Photovoltaic power station intelligent data monitoring system based on Internet of Things

Abstract

The invention discloses a photovoltaic power station intelligent data monitoring system based on the Internet of Things, relates to the technical field of photovoltaic operation monitoring, and improves the accuracy of anomaly detection of photovoltaic modules. The method comprises the following steps: extracting a multi-dimensional data set from a historical component operation record, establishing a plurality of isolation detection trees, inputting the multi-dimensional data set into the isolation detection trees, further generating a plurality of normal numerical value combinations and abnormal numerical value combinations, and establishing a component operation analysis network according to the normal numerical value combinations and the abnormal numerical value combinations. And inputting the real-time module operation record into the module operation analysis network, further generating a predicted operation numerical value interval corresponding to a plurality of operation data of the photovoltaic module in the next operation detection period, and performing operation monitoring on the photovoltaic module according to the predicted operation numerical value interval.

Description

Technical Field

[0001] The present invention relates to the technical field of photovoltaic operation monitoring, and in particular to an intelligent data monitoring system for photovoltaic power stations based on the Internet of Things. Background Art

[0002] As global demand for clean energy continues to grow, photovoltaic power generation, as a sustainable energy solution, is becoming a vital component of the energy sector. As the core equipment of photovoltaic power generation systems, the stable operation of photovoltaic modules is crucial to ensuring the efficiency and reliability of photovoltaic power generation. However, during actual operation, photovoltaic modules are affected by various factors, such as ambient temperature, light intensity, and humidity. These factors can cause performance degradation or even failure of photovoltaic modules.

[0003] Traditional methods for monitoring photovoltaic (PV) panel operation rely primarily on manual inspections and simple sensor data collection. This approach is not only inefficient but also struggles to accurately and timely identify potential issues with PV panels. Furthermore, due to the complexity and multidimensionality of PV panel operation data, traditional data analysis methods struggle to effectively uncover patterns within the data, making it difficult to accurately predict and assess the operating status of PV panels. Therefore, we propose an IoT-based intelligent data monitoring system for photovoltaic power plants. Summary of the Invention

[0004] In order to solve the above technical problems, the purpose of the present invention is to provide an intelligent data monitoring system for photovoltaic power plants based on the Internet of Things.

[0005] In order to achieve the above object, the present invention provides the following technical solutions: An IoT-based intelligent data monitoring system for photovoltaic power plants, including IoT sensing and edge detection. The IoT sensing end is used to set an operation detection cycle, and then collect multiple operating data and environmental parameters of the photovoltaic components in each operation detection cycle through various sensors, and generate historical component operation records and real-time component operation records. At the same time, the operation of the photovoltaic components is monitored according to the predicted operation value range from the edge detection end; The edge detection end is used to extract a multidimensional data set from the historical component operation records, establish a plurality of isolation detection trees, input the multidimensional data set into the isolation detection trees, and then generate a plurality of normal value combinations and abnormal value combinations; A component operation analysis network is established based on normal value combinations and abnormal value combinations, and real-time component operation records are input into the component operation analysis network to generate predicted operation value intervals corresponding to several types of operation data of photovoltaic components during the next operation detection cycle, and abnormal predictions of the predicted operation value intervals are marked.

[0006] Furthermore, the process of collecting the operating data and environmental parameters includes: The IoT sensing terminal sets a number for each photovoltaic module in the photovoltaic power station, where the numbers are a1, a2, ..., a n , n is a natural number greater than 0; Install multiple sensors of the same model on each photovoltaic module, set an operation detection cycle for all sensors, and each time an operation detection cycle starts, each sensor collects multiple operating data and environmental data of the photovoltaic module where it is located; Whenever an operation detection cycle ends, the IoT perception end sets a number for the collected data of each sensor and integrates the data with the same number into a component operation record.

[0007] Furthermore, the process of establishing the isolation detection tree includes: From all historical component operation records of each photovoltaic component, one type of historical environmental data and other historical operation data are selected in turn to generate a multidimensional data set S, wherein the multidimensional data set S = [s1, s2, ..., s i , h j ],s i and h j Represent the i-th type of operating data and the j-th type of environmental data, respectively, where i and j are natural numbers greater than 10; Several isolation detection trees are established based on the total number of historical component operation records of photovoltaic modules. A combination of operation data and historical environmental data in a multidimensional dataset is randomly selected as the detection feature dimension combination of an isolation detection tree. A dynamic segmentation threshold combination and positive and negative deviation unit values ​​are set for the isolation detection trees with the same detection feature dimension combination. The size of the positive and negative deviation unit values ​​is ±1% of the initial value of the dynamic segmentation threshold combination. The isolation detection tree consists of a root node and several leaf nodes.

[0008] Furthermore, the initial value of the dynamic segmentation threshold combination of each isolation detection tree is the middle value of the corresponding type of environmental data and operation data in the historical component operation records, that is, the same type of historical operation data or historical environment data in the historical component operation records are sorted by size, and the middle value of the sorting results is selected as the initial value of the dynamic segmentation threshold combination.

[0009] Furthermore, the generation process of normal value combinations and abnormal value combinations includes: Each multidimensional dataset is sequentially input into each isolation detection tree, and then iterated through the isolation forest algorithm so that the root node of each isolation detection tree has only one multidimensional dataset, and the multidimensional datasets in each isolation detection tree are different; During the iterative operation, whenever a dynamic segmentation threshold combination is used to allow an isolation detection tree to pass through a multidimensional data set and reach the root node, the corresponding dynamic segmentation threshold combination is set to an unselectable value, and then the dynamic segmentation threshold combinations of other isolation detection trees are adjusted by positive and negative deviation unit values ​​until a multidimensional data set that has not reached the root node before reaches the root node; Obtain the anomaly threshold score of each isolation detection tree, set the anomaly determination threshold, and then determine the size relationship between the anomaly threshold score of each isolation detection tree and the anomaly determination threshold. If the anomaly threshold score is greater than or equal to the anomaly determination threshold, then determine that the corresponding dynamic segmentation threshold combination is an abnormal value combination; If the abnormal threshold score is less than the abnormal judgment threshold, the corresponding dynamic segmentation threshold combination is judged to be a normal value combination; The above process of generating normal value combinations and abnormal value combinations is repeated to sequentially obtain normal value combinations and abnormal value combinations between different types of environmental data and various operating data.

[0010] Furthermore, the process of establishing the component operation analysis network includes: Establish a multidimensional coordinate system, map the abnormal value combinations and normal data combinations under various environmental data into the same multidimensional coordinate system, and use the environmental data in each abnormal value combination or normal data combination as the head node, and the rest of the tail nodes at the same level as the operating data to generate a data connection tree corresponding to the abnormal value combination or normal data combination; According to the data values ​​corresponding to the head nodes of each data connection tree, each data connection tree is fitted and clustered, and a data connection network corresponding to the combination of environmental data and operation data is generated according to the fitted clusters; The fitting clustering is to distribute the same environmental data in a multidimensional coordinate system, connect the head nodes in sequence according to the numerical degree of the head nodes, and simultaneously connect the tail nodes of different data connection trees according to the head node connection process; Then, the data connection networks corresponding to the abnormal numerical combinations and normal data combinations under different environmental data are fitted and clustered to obtain the component operation analysis network of the corresponding photovoltaic components, and different center box selection radii are set for various operating data and environmental data in the component operation analysis network.

[0011] Furthermore, the process of establishing the predicted operating value interval for the next operating detection cycle includes: Whenever an operation detection cycle begins, the real-time component operation records of each photovoltaic component at the end of the previous operation detection cycle are input into the corresponding component operation analysis network; Then, the component operation analysis network extracts multiple real-time environmental data from the real-time component operation records. Taking each real-time environmental data as the center value, the component operation analysis network retrieves the environmental data within the center selection radius and records it as the detection environmental data. Then, the data connection tree is retrieved according to the detection environment data, and cluster fitting is performed on all data connection trees to generate a state detection network corresponding to the real-time environment data. The center radius of various operation data is selected, and the real-time operation data is used as the center value to segment the predicted operation value intervals of several types of operation data from the state detection network. If the predicted operation value interval contains an interval segment with an abnormal value combination, an abnormal prediction label is set for the interval segment in the time interval corresponding to the predicted operation value interval. Otherwise, no operation is performed. According to the numbers of the real-time component operation records, the predicted operation value ranges of various operation data are marked with numbers.

[0012] Furthermore, the process of monitoring the operation of the photovoltaic module according to the predicted operation value range includes: In the next operation detection cycle, the IoT sensing end retrieves the predicted operation value interval according to the number to monitor the operation of each photovoltaic module, and performs maintenance operations on the photovoltaic modules in the time interval corresponding to the interval segment with the abnormal prediction mark until the end of the corresponding operation detection cycle.

[0013] Compared with the prior art, the present invention has the following beneficial effects: 1. The present invention establishes an isolation detection tree and a component operation analysis network to conduct in-depth mining and analysis of historical component operation records. It can accurately identify normal and abnormal value combinations and generate a predicted operation value range for the next operation detection cycle, providing powerful decision-making support for the operation and maintenance of photovoltaic modules.

[0014] 2. By making abnormal prediction annotations on the predicted operating value interval, the present invention can detect potential faults of photovoltaic components in advance, take timely measures to repair and replace them, avoid further expansion of the fault, and reduce the operation and maintenance costs and downtime losses of the photovoltaic power station. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 This is a system structure diagram of the present invention. DETAILED DESCRIPTION

[0016] In order to further illustrate the technical means and effects adopted by the present invention to achieve the predetermined purpose of the invention, the specific implementation methods, structures, features and effects of the present invention are described in detail below in conjunction with the accompanying drawings and preferred embodiments.

[0017] like Figure 1As shown in the figure, the intelligent data monitoring system for photovoltaic power stations based on the Internet of Things includes an Internet of Things sensing end and an edge detection end; The IoT sensing end is used to set an operation detection cycle, and then collect multiple operating data and environmental parameters of the photovoltaic components in each operation detection cycle through various sensors, and generate historical component operation records and real-time component operation records. At the same time, the operation of the photovoltaic components is monitored according to the predicted operation value range from the edge detection end; The edge detection end is used to extract a multidimensional data set from the historical component operation records, establish a plurality of isolation detection trees, input the multidimensional data set into the isolation detection trees, and then generate a plurality of normal value combinations and abnormal value combinations; A component operation analysis network is established based on normal value combinations and abnormal value combinations, and real-time component operation records are input into the component operation analysis network to generate predicted operation value intervals corresponding to several types of operation data of photovoltaic components during the next operation detection cycle, and abnormal predictions of the predicted operation value intervals are marked.

[0018] Furthermore, the IoT sensing terminal sets a number for each photovoltaic component in the photovoltaic power station, wherein the numbers are a1, a2, ..., a n , n is a natural number greater than 0; Install multiple sensors of the same model on each photovoltaic module. These sensors include voltage sensors, current sensors, temperature sensors, and light intensity sensors. Connect all sensors to the IoT sensing end, which then divides the sensors on the photovoltaic modules into operating sensors and environmental sensors. An operation detection cycle is set for all sensors. Each time an operation detection cycle starts, each sensor collects multiple operating data and environmental data of the PV module where it is located; The operating data includes voltage operating values, current operating values, etc., and the environmental data includes light intensity change values, temperature change values, etc. Whenever an operation detection cycle ends, the IoT perception end sets a number for the collected data of each sensor and integrates the data with the same number into a component operation record.

[0019] Furthermore, the IoT sensing end synchronizes the component operation records of each operation detection cycle with the edge detection end; The edge detection end is provided with an anomaly detection module and an operation prediction module; The anomaly detection module is used to extract a multidimensional dataset from historical component operation records and establish several isolation detection trees. The multidimensional dataset is input into the isolation detection trees to generate several normal value combinations and abnormal value combinations. Then, a component operation analysis network is established based on the normal value combinations and abnormal value combinations. The specific process includes: From all historical component operation records of each photovoltaic component, one type of historical environmental data and other historical operation data are selected in turn to generate a multidimensional data set S, wherein the multidimensional data set S = [s1, s2, ..., s i , h j ],s i and h j Represent the i-th type of operating data and the j-th type of environmental data, respectively, where i and j are natural numbers greater than 10; Several isolation detection trees are established based on the total number of historical component operation records of photovoltaic modules. A combination of operation data and historical environmental data in a multidimensional dataset is randomly selected as the detection feature dimension combination of an isolation detection tree. A dynamic segmentation threshold combination and positive and negative deviation unit values ​​are set for the isolation detection trees with the same detection feature dimension combination. The size of the positive and negative deviation unit values ​​is ±1% of the initial value of the dynamic segmentation threshold combination. The isolation detection tree consists of a root node and several leaf nodes; It should be noted that the initial value of the dynamic segmentation threshold combination of each isolation detection tree is the median value of the corresponding type of environmental data and operation data in the historical component operation record. That is, the same type of historical operation data or historical environmental data in the historical component operation record is sorted by size, and the median value of the sorting result is selected as the initial value of the dynamic segmentation threshold combination; Each multidimensional dataset is sequentially input into each isolation detection tree, and then iterated through the isolation forest algorithm so that the root node of each isolation detection tree has only one multidimensional dataset, and the multidimensional datasets in each isolation detection tree are different; During the iterative operation, whenever a dynamic segmentation threshold combination is used to allow an isolation detection tree to pass through a multidimensional data set and reach the root node, the corresponding dynamic segmentation threshold combination is set to an unselectable value, and then the dynamic segmentation threshold combinations of other isolation detection trees are adjusted by positive and negative deviation unit values ​​until a multidimensional data set that has not reached the root node before reaches the root node; Obtain the abnormal threshold score of each isolation detection tree. The calculation formula of the abnormal threshold score is: ; in represents the abnormal threshold score of the kth isolation detection tree, is the correction parameter, K is the total number of historical component operation records, represents the average path length in the kth isolation detection tree, represents the path length of the x-th multidimensional dataset in the k-th isolation detection tree; Set an abnormality determination threshold, and then determine the relationship between the abnormality threshold score of each isolation detection tree and the abnormality determination threshold. If the abnormality threshold score is greater than or equal to the abnormality determination threshold, the corresponding dynamic segmentation threshold combination is determined to be an abnormal value combination; If the abnormal threshold score is less than the abnormal judgment threshold, the corresponding dynamic segmentation threshold combination is judged to be a normal value combination; The above process of generating normal value combinations and abnormal value combinations is repeated to sequentially obtain normal value combinations and abnormal value combinations between different types of environmental data and various operating data.

[0020] Furthermore, a multidimensional coordinate system is established to map abnormal value combinations and normal data combinations under various environmental data into the same multidimensional coordinate system, and a data connection tree corresponding to the abnormal value combination or normal data combination is generated with the environmental data in each abnormal value combination or normal data combination as the head node and the rest of the tail nodes at the same level as the operating data; According to the data values ​​corresponding to the head nodes of each data connection tree, each data connection tree is fitted and clustered, and a data connection network corresponding to the combination of environmental data and operation data is generated according to the fitted clusters; The fitting clustering is to distribute the same environmental data in a multidimensional coordinate system, connect the head nodes in sequence according to the numerical degree of the head nodes, and simultaneously connect the tail nodes of different data connection trees according to the head node connection process; Then, the data connection network corresponding to the abnormal value combination and normal data combination under different environmental data is fitted and clustered to obtain the component operation analysis network of the corresponding photovoltaic module, and different center box selection radii are set for various operating data and environmental data in the component operation analysis network; It should be noted that, whenever an operation detection cycle ends, the anomaly detection module updates the component operation analysis network based on the component operation records from the IoT perception end.

[0021] Furthermore, the operation prediction module is used to input the real-time component operation records into the component operation analysis network, thereby generating predicted operation value intervals corresponding to several types of operation data of the photovoltaic components during the next operation detection cycle, and annotating abnormal predictions of the predicted operation value intervals. The specific process includes: The anomaly detection module synchronizes the component operation analysis network with the operation prediction module, and then each time an operation detection cycle begins, the operation prediction module inputs the real-time component operation record of each photovoltaic component at the end of the previous operation detection cycle into the corresponding component operation analysis network; Then, the component operation analysis network extracts multiple real-time environmental data from the real-time component operation records. Taking each real-time environmental data as the center value, the component operation analysis network retrieves the environmental data within the center selection radius and records it as the detection environmental data. Then, the data connection tree is retrieved according to the detection environment data, and cluster fitting is performed on all data connection trees to generate a state detection network corresponding to the real-time environment data. The center radius of various operation data is selected, and the real-time operation data is used as the center value to segment the predicted operation value intervals of several types of operation data from the state detection network. If the predicted operation value interval contains an interval segment with an abnormal value combination, an abnormal prediction label is set for the interval segment in the time interval corresponding to the predicted operation value interval. Otherwise, no operation is performed. According to the numbers in the real-time component operation records, the predicted operation value intervals of various operation data are marked with numbers and sent to the IoT perception end; Then, in the next operation detection cycle, the IoT sensing end retrieves the predicted operation value interval according to the number to monitor the operation of each photovoltaic module, and performs maintenance operations on the photovoltaic modules in the time interval corresponding to the interval segment with the abnormal prediction mark until the end of the corresponding operation detection cycle.

[0022] The above description is merely a preferred embodiment of the present invention and does not constitute any form of limitation to the present invention. Although the present invention has been disclosed as a preferred embodiment as above, it is not intended to limit the present invention. Any person skilled in the art can make some changes or modifications to equivalent embodiments using the technical contents disclosed above without departing from the scope of the technical solution of the present invention. However, any indirect modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention are still within the scope of the technical solution of the present invention.

Claims

1. The photovoltaic power station intelligent data monitoring system based on the Internet of Things is characterized by: Including IoT perception end and edge detection end; The IoT sensing end is used to set an operation detection cycle, collect multiple operating data and environmental parameters of the photovoltaic components in each operation detection cycle through various sensors, and generate historical component operation records and real-time component operation records. At the same time, the operation of the photovoltaic components is monitored according to the predicted operation value range from the edge detection end; The edge detection end is used to extract a multidimensional data set from the historical component operation records, establish a plurality of isolation detection trees, input the multidimensional data set into the isolation detection trees, and then generate a plurality of normal value combinations and abnormal value combinations; A component operation analysis network is established based on normal value combinations and abnormal value combinations, and real-time component operation records are input into the component operation analysis network to generate predicted operation value intervals corresponding to several types of operation data of photovoltaic components during the next operation detection cycle, and abnormal predictions of the predicted operation value intervals are marked.

2. The photovoltaic power station intelligent data monitoring system based on the Internet of Things according to claim 1 is characterized in that: The process of collecting the operating data and environmental parameters includes: Each photovoltaic module in the photovoltaic power station is numbered, where the numbers are a1, a2, ..., a n , n is a natural number greater than 0; Multiple sensors of the same model are installed on each photovoltaic module, and an operation detection cycle is set for all sensors. Whenever an operation detection cycle begins, the sensor collects multiple operation data and environmental data of the photovoltaic module. The Internet of Things perception end sets a number for the collected data of each sensor and integrates the data with the same number into a module operation record.

3. The photovoltaic power station intelligent data monitoring system based on the Internet of Things according to claim 2 is characterized in that: The process of establishing the isolation detection tree includes: From all historical component operation records of each photovoltaic component, select one type of historical environmental data and other historical operation data in turn to generate a multidimensional data set; Establish several isolation detection trees based on the total number of historical component operation records of photovoltaic modules, and randomly select a combination of operation data and historical environmental data in the multidimensional dataset as the detection feature dimension combination of an isolation detection tree. Set a dynamic segmentation threshold combination and positive and negative deviation unit values ​​for the isolation detection trees with the same detection feature dimension combination; The isolation detection tree consists of a root node and several leaf nodes.

4. The photovoltaic power station intelligent data monitoring system based on the Internet of Things according to claim 3 is characterized in that: The same historical operation data or historical environment data in the historical component operation records are sorted by size, and the middle value of the sorting result is selected as the initial value of the dynamic segmentation threshold combination.

5. The photovoltaic power station intelligent data monitoring system based on the Internet of Things according to claim 3 is characterized in that: The process of generating normal value combinations and abnormal value combinations includes: The multidimensional dataset is sequentially input into each isolation detection tree, and the isolation forest algorithm is used for iterative operation, so that the root node of each isolation detection tree has only one multidimensional dataset, and the multidimensional datasets in each isolation detection tree are different; During the iterative operation, whenever a dynamic segmentation threshold combination is used to allow an isolation detection tree to pass through a multidimensional data set and reach the root node, the corresponding dynamic segmentation threshold combination is set to an unselectable value, and then the dynamic segmentation threshold combinations of other isolation detection trees are adjusted by positive and negative deviation unit values ​​until a multidimensional data set that has not reached the root node before reaches the root node; Obtain the anomaly threshold score of each isolation detection tree, set the anomaly judgment threshold, and judge the size relationship between the anomaly threshold score of each isolation detection tree and the anomaly judgment threshold. If the anomaly threshold score is greater than or equal to the anomaly judgment threshold, then the corresponding dynamic segmentation threshold combination is judged to be an abnormal numerical combination. If the anomaly threshold score is less than the anomaly judgment threshold, then the corresponding dynamic segmentation threshold combination is judged to be a normal numerical combination.

6. The photovoltaic power station intelligent data monitoring system based on the Internet of Things according to claim 5 is characterized in that: The process of establishing the component operation analysis network includes: Establish a multidimensional coordinate system, map the abnormal value combinations and normal data combinations under various environmental data into the same multidimensional coordinate system, and use the environmental data in each abnormal value combination or normal data combination as the head node, and the rest of the tail nodes at the same level as the operating data to generate a data connection tree corresponding to the abnormal value combination or normal data combination; According to the data values ​​corresponding to the head nodes of each data connection tree, each data connection tree is fitted and clustered, and a data connection network corresponding to the combination of environmental data and operation data is generated according to the fitted clusters; Then, the data connection networks corresponding to the abnormal numerical combinations and normal data combinations under different environmental data are fitted and clustered to obtain the component operation analysis network of the corresponding photovoltaic components, and different center box selection radii are set for various operating data and environmental data in the component operation analysis network.

7. The photovoltaic power station intelligent data monitoring system based on the Internet of Things according to claim 6 is characterized in that: The process of establishing the predicted operating value range for the next operating detection cycle includes: Whenever an operation detection cycle begins, the real-time component operation records of each photovoltaic component at the end of the previous operation detection cycle are input into the corresponding component operation analysis network; The component operation analysis network extracts multiple real-time environmental data from the real-time component operation records. Taking each real-time environmental data as the center value, the environmental data within the center selection radius is retrieved from the component operation analysis network and recorded as the detection environment data. Retrieve the data connection tree based on the detection environment data, perform cluster fitting on all data connection trees to generate a state detection network corresponding to the real-time environment data, and select the radius based on the center of various operation data. With the real-time operation data as the center value, segment the predicted operation value intervals of several types of operation data from the state detection network. If the predicted operation value interval contains interval segments with abnormal value combinations, set an abnormal prediction label on the interval segment corresponding to the time interval of the predicted operation value interval. Otherwise, do nothing. According to the numbers of the real-time component operation records, the predicted operation value ranges of various operation data are marked with numbers.

8. The photovoltaic power station intelligent data monitoring system based on the Internet of Things according to claim 7 is characterized in that: The process of monitoring the operation of PV modules based on the predicted operating value range includes: In the next operation detection cycle, the IoT sensing end retrieves the predicted operation value interval according to the number to monitor the operation of each photovoltaic module, and performs maintenance operations on the photovoltaic modules in the time interval corresponding to the interval segment with the abnormal prediction mark until the end of the corresponding operation detection cycle.

Images