted wind speed and direction of the next moment by using the time series prediction model, the time series prediction model also needs to be trained. The training method of the time series prediction model includes: collecting the wind speed sequence and direction sequence in a historical time period, and taking the wind speed and direction of the next adjacent moment in the historical time period as a numerical label; inputting the wind speed sequence and direction sequence in the historical time period into the time series prediction model to obtain an output result; based on the mean square error loss between the output result and the numerical label, the time series prediction model is iterated for multiple times, until the number of iterations is greater than a preset number, or the mean square error loss is less than a preset loss, the training is completed. The value of the preset number is 300; the value of the preset loss is 0.01.

[0038] In this way, the time series prediction model learns the changes of wind speed and direction, and accurately obtains the predicted wind speed and direction of the next moment. ​​

[0039] S103, taking the orientation of the photovoltaic support and the inclination angle of the solar panel at the current time as the attitude information, calculating the attention degree of each connection part according to the attitude information and the predicted wind direction.

[0040] In one embodiment, the photovoltaic support can adjust the orientation of the photovoltaic support by rotation, and the solar panel mounted on the photovoltaic support can adjust the inclination angle of the solar panel by adjusting the pitch angle, and the orientation of the photovoltaic support and the inclination angle of the solar panel at the current time are taken as the attitude information.

[0041] It can be understood that in the case where the wind direction remains unchanged, the stress conditions of the photovoltaic support under different attitude information are different, which in turn leads to different stress sizes of each connection part under different attitude information, and the connection part with larger stress is more likely to have a risk of falling due to larger wind force. Therefore, the stress size of each connection part is related to the attitude information and the predicted wind direction, and a larger attention degree should be allocated to the connection part with larger stress.

[0042] Specifically, calculating the attention degree of each connection part according to the attitude information and the predicted wind direction includes: inputting the attitude information and the predicted wind direction into a ranking model, outputting a ranking result of the stress size of each connection part, and the attention degree of each connection part is negatively correlated with the ranking in the ranking result.

[0043] Wherein, the attention degree of the connection part satisfies the relationship: ; The ranking of the connection part in the ranking result, is the number of each connection part.

[0044] Exemplarily, if the photovoltaic support includes A, B, C, D and E in total, i.e. 5 connection parts, and the connection parts are arranged in descending order of stress size to obtain a ranking result, the ranking result is , that is to say, the ranking of D is 1 and the ranking of C is 5; then the attention degree of the connection part D is , and the attention degree of the connection part C is . A larger attention degree is allocated to the connection part with a higher ranking.

[0045] Wherein, the ranking model adopts a fully connected neural network, including an input layer, a hidden layer and an output layer; the input layer is used to receive the attitude information of the photovoltaic support at the current time and the predicted wind direction at the next time, the hidden layer is used to extract features of the attitude information and the predicted wind direction, and the feature extraction result is input to the output layer to obtain the ranking score of each connection part, and the ranking result is obtained in descending order of the ranking score.

[0046] ​In one embodiment, before using the ranking model to obtain the attention level of each connection part, the ranking model needs to be trained to ensure the accuracy of the attention level for each connection part. The training method of the ranking model includes: collecting multiple training samples, which include attitude information and predicted wind direction, as well as the force values ​​of each connection part under the attitude information and predicted wind direction; arranging each connection part in descending order of force values ​​to obtain ranking labels; inputting the training samples into the ranking model to obtain ranking results; iteratively training the ranking model based on the ranking loss between the ranking results and the ranking labels until the ranking loss is less than a loss threshold, or the number of iterations is greater than a number threshold, to obtain the trained ranking model. The number threshold is set to 300; the loss threshold is set to 0.01.

[0047] The sorting loss can be any existing sorting loss such as ListNet loss or listMLE loss, and this application does not impose any restrictions.

[0048] Thus, based on the current attitude information of the photovoltaic support and the predicted wind direction for the next moment, the attention level of each connection part is predicted. The greater the attention level, the greater the force on the corresponding connection part under the attitude information and predicted wind direction, and the more likely the connection part is to tilt.

[0049] S104. Calculate the predicted wind load for the next moment based on the predicted wind speed, predicted wind direction, and attitude information, and calculate the local risk value of each connection point. The local risk value is positively correlated with the predicted wind load and negatively correlated with stability.

[0050] In one embodiment, the local risk value is used to characterize the impact of the predicted wind load at the next moment on each connection point. Compared to the solar panels mounted on the photovoltaic support, the surface area of ​​the photovoltaic support and the wind is much smaller than the surface area of ​​the solar panels and the wind. The force of wind on the photovoltaic support is mainly caused by the contact surface between the solar panels and the wind. Therefore, to reduce the amount of calculation, when considering the force of wind on the photovoltaic support, the predicted wind load of the solar panels is regarded as the force of wind on the photovoltaic support.

[0051] Predicted wind load at the next moment Satisfying the relation: ;in, air density, For windward area, The drag coefficient of the solar panel. To predict wind speed, the windward area for , The area of ​​the solar panel. and These are the normal vector of the solar panel and the predicted wind direction, respectively. The normal vector of the solar panel is related to the attitude information.

[0052] wherein the drag coefficient of the solar panel can be obtained according to the model of the solar panel, which is not described herein.

[0053] Thus, the predicted wind load at the next moment is obtained, which can be regarded as the force of the wind on the photovoltaic support at the next moment.

[0054] In one embodiment, the greater the predicted wind load, the greater the force borne by each connection site, and the greater the local risk value of each connection site, so the local risk value of each connection site is positively correlated with the predicted wind load; the greater the stability of the connection site, the greater the wind resistance of the connection site, and the smaller the local risk value of the connection site, so the local risk value of each connection site is positively correlated with the stability.

[0055] connection site local risk value is: ; is the predicted wind load, is the stability of the connection site .

[0056] Thus, the local risk value of each connection site at the next moment is obtained by comprehensively considering the attitude information of the photovoltaic support at the current moment, and the predicted wind speed and the predicted wind direction at the next moment, and the local risk value is used to represent the influence degree of the predicted wind load at the next moment on each connection site.

[0057] S105, the overall risk value is obtained by weighted sum of the risk values of each connection site using the normalized attention degree.

[0058] In one embodiment, since the predicted wind load of the solar panel is directly used as the force of the wind on each connection site in the process of calculating the local risk value, the difference between the attitude information and the force of each connection site under the predicted wind direction is not considered, therefore, the overall risk value is obtained by weighted sum of the risk values of each connection site using the normalized attention degree. The overall risk value can truly and accurately reflect the influence of the wind on the photovoltaic support.

[0059] overall risk value is: ; is the total number of connection sites, is the attention degree of the connection site , is the sum of the attention degrees of each connection site, is the local risk value of the connection site .

[0060] S106, when the overall risk value is greater than the risk threshold, an early warning is issued.

[0061] In one embodiment, the risk threshold is set to 0.8 times the maximum safe wind load.

[0062] According to a second aspect of this application, a wind-resistant early warning system for a flexible photovoltaic support is also provided. Figure 2 This is a structural block diagram of a wind-resistant early warning system for a flexible photovoltaic support according to an embodiment of this application. Figure 2 As shown, the system 50 includes a processor and a memory. The memory stores computer program instructions, which, when executed by the processor, implement a wind-resistant early warning method for a flexible photovoltaic support according to the first aspect of this application. The system also includes other components well-known to those skilled in the art, such as a communication bus and a communication interface. Their configuration and functions are known in the art and will not be described further here.

[0063] It should be noted that, for those skilled in the art, several modifications and improvements can be made without departing from the concept of this application, and these all fall within the scope of protection of this application.

Claims

1. A wind resistance warning method for a flexible photovoltaic support, characterized in that, The early warning method comprises: According to the vibration data of the vibration sensor of each connection part of the photovoltaic support, the stability of each connection part is determined; the predicted wind speed and the predicted wind direction at the next moment are obtained by using the time sequence prediction model; The orientation of the photovoltaic support and the inclination angle of the cell panel at the current moment are taken as the attitude information, and the attention degree of each connection part is calculated according to the attitude information and the predicted wind direction, including: inputting the attitude information and the predicted wind direction into the sorting model to output the sorting result of the force size of each connection part, the attention degree of each connection part is negatively correlated with the ranking in the sorting result, and in the sorting result, each connection part is sorted in descending order of force size; According to the predicted wind speed, the predicted wind direction and the attitude information, the predicted wind load at the next moment is calculated, and the local risk value of each connection part is calculated, the local risk value is positively correlated with the predicted wind load and negatively correlated with the stability; The risk values of each connection part are weighted and summed by using the normalized attention degree to obtain the overall risk value; In response to the overall risk value being greater than the risk threshold, an early warning is issued.

2. The wind warning method of a flexible photovoltaic support according to claim 1, characterized in that, According to the vibration data of the vibration sensor of each connection part of the photovoltaic support, the stability of each connection part is determined, including: Performing Fourier transform on the vibration data to obtain the energy value of each connection part at different vibration frequencies; the normalized energy value is taken as the weight to calculate the center vibration frequency of each connection part, and the stability of each connection part is negatively correlated with the center vibration frequency.

3. The wind alert method of claim 2, wherein, Connection site Stability of the connection site Is: ; for the connection site of the center of the vibration frequency, is the maximum value of the vibration frequency.

4. The wind warning method of a flexible photovoltaic support according to claim 1, characterized in that, The training method of the time sequence prediction model comprises: Collecting wind speed sequence and wind direction sequence in a historical time period, and taking the wind speed and direction of the next adjacent moment in the historical time period as numerical labels; The wind speed sequence and wind direction sequence in the historical time period are input into the time sequence prediction model to obtain the output result; the time sequence prediction model is iterated multiple times based on the mean square error loss between the output result and the numerical label, until the iteration number is greater than the preset number, or the mean square error loss is less than the preset loss, the training is completed.

5. The wind alert method for a flexible photovoltaic support according to claim 1, characterized in that, Connection points The attention degree is The relationship is satisfied: ; The connection points The rank in the ranking result, The number of connection points.

6. The wind alert method for a flexible photovoltaic support according to claim 1, wherein The training method of the sorting model comprises: Collecting a plurality of training samples, the training samples comprising attitude information and predicted wind direction, and force values of each connection part under the attitude information and predicted wind direction, arranging each connection part in descending order of force value to obtain a sorting label; The training sample is input into the sorting model to obtain the sorting result, and the sorting model is iterated trained according to the sorting loss between the sorting result and the sorting label, until the sorting loss is less than the loss threshold, or the iteration number is greater than the number threshold, the trained sorting model is obtained.

7. The wind alert method of claim 1, wherein, predicted wind load for the next time instant satisfies the relation: ; wherein, is the air density, is the windward area, is the drag coefficient of the panel, is the predicted wind speed, the windward area is the , is the panel area, and are the normal vector of the panel and the predicted wind direction, respectively, the normal vector of the panel being related to the attitude information.

8. The wind warning method of a flexible photovoltaic support according to claim 1, characterized in that, Connection points The local risk value is: ; is the predicted wind load, is the stability of the connection points .

9. The wind alert method of claim 1, wherein, Overall risk value is: ; is the total number of connection sites, is the connection site of interest, is the sum of the connection site of interest local risk value.

10. A wind alert system for a flexible photovoltaic racking, characterized in that, A processor and a memory are included, the memory stores computer program instructions, when the computer program instructions are executed by the processor, a wind resistance early warning method for a flexible photovoltaic support according to any one of claims 1 to 9 is realized.

Images