Distributed photovoltaic operation fault risk scene intelligent monitoring and optimizing platform

Abstract

The invention relates to the field of photovoltaic power generation, in particular to a distributed photovoltaic operation fault risk scene intelligent monitoring and optimizing platform. The monitoring module monitors related information of photovoltaic facilities in real time and transmits the related information to the feature classification module; the feature classification module classifies the information collected by the photovoltaic monitoring module and transmits the classified data to the fault risk analysis module; the fault risk analysis module performs fault risk identification according to the information transmitted by the feature classification module, and sends an identification result to the optimization platform; the optimization platform generates a user operable interface, formulates a risk value map, processes and analyzes faults, sends early warning information, and visualizes monitoring module data and fault risk identification results. According to the method, the data is analyzed through the clustering algorithm, the fault risk existing in the photovoltaic facility is effectively identified, the method is more accurate and efficient, potential faults can be early warned in advance, and losses caused by the faults are avoided or reduced.

Description

Technical Field

[0001] The present invention relates to the field of photovoltaic power generation, and in particular to an intelligent monitoring and optimization platform for distributed photovoltaic operation failure risk scenarios. Background Art

[0002] With the growing global demand for clean energy, distributed photovoltaics have been widely used and developed. Distributed photovoltaic systems are usually installed in various scenarios, such as rooftops, industrial parks, and rural areas. In order to ensure the efficient and stable operation of these photovoltaic systems, real-time monitoring and fault diagnosis are essential. At the same time, with the development of intelligent technology, it is possible to use advanced sensors, data analysis, and artificial intelligence algorithms to achieve intelligent monitoring and optimization. Through the intelligent monitoring platform, fault risks can be discovered in a timely manner, the reliability and power generation efficiency of the system can be improved, and the sustainable development of the distributed photovoltaic industry can be promoted.

[0003] The prior art still has the following defects:

[0004] 1. Insufficient data fusion: Different types of data, such as operating parameters and environmental data, may not be well integrated and correlated, affecting the overall judgment of the fault.

[0005] 2. Weak adaptability to scenarios and inaccurate algorithm recognition: It is difficult to adapt to complex and changeable distributed photovoltaic installation scenarios, such as different climatic conditions and geographical locations. In terms of fault risk identification, it mainly relies on preset rules and thresholds for judgment, lacks flexibility and adaptability, and is difficult to cope with complex and changeable operating environments and fault types.

[0006] 3. Lack of effective optimization and intuitive reflection: The speed of identifying faults is slow. After the fault risk is discovered, the optimization suggestions and measures provided may not be perfect. The relevant staff cannot immediately understand where the problem occurred in the photovoltaic facilities and what type of problem it is, which affects the operation of the photovoltaic facilities and causes losses. Summary of the invention

[0007] In order to solve the above technical problems, the present invention provides a distributed photovoltaic operation fault risk scenario intelligent monitoring and optimization platform to solve the problems in the existing background technology.

[0008] The technical solution of the present invention is as follows:

[0009] An intelligent monitoring and optimization platform for distributed photovoltaic operation fault risk scenarios, including:

[0010] A monitoring module, which monitors photovoltaic facility-related information in real time, transmits environmental information to the optimization platform, and transmits operating status parameter information to the feature classification module;

[0011] A feature classification module, which classifies and adds feature identifiers to the information collected by the photovoltaic monitoring module, and transmits the classified data to the fault risk analysis module;

[0012] A fault risk analysis module, which identifies fault risks based on the information transmitted by the feature classification module and sends the identification results to the optimization platform;

[0013] Optimization platform; the optimization platform generates a user-operable interface, formulates risk value maps, fault handling analysis and sends early warning information based on the environment transmitted by the monitoring module and the identification results of the fault risk analysis module, and visualizes the monitoring module data and fault risk identification results.

[0014] A preferred technical solution is that the monitoring module uses sensors and monitoring to monitor the status data of photovoltaic facilities in real time, including environmental information and operating status information;

[0015] The environmental information includes geographical location, ambient temperature, climate information, light intensity change information and shadow area change information;

[0016] The operating status information includes photovoltaic facility temperature, voltage, current, power, inverter DC input power and photovoltaic facility appearance component image data.

[0017] A preferred technical solution is that the feature classification module adds features to the collected data to facilitate subsequent identification and enhance the applicability of the system. The feature classification method is as follows:

[0018] The feature classification module classifies the collected data according to the time sequence and classifies the feature values ​​of the data; first, the data set H is divided according to each temperature range;

[0019] In each set H, each piece of light intensity data and the photovoltaic area data exposed to light are combined to obtain the photovoltaic power generation data collected in the combination, and they are divided into a small set G. In each set G:

[0020] The current is marked as feature 1 A, the voltage is marked as feature 2 B, the generated power is marked as feature 2 C, and the inverter DC input power is marked as feature 3 D;

[0021] Appearance component image data changes are divided into changes and no changes. Changes are directly recorded as failure risks.

[0022] The set G is represented as G(A, B, C, D);

[0023] A preferred technical solution, the fault analysis module uses a clustering algorithm (KMeans) to identify fault risks according to feature identification;

[0024] Take each data set G after adding identification and classification in the feature adding module as a data point;

[0025] Randomly generate multiple clusters and cluster centers;

[0026] Obtain a data set of the normal working state of photovoltaic facilities from enterprises related to the photovoltaic facilities or a data set of the normal power generation state of photovoltaic facilities in historical records, and use the set as a data point as the center of one of the clusters;

[0027] G K (A K ,B K ,C K ,D K ) represents the data point at the center point, and K represents the center of the cluster;

[0028] For each data point, calculate its distance d to the center of each cluster using the Euclidean distance formula:

[0029]

[0030] Assign data points to the closest cluster according to the distance d, and assign data points with similar characteristics to the same cluster;

[0031] The data points are allocated to multiple clusters according to the distance between the data points. If the distances are the same, all the data points are allocated to the cluster centered on the data point of the normal power generation status of the photovoltaic facility.

[0032] Cluster center update: except for the cluster center of the data point of the normal power generation state of the photovoltaic facility, the average value of the characteristic value of each data point in the other cluster is calculated as the new cluster center; the calculation method is:

[0033]

[0034] G′ K is the new cluster center; i is the number of data points in the cluster;

[0035] Similarly, calculate the new cluster center values ​​of other features until the cluster center no longer changes or reaches the set number of iterations;

[0036] Finally, it is assigned to a cluster different from the cluster centered on the normal working status data, and the PV facility at this data point is considered to be at risk of failure.

[0037] A preferred technical solution, the optimization platform module includes a storage unit, an operation scenario optimization unit, a fault identification optimization unit, a decision optimization unit, a user interface unit, and an early warning unit;

[0038] The storage unit is used to receive and store monitoring data from the monitoring module, fault risk records identified by the fault analysis module, historical fault data of photovoltaic facilities and photovoltaic equipment fault type processing;

[0039] The operation scenario optimization unit is used to calculate the risk value of the geographical location;

[0040] The fault identification optimization unit is used to optimize the analysis of fault risks;

[0041] The decision optimization unit is used to analyze the failure risk response strategy;

[0042] The user interface unit is used to generate a user operation interface, visualize system content, and receive relevant photovoltaic data;

[0043] The early warning unit is used to issue an early warning according to the recognition result.

[0044] A preferred technical solution is that the operation scenario optimization unit calculates the risk value of the geographical location according to the geographical location information of the monitoring module and the identification result of the fault risk analysis module, and generates a distributed photovoltaic risk value map;

[0045] Each distributed photovoltaic is marked according to the identification results, with "1" indicating risk and "0" indicating no risk;

[0046] Count the number of distributed photovoltaics marked with "1" at each geographical location;

[0047] Select geographical locations with more “0” marks for allocation, or prioritize allocating distributed photovoltaics marked with “0” to these locations;

[0048] There are m distributed photovoltaics, with x m represents the risk mark of the mth distributed photovoltaic, x m is 0 or 1. For a geographical location L, its risk value Z L It can be obtained by summing the risk markers of all distributed photovoltaics at this location:

[0049]

[0050] The risk value of the geographical location L is calculated.

[0051] A preferred technical solution, the fault identification optimization unit is used to collect the operating status data of the photovoltaic facility when the fault is triggered, divide the set of relevant data records of the photovoltaic facility with confirmed fault into new data points, and transmit them to the fault risk identification module. The fault risk identification module uses the new data point as the center of the cluster for further accurate identification.

[0052] A preferred technical solution, the decision optimization unit calculates the probability of failure and maintenance means by statistical analysis method;

[0053] F represents the set of fault types, G represents the set of data when the fault occurs, and T represents the set of processing methods;

[0054] For a certain fault type f, its occurrence frequency can be expressed as:

[0055]

[0056] The degree of correlation between the cause of the fault G∈F and the fault type f can be expressed by conditional probability:

[0057]

[0058] The corresponding relationship between the processing method T∈F and the fault type f can be expressed as:

[0059]

[0060] Based on the photovoltaic data of the fault risk identified by the fault risk analysis module, the fault type with the highest probability and the treatment method with the highest probability under this type are calculated as the output of the maintenance suggestion.

[0061] A preferred technical solution is that the user interface unit generates an operation interface and visualizes the system content for relevant personnel to input and view relevant information;

[0062] The input information includes photovoltaic test data of photovoltaic manufacturers, photovoltaic historical operation data, contact information of relevant staff, photovoltaic fault type information and photovoltaic fault handling means information.

[0063] The viewed information includes the photovoltaic facility status data and fault risk identification results monitored in real time by the monitoring module, and a map is made and displayed according to the geographical location risk value information calculated by the operation scenario optimization unit;

[0064] A preferred technical solution is that the early warning unit issues an early warning when the fault analysis module identifies a fault risk. The early warning means includes sending the fault risk information to relevant personnel of the photovoltaic facility using text messages according to the contact information input through the operation interface. The content of the text message will describe in detail the fault risk identification results, location and monitoring image information.

[0065] Compared with the prior art, the present invention has the following beneficial effects:

[0066] 1. The present invention can effectively identify the failure risks of photovoltaic facilities by monitoring the working status data of photovoltaic facilities in real time and analyzing the data using a clustering algorithm. Compared with traditional fault detection methods, the method of the present invention is more accurate and efficient, and can warn of potential faults in advance, thereby avoiding or reducing losses caused by faults.

[0067] 2. The present invention realizes the comprehensive management of monitoring data, fault risk records, historical fault data and fault handling types by optimizing the design of platform modules. The operation scenario optimization unit can provide a scientific basis for the allocation of distributed photovoltaics based on geographic location information and fault risk analysis results, thereby improving the overall operation efficiency of photovoltaic facilities. At the same time, the fault identification optimization unit further improves the accuracy of fault identification by collecting data when the fault is triggered.

[0068] 3. The present invention, through the design of the user interface unit and the early warning unit, enables relevant personnel to easily input and view relevant information, and to issue timely early warnings when fault risks are identified, thereby improving the timeliness and effectiveness of fault handling. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] Figure 1 It is a flow chart of the present invention.

[0070] Figure 2 It is a flow chart of the fault risk analysis module of the present invention.

[0071] Figure 3 It is the optimization platform flow chart. DETAILED DESCRIPTION

[0072] The following embodiments of the present invention are described in further detail in conjunction with the accompanying drawings and examples. The following examples are used to illustrate the present invention, but are not intended to limit the scope of the present invention.

[0073] like Figure 1-3 As shown, the present invention provides a distributed photovoltaic operation fault risk scenario intelligent monitoring and optimization platform, including:

[0074] A monitoring module, which monitors photovoltaic facility-related information in real time, transmits environmental information to the optimization platform, and transmits operating status parameter information to the feature classification module;

[0075] A feature classification module, which classifies and adds feature identifiers to the information collected by the photovoltaic monitoring module, and transmits the classified data to the fault risk analysis module;

[0076] A fault risk analysis module, which identifies fault risks based on the information transmitted by the feature classification module and sends the identification results to the optimization platform;

[0077] Optimization platform; the optimization platform generates a user-operable interface, formulates risk value maps, fault handling analysis and sends early warning information based on the environment transmitted by the monitoring module and the identification results of the fault risk analysis module, and visualizes the monitoring module data and fault risk identification results.

[0078] Embodiment 1: Figure 1-3 As shown, in this embodiment, the monitoring module uses sensors and monitoring to monitor the status data of photovoltaic facilities in real time, including environmental information and operating status information;

[0079] Environmental information includes geographic location, ambient temperature, climate information, light intensity change information, and shadow area change information;

[0080] The operating status information includes the photovoltaic facility temperature, voltage, current, power, inverter DC input power and photovoltaic facility appearance component image data.

[0081] The feature classification module adds features to the collected data to facilitate subsequent identification and enhance the applicability of the system. The feature classification method is as follows:

[0082] First, divide the data set H according to each temperature range. Then, in each set H, combine each piece of light intensity data with the photovoltaic area data exposed to light to obtain the photovoltaic power generation data collected in the combination. Divide them into a small set G, each set G contains the current mark, the voltage mark as feature 2 B, the power generation mark as feature 2 C, the inverter DC input power mark as feature 3 D, and the appearance component image data changes are divided into changes and no changes.

[0083] The fault analysis module uses a clustering algorithm (such as KMeans) to identify fault risks based on feature identification. Each data set G after adding identification and classification in the module is used as a data point. Multiple clusters and cluster centers are randomly generated, and the normal power generation status data set of photovoltaic facilities in the normal working status data set history record of photovoltaic facilities is used as the center of a cluster.

[0084] The distance between the data points and the centers of each cluster is calculated according to the Euclidean distance formula, and data points with similar characteristics are placed in the same cluster. If the data points are at the same distance from multiple cluster centers, all of them are assigned to the cluster centered on the data point of the normal power generation status of the photovoltaic facility. The cluster center is updated, and the average value of the characteristic values ​​of the data points in each cluster is calculated as the new cluster center. Finally, data points assigned to clusters different from the cluster centered on the normal working status data are considered to have a risk of failure.

[0085] Here is a specific example: We have collected data from distributed photovoltaic power generation from the same manufacturer, and the data sets for five operating scenarios are as follows:

[0086] Suppose there are 5 sets of data as follows:

[0087] Data group 1: ambient temperature 25°C, light intensity 800 lux, photovoltaic area 10 square meters, current 5A, voltage 20V, power generation power 100W, inverter DC input power 80W, and appearance component image data unchanged.

[0088] Data group 2: ambient temperature 20°C, light intensity 600 lux, photovoltaic area 8 square meters, current 4A, voltage 18V, power generation power 72W, inverter DC input power 60W, and the appearance component image data remained unchanged.

[0089] Data group 3: ambient temperature 30°C, light intensity 1000 lux, photovoltaic area 12 square meters, current 6A, voltage 22V, power generation power 132W, inverter DC input power 100W, and the appearance component image data remained unchanged.

[0090] Data group 4: ambient temperature 22°C, light intensity 700 lux, photovoltaic area 9 square meters, current 4.5A, voltage 19V, power generation power 85.5W, inverter DC input power 70W, and appearance component image data remained unchanged.

[0091] Data group 5: ambient temperature 28°C, light intensity 900 lux, photovoltaic area 11 square meters, current 5.5A, voltage 21V, power generation 115.5W, inverter DC input power 90W, the appearance component image data has changed, which is recorded as a failure risk.

[0092] First, the data set H is divided according to the temperature range. Assuming that the temperature range is 20-30℃, these 5 groups of data are all in this set H.

[0093] Then, in set H, for each group of data, combine them according to light intensity and photovoltaic area to get a small set G:

[0094] Set G1: (5A, 20V, 100W, 80W)

[0095] Set G2: (4A, 18V, 72W, 60W)

[0096] Set G3: (6A, 22V, 132W, 100W)

[0097] G4: (4.5A, 19V, 85.5W, 70W)

[0098] G5: (5.5A, 21V, 115.5W, 90W)

[0099] Two clusters and cluster centers are randomly generated. It is assumed that cluster center 1 is (4A, 18V, 70W, 60W) and cluster center 2 is the normal working status data obtained from the enterprise (5A, 20V, 100W, 80W).

[0100] First iteration:

[0101] Calculate the distance d from each data point to the cluster center:

[0102] d(number 1-cluster 1)=√[(5-4) 2 +(20-18) 2 +(100-70) 2 +(80-60) 2 ]=√(1+4+900+400)=√1305≈36.1.

[0103] d(number 2-cluster 1)=√[(4-4) 2 +(18-18) 2 +(72-70) 2 +(60-60) 2 ]=√4=2.

[0104] d(number 3-cluster 1)=√[(6-4) 2 +(22-18) 2 +(132-70) 2 +(100-60) 2 ]=√(4+16+3844+1600)=√5464≈73.9.

[0105] d(number 4-cluster 1)=√[(4.5-4) 2 +(19-18) 2 +(85.5-70) 2 +(70-60) 2 ]=√(0.25+1+20.25+100)=√121.5≈11.0.

[0106] d(number 5-cluster 1)=0 (because it has been marked as a failure risk, it is directly assigned to the failure cluster).

[0107] The following is to calculate the distance d from each data point to the cluster center 2 (normal working state data (5A, 20V, 100W, 80W)):

[0108] d(number 1-cluster 2) = √[(5-5) 2 +(20-20) 2 +(100-100) 2 +(80-80) 2 ]=0.

[0109] d(number2-cluster2)=√[(4-5) 2 +(18-20) 2 +(72-100) 2 +(60-80) 2 ]=√(1+4+784+400)=√1189≈34.5.

[0110] d(number 3 - cluster 2) = √[(6-5) 2 +(22-20) 2 +(132-100) 2 +(100-80) 2 ]=√(1+4+961+400)=√1366≈37.0.

[0111] d(number 4-cluster 2)=√[(4.5-5) 2 +(19-20) 2 +(85.5-100) 2 +(70-80) 2 ]=√(0.25+1+20.25+100)=√121.5≈11.0.

[0112] d(number 5-cluster 2)=0 (because it has been marked as a failure risk, it is directly assigned to the failure cluster).

[0113] Data group 1 is assigned to the cluster where cluster center 1 is located, data group 2 is assigned to the cluster where cluster center 2 is located, data group 3 is assigned to the cluster where cluster center 1 is located, data group 4 is assigned to the cluster where cluster center 2 is located, and data group 5 is assigned to the fault cluster.

[0114] Second iteration:

[0115] The new cluster center 1 becomes (5.5A, 21V, 115.5W, 90W), and cluster center 2 remains unchanged.

[0116] The distances and assignments continue to be calculated, with the same results as in the first iteration.

[0117] After multiple iterations, the results remained stable, and finally data groups 1 and 3 were assigned to the cluster where cluster center 1 was located, which was considered to be at risk of failure, and data groups 2 and 4 were assigned to the cluster where cluster center 2 was located, which was considered normal.

[0118] The reasons for multiple iterations are mainly the following:

[0119] Make the allocation of data points more accurate. As iterations proceed, data points may be readjusted according to more precise cluster centers, improving the accuracy of fault identification.

[0120] Adapting to the complexity and diversity of data, actual data may have various fluctuations and changes, and multiple iterations can better handle these situations to achieve more reliable results.

[0121] By monitoring the working status data of photovoltaic facilities in real time and analyzing the data using clustering algorithms, the failure risks of photovoltaic facilities can be effectively identified. Compared with traditional fault detection methods, the method of the present invention is more accurate and efficient, and can warn of potential faults in advance, thereby avoiding or reducing losses caused by faults.

[0122] Embodiment 2: According to Figure 1-3 As shown: The operation scenario optimization unit calculates the risk value of the geographical location based on the geographical location information of the monitoring module and the identification result of the fault risk analysis module, and generates a distributed photovoltaic risk value map.

[0123] Each distributed photovoltaic is marked according to the identification results, with "1" indicating risk and "0" indicating no risk;

[0124] Count the number of distributed photovoltaics marked with "1" at each geographical location;

[0125] Select geographical locations with more “0” marks for allocation, or prioritize allocating distributed photovoltaics marked with “0” to these locations.

[0126] There are 5 distributed photovoltaics, and their risk markers are:

[0127] Distributed photovoltaic 1: marked as 1.

[0128] Distributed photovoltaic 2: marked as 0.

[0129] Distributed photovoltaic 3: marked as 0.

[0130] Distributed photovoltaic 4: marked as 1.

[0131] Distributed photovoltaic 5: marked as 0.

[0132] For geographical location L, there are distributed photovoltaics 1, 2, and 3 at this location.

[0133] Then the risk value of geographical location L = 1+0+0=1.

[0134] The fault identification optimization unit collects the operating status data of the PV facility when a fault is triggered, divides it into new data points and transmits it to the fault risk identification module. The new data point becomes the new cluster center.

[0135] Now, a new photovoltaic facility operating status data (3A, 15V, 80W, 60W) is collected when a fault is triggered. It is divided into new data points and used as a new cluster center.

[0136] In subsequent fault identification, the data of this new cluster center will be used as a reference to more accurately identify similar fault conditions. For example, when the operating status of a distributed photovoltaic is monitored to be close to (3A, 15V, 80W, 60W), it can be judged more quickly that there is a fault risk.

[0137] Decision optimization unit, assuming that the fault type set F = {f1, f2}, the data set when the fault occurs G = {g1, g2, g3}, and the processing means set T = {t1, t2}.

[0138] For fault type f1, the frequency of occurrence P(f1)=0.6, the correlation degree between fault cause g1 and fault type f1 P(G|f1)=0.8, and the corresponding relationship between processing method t1 and fault type f1 P(T|f1)=0.7.

[0139] Based on the photovoltaic data of the fault risk identified by the fault risk analysis module, it is calculated that the fault type with the highest probability is f1, and the treatment method with the highest probability under this type is t1, which is output as the maintenance suggestion.

[0140] By optimizing the design of the platform, comprehensive management of monitoring data, fault risk records, historical fault data and fault handling types is achieved. Among them, the operation scenario optimization unit can provide a scientific basis for the allocation of distributed photovoltaics based on geographic location information and fault risk analysis results, thereby improving the overall operation efficiency of photovoltaic facilities. At the same time, the fault identification optimization unit further improves the accuracy of fault identification by collecting data when the fault is triggered.

[0141] The user interface unit generates an operation interface. The input information includes photovoltaic test data and photovoltaic historical operation data of a photovoltaic manufacturer. The viewing information includes photovoltaic facility status data monitored in real time by the monitoring module, the operating status of each distributed photovoltaic and fault risk identification results, and a map display made according to the risk value information of the geographical location.

[0142] When the early warning unit identifies a fault, it will send the fault risk information to relevant personnel via SMS according to the input contact information. The SMS content describes in detail the fault risk identification results, location and other information.

[0143] Through the design of the user interface unit and the early warning unit, relevant personnel can easily input and view relevant information, and issue timely warnings when fault risks are identified, thereby improving the timeliness and effectiveness of fault handling.

[0144] The embodiments of the present invention are provided for the purpose of illustration and description. Although the embodiments of the present invention have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limitations of the present invention. Ordinary technicians in this field can change, modify, replace and modify the above embodiments within the scope of the present invention.

Claims

1. An intelligent monitoring and optimization platform for distributed photovoltaic operation failure risk scenarios, characterized by: include: A monitoring module, which monitors photovoltaic facility-related information in real time, transmits environmental information to the optimization platform, and transmits operating status parameter information to the feature classification module; A feature classification module, which classifies the information collected by the photovoltaic monitoring module and transmits the classified data to the fault risk analysis module; A fault risk analysis module, which identifies fault risks based on the information transmitted by the feature classification module and sends the identification results to the optimization platform; Optimize the platform; The optimization platform generates a user-operable interface, formulates risk value maps, fault handling analysis and sends warning information according to the environment transmitted by the monitoring module and the identification results of the fault risk analysis module, and visualizes the monitoring module data and fault risk identification results.

2. A distributed photovoltaic operation fault risk scenario intelligent monitoring and optimization platform as claimed in claim 1, characterized in that: The monitoring module uses sensors and monitors real-time status data of photovoltaic facilities, including environmental information and operating status information; The environmental information includes geographical location, ambient temperature, climate information, light intensity change information and shadow area change information; The operating status information includes photovoltaic facility temperature, voltage, current, power, inverter DC input power and photovoltaic facility appearance component image data.

3. A distributed photovoltaic operation fault risk scenario intelligent monitoring and optimization platform as claimed in claim 1, characterized in that: The feature classification module adds features to the collected data, and the classification method is as follows: The feature classification module classifies the collected data according to the time sequence and classifies the feature values ​​of the data; first, the data set H is divided according to each temperature range; In each set H, each piece of light intensity data and the photovoltaic area data exposed to light are combined to obtain the photovoltaic power generation data collected in the combination, and they are divided into a small set G. In each set G: The current is marked as feature 1 A, the voltage is marked as feature 2 B, the generated power is marked as feature 2 C, and the inverter DC input power is marked as feature 3 D; Appearance component image data changes are divided into changes and no changes. Changes are directly recorded as failure risks. The set G is represented by G(A, B, C, D).

4. A distributed photovoltaic operation fault risk scenario intelligent monitoring and optimization platform as claimed in claim 1, characterized in that: The fault analysis module uses a clustering algorithm (KMeans) to identify fault risks based on feature identification: Take each data set G after adding identification and classification in the feature adding module as a data point; Randomly generate multiple clusters and cluster centers; Obtain a data set of the normal working state of photovoltaic facilities from enterprises related to the photovoltaic facilities or a data set of the normal power generation state of photovoltaic facilities in historical records, and use the set as a data point as the center of one of the clusters; G K (A K ,B K ,C K ,D K ) represents the data point at the center point, and K represents the center of the cluster; For each data point, calculate its distance d to the center of each cluster using the Euclidean distance formula: Assign data points to the closest cluster according to the distance d, and assign data points with similar characteristics to the same cluster; The data points are allocated to multiple clusters according to the distance between the data points. If the distances are the same, all the data points are allocated to the cluster centered on the data point of the normal power generation status of the photovoltaic facility. Cluster center update: except for the cluster center of the data point of the normal power generation state of the photovoltaic facility, the average value of the characteristic value of each data point in the other cluster is calculated as the new cluster center; the calculation method is: G′ K is the new cluster center; i is the number of data points in the cluster; Similarly, calculate the new cluster center values ​​of other features until the cluster center no longer changes or reaches the set number of iterations; Finally, it is assigned to a cluster different from the cluster centered on the normal working status data, and the PV facility at this data point is considered to be at risk of failure.

5. A distributed photovoltaic operation fault risk scenario intelligent monitoring and optimization platform as claimed in claim 1, characterized in that: The optimization platform module includes a storage unit, an operation scenario optimization unit, a fault identification optimization unit, a decision optimization unit, a user interface unit, and an early warning unit; The storage unit is used to receive and store monitoring data from the monitoring module, fault risk records identified by the fault analysis module, historical fault data of photovoltaic facilities and photovoltaic equipment fault type processing; The operation scenario optimization unit is used to calculate the risk value of the geographical location; The fault identification optimization unit is used to optimize the analysis of fault risks; The decision optimization unit is used to analyze the failure risk response strategy; The user interface unit is used to generate a user operation interface, visualize system content, and receive relevant photovoltaic data; The early warning unit is used to issue an early warning according to the recognition result.

6. A distributed photovoltaic operation fault risk scenario intelligent monitoring and optimization platform as claimed in claim 5, characterized in that: The operation scenario optimization unit calculates the risk value of the geographical location according to the geographical location information of the monitoring module and the identification result of the fault risk analysis module, and generates a distributed photovoltaic risk value map; Mark each distributed photovoltaic system according to the identification results, with "1" indicating risk and "0" indicating no risk; Count the number of distributed photovoltaics marked with "1" at each geographical location; Select geographical locations with more "0" marks for allocation, or prioritize allocating distributed photovoltaics marked "0" to these locations; There are m distributed photovoltaics, with x m represents the risk mark of the mth distributed photovoltaic, x m is 0 or 1. For a geographical location L, its risk value Z L It can be obtained by summing the risk markers of all distributed photovoltaics at this location: The risk value of the geographical location L is calculated.

7. A distributed photovoltaic operation fault risk scenario intelligent monitoring and optimization platform as claimed in claim 5, characterized in that: The fault identification optimization unit is used to collect the operating status data of the photovoltaic facility when the fault is triggered, divide the set of relevant data records of the photovoltaic facility with confirmed fault into new data points, and transmit them to the fault risk identification module. The fault risk identification module uses the new data point as the center of the cluster for further accurate identification.

8. A distributed photovoltaic operation fault risk scenario intelligent monitoring and optimization platform as claimed in claim 5, characterized in that: The decision optimization unit obtains the photovoltaic facility related fault types and treatment methods through the user interface unit; uses statistical analysis methods to calculate the probability of faults and maintenance methods; F represents the set of fault types, G represents the set of data when the fault occurs, and T represents the set of processing methods; For a certain fault type f, its occurrence frequency can be expressed as: The degree of correlation between the cause of the fault G∈F and the fault type f can be expressed by conditional probability: The corresponding relationship between the processing method T∈F and the fault type f can be expressed as: Based on the photovoltaic data of the fault risk identified by the fault risk analysis module, the fault type with the highest probability and the treatment method with the highest probability under this type are calculated as the output of the maintenance suggestion.

9. A distributed photovoltaic operation fault risk scenario intelligent monitoring and optimization platform as claimed in claim 5, characterized in that: The user interface unit generates an operation interface and visualizes the system content for relevant personnel to input and view relevant information; The input information includes photovoltaic test data of photovoltaic manufacturers, photovoltaic historical operation data, contact information of relevant staff, photovoltaic fault type information and photovoltaic fault handling means information; The viewed information includes the photovoltaic facility status data and fault risk identification results monitored in real time by the monitoring module, and a map is made and displayed based on the geographical location risk value information calculated by the operation scenario optimization unit.

10. A distributed photovoltaic operation fault risk scenario intelligent monitoring and optimization platform as claimed in claim 5, characterized in that: The early warning unit issues an early warning when the fault analysis module identifies a fault risk. The early warning means includes sending the fault risk information to relevant personnel of the photovoltaic facility using text messages according to the contact information input in the operation interface. The content of the text message will describe in detail the fault risk identification results, location and monitoring image information.

Images