Photovoltaic power supply breeding detection method and system based on AI identification technology

Through the photovoltaic power supply aquaculture detection method based on AI recognition technology, sensors and deep learning algorithms are used to monitor and diagnose the status of the photovoltaic system in real time, solving the problem of low maintenance and fault diagnosis efficiency in the existing technology, and achieving efficient and stable operation of the photovoltaic power supply system .

CN118971799APending Publication Date: 2024-11-15FUJIAN MINGAO ELECTRIC POWER ENERGY GROUP CO LTD
View PDF 0 Cites 2 Cited by

Patent Information

Application Number
CN202411070666.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-06
Publication Date
2024-11-15

AI Technical Summary

Technical Problem

The existing photovoltaic power supply system maintenance methods and fault diagnosis mode are inefficient, and it is impossible to quickly and effectively detect faults and provide countermeasures, which reduces the working efficiency of the system.

Method used

Using photovoltaic powered aquaculture detection method based on AI recognition technology, the state of photovoltaic panels is monitored in real time by installing sensors, data processing and deep learning algorithms to train models, monitor and diagnose abnormal situations in real time, issue alarms and provide decision-making suggestions.

Benefits of technology

It improves the intelligence level and working efficiency of the photovoltaic power supply system, can detect abnormalities in a timely manner and provide handling suggestions, improving the stability of the system.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118971799A_ABST
    Figure CN118971799A_ABST
Patent Text Reader

Abstract

The invention discloses a photovoltaic power supply breeding detection method and system based on an AI identification technology, and the method comprises the following steps: S1, installing a sensor, and monitoring the working states of a photovoltaic panel in real time, such as illumination intensity, voltage, current, temperature and other data; the collected data is transmitted to the data processing unit and the monitoring software through a wireless network or wired connection; s2, the data processing unit preprocesses the collected data, the preprocessing content comprises data cleaning, denoising and calibration, and the processed data is divided into a training set and a test set; according to the invention, based on the AI identification technology, the operation state of the photovoltaic power supply system can be monitored, when an abnormal condition is detected, an alarm can be timely given to remind personnel and corresponding decision suggestions can be given, a user can conveniently and timely carry out processing, the automatic detection and troubleshooting mode, the intelligent degree and the working efficiency are effectively improved, and the user experience is improved. Therefore, the working stability of the photovoltaic power supply system is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of photovoltaic power supply aquaculture detection, and in particular to a photovoltaic power supply aquaculture detection method and system based on AI recognition technology. Background Art

[0002] The photovoltaic power supply system uses solar energy to generate electricity, has no emissions or pollution, and is one of the clean energy sources. The application of photovoltaic power supply systems in farms can reduce energy consumption, reduce the impact on the environment, and meet the requirements of sustainable development. In order to improve the stability of the photovoltaic power supply system, the existing photovoltaic power supply system needs to be maintained frequently. The maintenance methods and fault diagnosis modes in the existing technology are passive and blindly performed by maintenance personnel. Once a fault occurs, it is necessary to check, confirm and eliminate it one by one. The troubleshooting efficiency and intelligence level are low. It is impossible to quickly and effectively detect the status and faults of the photovoltaic power supply system and quickly give countermeasures. At the same time, it reduces the working efficiency of the photovoltaic power supply system. For this reason, we propose a photovoltaic power supply aquaculture detection method and system based on AI recognition technology. Summary of the invention

[0003] Based on the technical problems existing in the background technology, the present invention proposes a photovoltaic-powered farming detection method and system based on AI recognition technology.

[0004] The present invention proposes a photovoltaic power supply aquaculture detection method based on AI recognition technology, comprising the following steps:

[0005] S1: Install sensors to monitor the working status of photovoltaic panels in real time, such as light intensity, voltage, current, temperature and other data, and transmit the collected data to the data processing unit and monitoring software through wireless network or wired connection;

[0006] S2: The data processing unit preprocesses the collected data, including data cleaning, denoising and calibration, and divides the processed data into a training set and a test set;

[0007] S3: Build an AI model and use deep learning algorithms to train and optimize the AI ​​model in combination with the training set and test set in S2;

[0008] S4: Deploy the trained AI model to the detection system to monitor the operating status of the photovoltaic power supply system in real time. When the detection system detects an abnormal situation, such as insufficient light, abnormal voltage or current fluctuation, it immediately issues an alarm to notify relevant personnel;

[0009] S5: Users can view the detection data and images of the photovoltaic power supply system in real time through the monitoring software;

[0010] S6: Based on the results of AI model recognition, the detection system gives corresponding decision suggestions through the decision-making mechanism;

[0011] S7: Based on the decision suggestions in S6, the user can take corresponding measures, such as repairing equipment, adjusting the angle of photovoltaic panels, etc., to ensure the normal operation of the photovoltaic power supply system.

[0012] Preferably, in said S1, the installed sensors include voltage sensors, current sensors, temperature sensors, light sensors, radiation sensors, solar irradiance sensors and weather sensors, wherein the voltage sensor is used to measure the DC voltage output by the photovoltaic panel, and can monitor the output of the photovoltaic cell group of the photovoltaic panel;

[0013] The current sensor is used to measure the current output by the photovoltaic panel. It can monitor the change and stability of the current and evaluate the power generation capacity of the photovoltaic panel.

[0014] Temperature sensors are used to monitor the temperature of photovoltaic panels, which can help evaluate the working efficiency and power generation performance of photovoltaic panels and prevent overheating;

[0015] Light sensors are used to measure light intensity, which can evaluate the light received by photovoltaic panels and guide the adjustment and layout of photovoltaic panels;

[0016] Radiation sensors are used to detect the intensity of radiation energy, help evaluate the radiation received by photovoltaic panels, and understand the energy conversion efficiency of photovoltaic panels;

[0017] Solar irradiance sensors are used to measure solar radiation intensity, helping to assess the light received by photovoltaic panels and changes in solar radiation;

[0018] Weather sensors are used to detect weather conditions, such as sunny, cloudy, overcast, etc., which affect the lighting conditions and the power generation efficiency of photovoltaic panels.

[0019] Preferably, in S2, the training set occupies most of the data, including a large amount of real-time monitoring data, such as light intensity, voltage, current, temperature and other data, which are used to train the AI ​​model, and the test set includes unseen, new data samples, which are used to evaluate the performance of the AI ​​model.

[0020] Preferably, in S3, the AI ​​model is trained and optimized using a deep learning algorithm in combination with a training set and a test set, and the logical steps are as follows:

[0021] S301: Select a convolutional neural network as the initial model of the AI ​​model and initialize the parameters of the model;

[0022] S302: input the training set data into the initial model selected in S301, train the model, and optimize the model parameters through the back propagation algorithm so that the model can minimize the loss function;

[0023] Suppose we have a loss function L, which is the error between the model's prediction and the true label value. The loss function is usually calculated based on the model's output and the label value.

[0024] Its expression is: [L(y_{pred},y_{true})=\frac{1}{2}\sum_{i}(y_{pred}^{(i)}-y_{true}^{(i)})^2], where (y_{pred}) is the model prediction value, (y_{true}) is the true label value, (i) is the sample index, and the loss function L is calculated based on its expression, and the model parameters are optimized based on the loss function value;

[0025] S303: Input the test set data into the trained model to verify and evaluate the model, evaluate the generalization ability of the model on unseen data, calculate the performance indicators of the model, including accuracy, precision and recall, and evaluate the performance of the model;

[0026] S304: Based on the evaluation results and performance indicators of the test set, optimize and adjust the deficiencies of the model to improve the generalization ability and accuracy of the model, thereby obtaining a trained AI model.

[0027] Preferably, in S303, the accuracy rate refers to the proportion of samples correctly predicted by the AI ​​model, which is usually used to evaluate the overall classification accuracy, and its expression is [Accuracy=\frac{TP+TN}{TP+TN+FP+FN}], wherein TP represents the number of samples that are actually positive and predicted as positive by the model, TN represents the number of samples that are actually negative and predicted as negative by the model, FP represents the number of samples that are actually negative but misclassified as positive by the model, and FN represents the number of samples that are actually positive but misclassified as negative by the model;

[0028] Among them, data such as normal working state, high power generation state and no shadow occlusion state belong to the positive class samples. Positive class samples are the photovoltaic power supply system states that we hope the AI ​​model can correctly identify and judge. Abnormal state, fault state and shadow occlusion state are in the negative class samples. We hope that the AI ​​model can correctly exclude or identify them to ensure the normal operation of the system.

[0029] Precision refers to the proportion of samples that are actually positive among those predicted as positive by the AI ​​model. It is used to evaluate the prediction accuracy of the AI ​​model. The calculation formula is: [Precision = \frac{TP}{TP+FP}];

[0030] The recall rate refers to the ratio of the number of samples correctly predicted as positive by the AI ​​model to the number of all true positive samples. It is used to evaluate the recall rate of the AI ​​model. The calculation formula is: [Recall=\frac{TP}{TP+FN}].

[0031] Preferably, the logical steps of S6 are as follows:

[0032] S601: Confirm the detected abnormality according to the result of AI model recognition, including information such as abnormality type, location, severity, etc.;

[0033] S602: Based on the abnormal situation detected in S601, the detection system gives corresponding suggestions or measures according to preset rules and decision strategies. The decision suggestions may include adjusting the angle of the photovoltaic panel, replacing damaged parts, clearing shadows, and other action plans to help users make decisions and handle them;

[0034] S603: Depending on the abnormal situation, the detection system will provide personalized suggestions and make customized operational suggestions for the photovoltaic power supply system according to the specific situation, and notify the relevant personnel, such as farm owners, technicians, etc., of the decision suggestions in real time to ensure that the abnormal situation is handled and resolved in a timely manner.

[0035] Preferably, in said S602, the preset rules include light intensity being lower than a certain value, voltage being out of a normal range, and power generation suddenly decreasing;

[0036] The decision strategy formulates corresponding decision strategies according to different abnormal situations. For damaged photovoltaic panels, timely replacement or repair may be recommended; for shadow obstruction, it may be recommended to remove obstacles or adjust the position of photovoltaic panels, and set priorities and urgency to ensure that the detection system can quickly respond to important and urgent abnormal situations.

[0037] The present invention also proposes a photovoltaic power supply aquaculture detection system based on AI recognition technology, including a data acquisition module, a data processing module, an AI recognition module, an early warning module, a real-time monitoring module, a user feedback and management module, and a photovoltaic power supply system;

[0038] The data acquisition module is connected to the data processing module, the real-time monitoring module and the photovoltaic power supply system, the AI ​​recognition module is connected to the data processing module, the early warning module and the real-time monitoring module, and the user feedback and management module is connected to the real-time monitoring module and the photovoltaic power supply system.

[0039] Preferably, the data acquisition module is used to monitor the working status of the photovoltaic panel in real time and collect data of the photovoltaic panel;

[0040] The data processing module is used to pre-process the collected data;

[0041] The AI ​​recognition module is used to establish an AI model and deploy the AI ​​model to the detection system to monitor the operating status of the photovoltaic power supply system in real time, and when an abnormality is detected, an alarm is issued through the early warning module to notify relevant personnel;

[0042] The real-time monitoring module is used to display the results processed by the AI ​​recognition module to the user in real time, provide monitoring data and alarm information, and the user can view the status of the breeding scene and photovoltaic power supply system through the monitoring software and make corresponding processing and decisions in time;

[0043] The user feedback and management module is used to allow the user to provide feedback and management based on the monitoring results. The user feedback and management module is combined with the monitoring software to help the user to manage and maintain the photovoltaic power supply system in a timely manner.

[0044] Compared with the prior art, the present invention has the following beneficial effects:

[0045] The present invention is based on AI recognition technology, which can monitor the operating status of the photovoltaic power supply system. When an abnormal situation is detected, it can promptly alarm and remind personnel and give corresponding decision-making suggestions, so as to facilitate users to handle it in time. The automatic detection and troubleshooting method has effectively improved the intelligence level and work efficiency, thereby improving the working stability of the photovoltaic power supply system. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 This is a flow chart of a photovoltaic power supply aquaculture detection method based on AI recognition technology proposed by the present invention;

[0047] Figure 2 This is a block diagram of a photovoltaic-powered farming detection system based on AI recognition technology proposed by the present invention. DETAILED DESCRIPTION

[0048] The present invention will be further explained below in conjunction with specific embodiments.

[0049] Embodiment 1

[0050] Reference Figure 1 This embodiment proposes a photovoltaic power supply farming detection method based on AI recognition technology, including the following steps:

[0051] S1: Install sensors to monitor the working status of photovoltaic panels in real time, such as light intensity, voltage, current, temperature and other data, and transmit the collected data to the data processing unit and monitoring software through a wireless network or wired connection. The installed sensors include voltage sensors, current sensors, temperature sensors, light sensors, radiation sensors, solar irradiance sensors and weather sensors. The voltage sensor is used to measure the DC voltage output by the photovoltaic panel, which can monitor the output of the photovoltaic battery group of the photovoltaic panel;

[0052] The current sensor is used to measure the current output by the photovoltaic panel. It can monitor the change and stability of the current and evaluate the power generation capacity of the photovoltaic panel.

[0053] Temperature sensors are used to monitor the temperature of photovoltaic panels, which can help evaluate the working efficiency and power generation performance of photovoltaic panels and prevent overheating;

[0054] Light sensors are used to measure light intensity, which can evaluate the light received by photovoltaic panels and guide the adjustment and layout of photovoltaic panels;

[0055] Radiation sensors are used to detect the intensity of radiation energy, help evaluate the radiation received by photovoltaic panels, and understand the energy conversion efficiency of photovoltaic panels;

[0056] Solar irradiance sensors are used to measure solar radiation intensity, helping to assess the light received by photovoltaic panels and changes in solar radiation;

[0057] Weather sensors are used to detect weather conditions, such as sunny, cloudy, overcast, etc., which affect the lighting conditions and the power generation efficiency of photovoltaic panels;

[0058] S2: The data processing unit preprocesses the collected data, including data cleaning, denoising and calibration, and divides the processed data into a training set and a test set. The training set occupies most of the data, including a large amount of real-time monitoring data, such as light intensity, voltage, current, temperature and other data, which is used to train the AI ​​model. The test set includes unseen and new data samples, which are used to evaluate the performance of the AI ​​model.

[0059] S3: Build an AI model and use the deep learning algorithm to train and optimize the AI ​​model in combination with the training set and test set in S2. The logical steps are as follows:

[0060] S301: Select a convolutional neural network as the initial model of the AI ​​model and initialize the parameters of the model;

[0061] S302: input the training set data into the initial model selected in S301, train the model, and optimize the model parameters through the back propagation algorithm so that the model can minimize the loss function;

[0062] Suppose we have a loss function L, which is the error between the model's prediction and the true label value. The loss function is usually calculated based on the model's output and the label value.

[0063] Its expression is: [L(y_{pred},y_{true})=\frac{1}{2}\sum_{i}(y_{pred}^{(i)}-y_{true}^{(i)})^2], where (y_{pred}) is the model prediction value, (y_{true}) is the true label value, (i) is the sample index, and the loss function L is calculated based on its expression, and the model parameters are optimized based on the loss function value;

[0064] S303: Input the test set data into the trained model to verify and evaluate the model, evaluate the generalization ability of the model on unseen data, calculate the performance indicators of the model, including accuracy, precision and recall, and evaluate the performance of the model;

[0065] Accuracy refers to the proportion of samples correctly predicted by the AI ​​model, which is usually used to evaluate the overall classification accuracy. Its expression is [Accuracy=\frac{TP+TN}{TP+TN+FP+FN}], where TP represents the number of samples that are actually positive and predicted as positive by the model, TN represents the number of samples that are actually negative and predicted as negative by the model, FP represents the number of samples that are actually negative but misclassified as positive by the model, and FN represents the number of samples that are actually positive but misclassified as negative by the model.

[0066] Among them, data such as normal working state, high power generation state and no shadow occlusion state belong to the positive class samples. Positive class samples are the photovoltaic power supply system states that we hope the AI ​​model can correctly identify and judge. Abnormal state, fault state and shadow occlusion state are in the negative class samples. We hope that the AI ​​model can correctly exclude or identify them to ensure the normal operation of the system.

[0067] Precision refers to the proportion of samples that are actually positive among those predicted as positive by the AI ​​model. It is used to evaluate the prediction accuracy of the AI ​​model. The calculation formula is: [Precision = \frac{TP}{TP+FP}];

[0068] Recall rate refers to the ratio of the number of samples correctly predicted as positive by the AI ​​model to the number of all true positive samples. It is used to evaluate the recall rate of the AI ​​model. The calculation formula is: [Recall = \frac{TP}{TP+FN}];

[0069] S304: Based on the evaluation results and performance indicators of the test set, optimize and adjust the deficiencies of the model to improve the generalization ability and accuracy of the model, thereby obtaining a trained AI model;

[0070] S4: Deploy the trained AI model to the detection system to monitor the operating status of the photovoltaic power supply system in real time. When the detection system detects an abnormal situation, such as insufficient light, abnormal voltage or current fluctuation, it immediately issues an alarm to notify relevant personnel;

[0071] S5: Users can view the detection data and images of the photovoltaic power supply system in real time through the monitoring software;

[0072] S6: Based on the results of AI model recognition, the detection system gives corresponding decision suggestions through the decision-making mechanism. The logical steps are as follows:

[0073] S601: Confirm the detected abnormality according to the result of AI model recognition, including information such as abnormality type, location, severity, etc.;

[0074] S602: Based on the abnormal situation detected in S601, the detection system gives corresponding suggestions or measures according to preset rules and decision strategies. The decision suggestions may include adjusting the angle of the photovoltaic panel, replacing damaged parts, clearing shadows, and other action plans to help users make decisions and handle them. The preset rules include light intensity below a certain value, voltage outside the normal range, and sudden drop in power generation, etc.;

[0075] Decision-making strategies are formulated according to different abnormal situations. For damaged photovoltaic panels, timely replacement or repair may be recommended; for shadow occlusion, obstacles may be removed or the position of photovoltaic panels may be adjusted. Priorities and urgency levels are set to ensure that the detection system can quickly respond to important and urgent abnormal situations.

[0076] S603: Depending on the abnormal situation, the detection system will provide personalized suggestions, customize the operation suggestions for the photovoltaic power supply system according to the specific situation, and notify the relevant personnel, such as farm owners and technicians, of the decision suggestions in real time to ensure that the abnormal situation is handled and resolved in a timely manner;

[0077] S7: Based on the decision suggestions in S6, the user can take corresponding measures, such as repairing equipment, adjusting the angle of photovoltaic panels, etc., to ensure the normal operation of the photovoltaic power supply system;

[0078] This embodiment is based on AI recognition technology, can monitor the operating status of the photovoltaic power supply system, and can promptly alarm and remind personnel and give corresponding decision-making suggestions when abnormal situations are detected, so as to facilitate users to handle them in time. The automatic detection and troubleshooting method has effectively improved the degree of intelligence and work efficiency, thereby improving the working stability of the photovoltaic power supply system.

[0079] Embodiment 2

[0080] Reference Figure 2 ,This embodiment proposes a photovoltaic-powered farming detection system based on AI recognition technology, including a data acquisition module, a data processing module, an AI recognition module, an early warning module, a real-time monitoring module, a user feedback and management module, and a photovoltaic power supply system;

[0081] The data acquisition module is connected with the data processing module, the real-time monitoring module and the photovoltaic power supply system, wherein the data acquisition module is used to monitor the working status of the photovoltaic panel in real time and collect the data of the photovoltaic panel, and the data processing module is used to pre-process the collected data;

[0082] The AI ​​recognition module is connected with the data processing module, the early warning module and the real-time monitoring module. The AI ​​recognition module is used to establish an AI model and deploy the AI ​​model to the detection system to monitor the operating status of the photovoltaic power supply system in real time. When an abnormality is detected, the early warning module sends an alarm to notify relevant personnel.

[0083] The user feedback and management module is connected to the real-time monitoring module and the photovoltaic power supply system. The real-time monitoring module is used to display the results processed by the AI ​​recognition module to the user in real time, provide monitoring data and alarm information, and the user can view the status of the breeding scene and the photovoltaic power supply system through the monitoring software, and make corresponding processing and decisions in time;

[0084] The user feedback and management module is used to allow users to provide feedback and management based on monitoring results. The user feedback and management module is combined with the monitoring software to help users manage and maintain the photovoltaic power supply system in a timely manner.

[0085] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A photovoltaic power supply farming detection method based on AI recognition technology, It is characterized in that The following steps are involved: S1: Install sensors to monitor the working status of photovoltaic panels in real time, and transmit the collected data to the data processing unit and monitoring software through wireless network or wired connection; S2: The data processing unit preprocesses the collected data, including data cleaning, denoising and calibration, and divides the processed data into a training set and a test set; S3: Build an AI model and use deep learning algorithms to train and optimize the AI ​​model in combination with the training set and test set in S2; S4: Deploy the trained AI model to the detection system to monitor the operating status of the photovoltaic power supply system in real time. When the detection system detects an abnormal situation, it immediately issues an alarm to notify relevant personnel; S5: Users can view the detection data and images of the photovoltaic power supply system in real time through the monitoring software; S6: Based on the results of AI model recognition, the detection system gives corresponding decision suggestions through the decision-making mechanism; S7: Based on the decision recommendations in S6, the user can take corresponding measures to ensure the normal operation of the photovoltaic power supply system.

2. According to claim 1, a photovoltaic power supply farming detection method based on AI recognition technology, It is characterized in that In said S1, the installed sensors include voltage sensors, current sensors, temperature sensors, light sensors, radiation sensors, solar irradiance sensors and weather sensors, wherein the voltage sensor is used to measure the DC voltage output by the photovoltaic panel, and can monitor the output of the photovoltaic cell group of the photovoltaic panel; The current sensor is used to measure the current output by the photovoltaic panel. It can monitor the change and stability of the current and evaluate the power generation capacity of the photovoltaic panel. Temperature sensors are used to monitor the temperature of photovoltaic panels, which can help evaluate the working efficiency and power generation performance of photovoltaic panels and prevent overheating; Light sensors are used to measure light intensity, which can evaluate the light received by photovoltaic panels and guide the adjustment and layout of photovoltaic panels; Radiation sensors are used to detect the intensity of radiation energy, help evaluate the radiation received by photovoltaic panels, and understand the energy conversion efficiency of photovoltaic panels; Solar irradiance sensors are used to measure solar radiation intensity, helping to assess the light received by photovoltaic panels and changes in solar radiation; Weather sensors are used to detect weather conditions.

3. According to the photovoltaic power supply breeding detection method based on AI recognition technology as described in claim 1, It is characterized in that In S2, the training set occupies most of the data, including a large amount of real-time monitoring data, which is used to train the AI ​​model, and the test set includes unseen, new data samples, which are used to evaluate the performance of the AI ​​model.

4. According to claim 1, a photovoltaic power supply farming detection method based on AI recognition technology, It is characterized in that In S3, the AI ​​model is trained and optimized using a deep learning algorithm combined with a training set and a test set. The logical steps are as follows: S301: Select a convolutional neural network as the initial model of the AI ​​model and initialize the parameters of the model; S302: input the training set data into the initial model selected in S301, train the model, and optimize the model parameters through the back propagation algorithm so that the model can minimize the loss function; Suppose we have a loss function L, which is the error between the model's prediction and the true label value. The loss function is usually calculated based on the model's output and the label value. Its expression is: [L(y_{pred},y_{true})=\frac{1}{2}\sum_{i}(y_{pred}^{(i)}-y_{true}^{(i)})^2], where (y_{pred}) is the model prediction value, (y_{true}) is the true label value, (i) is the sample index, and the loss function L is calculated based on its expression, and the model parameters are optimized based on the loss function value; S303: Input the test set data into the trained model, verify and evaluate the model, evaluate the generalization ability of the model on unseen data, calculate the performance indicators of the model, including accuracy, precision and recall, and evaluate the performance of the model; S304: Based on the evaluation results and performance indicators of the test set, optimize and adjust the deficiencies of the model to improve the generalization ability and accuracy of the model, thereby obtaining a trained AI model.

5. According to claim 4, a photovoltaic power supply farming detection method based on AI recognition technology, It is characterized in that In S303, accuracy refers to the proportion of samples correctly predicted by the AI ​​model, which is usually used to evaluate the overall classification accuracy. Its expression is [Accuracy = \frac{TP+TN}{TP+TN+FP+FN}], where TP represents the number of samples that are actually positive and predicted as positive by the model, TN represents the number of samples that are actually negative and predicted as negative by the model, FP represents the number of samples that are actually negative but misclassified as positive by the model, and FN represents the number of samples that are actually positive but misclassified as negative by the model; The data of normal working state, high power generation state and no shadow occlusion state belong to the positive class. The positive class samples are the PV power supply system states that we hope the AI ​​model can correctly identify and judge. The abnormal state, fault state and shadow occlusion state are in the negative class. The negative class samples are expected to be correctly excluded or identified by the AI ​​model to ensure the normal operation of the system. Precision refers to the proportion of samples that are actually positive among those predicted as positive by the AI ​​model. It is used to evaluate the prediction accuracy of the AI ​​model. The calculation formula is: [Precision = \frac{TP}{TP+FP}]; The recall rate refers to the ratio of the number of samples correctly predicted as positive by the AI ​​model to the number of all true positive samples. It is used to evaluate the recall rate of the AI ​​model. The calculation formula is: [Recall=\frac{TP}{TP+FN}].

6. According to claim 1, a photovoltaic power supply farming detection method based on AI recognition technology, It is characterized in that The logical steps of S6 are as follows: S601: Confirm the detected abnormality according to the result of AI model recognition, including the abnormality type, location, and severity; S602: Based on the abnormal situation detected in S601, the detection system gives corresponding suggestions or measures according to preset rules and decision strategies to help users make decisions and handle the situation; S603: Depending on the abnormal situation, the detection system will provide personalized suggestions, customize the operation suggestions for the photovoltaic power supply system according to the specific situation, and notify the relevant personnel of the decision suggestions in real time to ensure that the abnormal situation is handled and resolved in a timely manner.

7. According to claim 6, a photovoltaic power supply farming detection method based on AI recognition technology, It is characterized in that In S602, the preset rules include light intensity being lower than a certain value, voltage being out of a normal range, and power generation suddenly decreasing; The decision strategy formulates corresponding decision strategies according to different abnormal situations. For damaged photovoltaic panels, timely replacement or repair may be recommended; for shadow obstruction, it may be recommended to remove obstacles or adjust the position of photovoltaic panels, and set priorities and urgency to ensure that the detection system can quickly respond to important and urgent abnormal situations.

8. A photovoltaic power supply farming detection system based on AI recognition technology, used to implement any method described in claims 1-7, It is characterized in that It includes data acquisition module, data processing module, AI recognition module, early warning module, real-time monitoring module, user feedback and management module and photovoltaic power supply system; The data acquisition module is connected to the data processing module, the real-time monitoring module and the photovoltaic power supply system, the AI ​​recognition module is connected to the data processing module, the early warning module and the real-time monitoring module, and the user feedback and management module is connected to the real-time monitoring module and the photovoltaic power supply system.

9. A photovoltaic-powered farming detection system based on AI recognition technology according to claim 8, It is characterized in that The data acquisition module is used to monitor the working status of the photovoltaic panel in real time and collect data of the photovoltaic panel; The data processing module is used to pre-process the collected data; The AI ​​recognition module is used to establish an AI model and deploy the AI ​​model to the detection system to monitor the operating status of the photovoltaic power supply system in real time, and when an abnormality is detected, an alarm is issued through the early warning module to notify relevant personnel; The real-time monitoring module is used to display the results processed by the AI ​​recognition module to the user in real time, provide monitoring data and alarm information, and the user can view the status of the breeding scene and photovoltaic power supply system through the monitoring software and make corresponding processing and decisions in time; The user feedback and management module is used to allow the user to provide feedback and management based on the monitoring results. The user feedback and management module is combined with the monitoring software to help the user to manage and maintain the photovoltaic power supply system in a timely manner.

Citation Information

Cited By

  • Intelligent breeding management platform based on big data driving

    CN121599419A

  • A smart aquaculture management platform driven by big data

    CN121599419B