A sensor network-based photovoltaic power station remote monitoring system

Abstract

The application relates to the technical field of photovoltaic power stations, and discloses a photovoltaic power station remote monitoring system based on a sensor network, which comprises a monitoring unit, a central control unit, a maintenance unit and the like.

Description

Technical Field

[0001] This application relates to the field of photovoltaic power plant technology, and in particular to a remote monitoring system for photovoltaic power plants based on sensor networks. Background Technology

[0002] Photovoltaic power generation, as a clean and renewable energy source, is receiving increasing attention and importance. Based on different application scenarios, photovoltaic power generation technology is mainly divided into two categories: centralized photovoltaic power plants and distributed photovoltaic power plants. Centralized photovoltaic power plants concentrate large-scale photovoltaic power generation facilities in areas with abundant solar resources, such as deserts and Gobi deserts. These power plants convert solar energy into electrical energy through large-scale photovoltaic panel arrays and transmit the electricity to load centers far from the power generation center through high-voltage transmission networks, meeting the electricity demand of a wide area.

[0003] In existing technologies, remote monitoring systems for photovoltaic power plants can only independently identify equipment faults, but cannot deeply determine the fault propagation path, accurately locate related equipment that may be affected and the derivative faults that may be caused. Furthermore, when inspecting components of a photovoltaic power plant, existing technologies do not consider that the probability of faults caused by components with different lifespans will also vary. Existing algorithms cannot integrate the real-time lifespan status of related equipment to calculate the dynamic risk probability on different propagation paths, thereby failing to achieve the transformation from "post-event alarm" to "pre-event prediction and accurate in-event tracing," resulting in low operation and maintenance efficiency and low reliability of the power plant. Summary of the Invention

[0004] To address the aforementioned technical issues, this application provides a remote monitoring system for photovoltaic power plants based on sensor networks. The system aims to obtain a simulated propagation path of a fault throughout the entire power plant system, accurately locate potentially affected related equipment and its potential derivative faults, integrate the real-time lifespan status of related equipment, and calculate the dynamic risk probability on different propagation paths, thereby greatly improving operation and maintenance efficiency and power plant reliability.

[0005] In some embodiments of this application, a remote monitoring system for photovoltaic power plants based on sensor networks is provided:

[0006] The monitoring unit is used to set up multiple equipment points, install sensor modules on the equipment points, and collect real-time operating data of each equipment point;

[0007] The central control unit is used to set up a fault association directed model and generate expected fault data based on real-time operating data and the fault association directed model.

[0008] The central control unit includes:

[0009] The first processing module is used to pre-define the fault association directed model;

[0010] The second processing module is used to calculate the aging degree of each device point and extract the corresponding fault association directed graph from the fault association directed model based on the aging degree.

[0011] The third processing module is used to generate real-time fault data for equipment points based on real-time operating data.

[0012] The fourth processing module is used to generate expected fault data based on the fault association directed graph and real-time fault data;

[0013] The maintenance unit is used to develop maintenance strategies based on expected fault data.

[0014] In some embodiments of this application, the first processing module is further configured to:

[0015] Each device point is set as a vertex of the fault association directed model, and the fault association relationship between each device point is set as the edge between each vertex of the fault association directed model.

[0016] The fault association relationship is constructed based on a preset fault cascade table;

[0017] Establish a comparison database and obtain comparison datasets for each device point;

[0018] The probability of each fault association at different aging degrees is calculated based on the fault data in the comparison dataset.

[0019] Generate a fault association directed model.

[0020] In some embodiments of this application, calculating the probability of various fault associations at different aging degrees includes:

[0021] Let the vertex sequence be D, D = (d1, d2, ... dn). i …d j …d n ), where n is the total number of vertices; d i For the i-th vertex; d j For the j-th vertex;

[0022] Get vertex d i Compare the fault data with the aging data in the dataset;

[0023] Set vertex d i The aging period T corresponding to different aging degrees q 老 ;

[0024] Vertex d i The fault during the aging detection period corresponding to aging degree q is set as S1;

[0025] Calculate the probability P(ij│q) of edge ij when the aging degree is q;

[0026] P(ij│q)=s2 / s1

[0027] Where s2 is the vertex d i The number of fruit defects within the aging detection period corresponding to aging degree q; s1 is the number of S1;

[0028] In some embodiments of this application, the setting vertex d i The aging period T corresponding to different aging degrees q 老 ,include:

[0029] Construct vertex d i Fault-aging image;

[0030] The fault-aging image is divided into Class I and Class II regions;

[0031] When q is located in the first type of region, T 老 =a1*c i *c 距i Where a1 is a type of coefficient; c i c is the total number of times a fault occurs in the region containing vertex i and q. 距i The time interval for the fault occurring in the region where q is located at vertex i;

[0032] When q is located in the second type of region, T 老 =a2*c i *c 距i ; where a2 is a type II coefficient.

[0033] In some embodiments of this application, the number of fruit faults of vertex i within the aging detection period corresponding to aging degree q includes:

[0034] Obtain the vertex d corresponding to S1 i The lower-level fault;

[0035] Construct a lower-level fault sequence J, J = (j1, j2, ... jj) m …j n ), where n is the total number of lower-level faults corresponding to S1, j m This is the m-th lower-level fault corresponding to S1;

[0036] When j m When the number of faults in the parent level is 1, j will be... m The fault is set to S1;

[0037] When j m When the number of superior faults is greater than 1, set jm The fault intensity sequence L for each superior fault, L = (l1, l2, ... l q …l n ), where l m For j m The fault intensity corresponding to the q-th superior fault; n is the fault intensity of j. m The number of all superior faults;

[0038] Calculate j m The severity of the failure;

[0039] When the fault intensity corresponding to the fault in S1 is equal to j m When the fault intensity is closest, j m The fault is set to S1.

[0040] In some embodiments of this application, the fault intensity includes:

[0041] l m =P q后 / P q前

[0042] Among them, P q后 For j m The average power of the vertex P during the period from the occurrence of the q-th superior fault until the next fault of the vertex containing the superior fault; q前 For j m The average power of the vertex during the period from before the occurrence of the qth superior fault to after the occurrence of the previous fault of the vertex where the superior fault is located.

[0043] In some embodiments of this application, the third processing module is further configured to:

[0044] Real-time acquisition of power data sequences from various device points;

[0045] Determine whether the power data of each device point meets the fault conditions;

[0046] If the conditions are met, it is determined that the device has experienced a real-time fault; if the conditions are not met, it is determined that the device has not experienced a real-time fault.

[0047] The fault condition is that the power data is continuously lower than the fault threshold for more than the first time window.

[0048] In some embodiments of this application, the fourth processing module is further configured to:

[0049] The device point where a real-time fault occurs is set as the fault source vertex D. 源 D 源 =(d 源1 d 源2 , ...d 源m …d源n ), where d 源m Let be the i-th fault source vertex; n is the total number of fault source vertices;

[0050] Traversing the fault-related directed graph d m Given all the edges pointed to, obtain the probability of each edge, where the probability represents the probability of a failure caused by vertex i;

[0051] Obtain the fault severity of all fault source vertices that cause resultant faults.

[0052] In some embodiments of this application, the maintenance module is further used for:

[0053] Set up points to be inspected;

[0054] Implement maintenance strategies for the points to be inspected;

[0055] The points to be inspected are equipment points that meet the inspection conditions.

[0056] In some embodiments of this application, the maintenance conditions include:

[0057] The maintenance conditions include: the first maintenance condition and the second maintenance condition;

[0058] The first maintenance condition is that a fault occurs at the equipment point;

[0059] The second maintenance condition is that the probability of a fault occurring at the equipment point is greater than the first threshold or the severity of the fault is greater than the second threshold.

[0060] Compared with existing technologies, the remote monitoring system for photovoltaic power plants based on sensor networks in this application has the following advantages:

[0061] A dynamic and self-evolving fault prediction and propagation analysis system has been established. This system integrates real-time sensor data with an advanced equipment lifespan database to accurately calculate the real-time health and remaining effective lifespan of each piece of equipment. More importantly, it incorporates a fault cascading effect model based on directed graphs and probabilistic graphs. When any equipment fails, the system can not only immediately identify it but also simulate the propagation path of the fault throughout the entire power plant system based on this model, accurately locating potentially affected related equipment and its possible derivative faults. The core algorithm of the model deeply integrates the real-time lifespan status of related equipment to calculate the dynamic risk probability on different propagation paths, thereby achieving a shift from "post-event alarm" to "pre-event prediction and precise in-event tracing," greatly improving operation and maintenance efficiency and power plant reliability. Attached Figure Description

[0062] Figure 1This application describes a remote monitoring system for a photovoltaic power plant based on a sensor network. Detailed Implementation

[0063] The specific embodiments of this application will be described in further detail below with reference to the accompanying drawings and examples. The following examples are used to illustrate this application, but are not intended to limit the scope of this application.

[0064] In the description of this application, it should be understood that the terms "center", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this application.

[0065] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this application, unless otherwise stated, "a plurality of" means two or more.

[0066] In the description of this application, it should be noted that, unless otherwise expressly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection between two components. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances.

[0067] like Figure 1 As shown in the figure, an operation and maintenance management system for a new energy power plant according to an embodiment of this application includes:

[0068] The monitoring unit is used to set up multiple equipment points, install sensor modules on the equipment points, and collect real-time operating data of each equipment point;

[0069] The central control unit is used to set up a fault association directed model and generate expected fault data based on real-time operating data and the fault association directed model.

[0070] The central control unit includes:

[0071] The first processing module is used to pre-define the fault association directed model;

[0072] The second processing module is used to calculate the aging degree of each device point and extract the corresponding fault association directed graph from the fault association directed model based on the aging degree.

[0073] The third processing module is used to generate real-time fault data for equipment points based on real-time operating data.

[0074] The fourth processing module is used to generate expected fault data based on the fault association directed graph and real-time fault data;

[0075] The maintenance unit is used to develop maintenance strategies based on expected fault data.

[0076] Specifically, the sensor module can collect various real-time operating data, including: temperature data, image data, current data, voltage data, and electrical power data.

[0077] Specifically, the fault association directed model sets all device points as vertices of a directed graph, and the association between faults generated by device points is used as the edge connecting two vertices. This edge is directional. For example, if a fault generated by vertex a induces a fault in vertex b, then edge ab is from a to b, and cannot be represented as from b to a. The edge pointing to a vertex is called an incoming edge, and the edge pointing out from a vertex is called an outgoing edge. There can be multiple incoming edges and outgoing edges. When a fault can induce a fault in b, b is called the lower point of a, and a is called the upper point of b.

[0078] The probability of each edge in the model changes with the degree of aging. Once the degree of aging of each device point is determined, the probability of each edge in the model will also be determined, thus obtaining a fault-related directed graph.

[0079] Specifically, real-time faults are faults detected by sensors based on real-time operational data.

[0080] Specifically, expected fault data refers to the probability of other equipment points failing due to the correlation between faults caused by real-time faults.

[0081] Specifically, a power generation device can contain multiple device points, and a photovoltaic module is set as one device point.

[0082] In some embodiments of this application, the first processing module is further configured to:

[0083] Each device point is set as a vertex in the fault association directed model, and the fault association relationship between each device point is set as the edge between each vertex in the fault association directed model.

[0084] Fault associations are constructed based on a pre-defined fault cascade table;

[0085] Establish a comparison database and obtain comparison datasets for each device point;

[0086] The probability of the correlation between various faults at different aging levels is calculated based on the fault data in the comparative dataset.

[0087] Generate a directed model of fault association.

[0088] Specifically, the health status threshold and the abnormal status threshold are determined based on the average value of the critical values ​​of the equipment points in historical monitoring data when they transition from a healthy state and an abnormal state to an abnormal state and a dangerous state, respectively.

[0089] Specifically, the preset fault cascading table is a summary of the fault relationships that can occur between various components, based on the experience of maintenance personnel and experts in this field.

[0090] Specifically, the comparison database consists of the operating data of selected photovoltaic power plants that meet preset conditions, including: the operating time of the power plant reaches its rated lifespan, such as 15 years; the scale of the power plant; the ambient light intensity and temperature; the operation and maintenance method is the same as this power plant; and the monthly average power generation is less than 5% of the monthly average power generation of this power plant. The comparison data of the entire operating cycle of each of its equipment points are collected.

[0091] Specifically, the comparative data includes fault data, power data, and aging data.

[0092] Specifically, the comparison database is a collection of comparison data for all device points; the comparison dataset is a collection of comparison data for a single device point.

[0093] Specifically, the equipment points are moved to the comparison power station in the exact same way as the equipment points of this power station, and then their operating data over the complete life cycle of the power station is extracted. The data includes electrical power.

[0094] In some embodiments of this application, the probability of various fault associations at different aging degrees is calculated, including:

[0095] Let the vertex sequence be D, D = (d1, d2, ... dn). i …d j …d n ), where n is the total number of vertices; d i For the i-th vertex; d j For the j-th vertex;

[0096] Get vertex d i Compare the fault data with the aging data in the dataset;

[0097] Set vertex d i The aging period T corresponding to different aging degrees q老 ;

[0098] Vertex d i The fault during the aging detection period corresponding to aging degree q is set as S1;

[0099] Calculate the probability P(ij│q) of edge ij when the aging degree is q;

[0100] P(ij│q)=s2 / s1

[0101] Where s2 is the vertex d i The number of fruit failures within the aging detection period corresponding to aging degree q; s1 is the number of S1.

[0102] Specifically, the vertex sequence contains all device points.

[0103] Specifically, the fault data includes the time when the fault occurred and the power at the time of the fault.

[0104] Specifically, the aging degree is a value describing the degree of degradation of a device. It is the ratio of the power of the device under normal operating conditions to the rated power of the device. The aging degree is an array that changes from 0 to 1, increasing in increments of 0.01, such as: 0.01, 0.02, 0.03...0.09, 1. When calculating the aging degree, the original result is retained to two decimal places and rounded to obtain a modified result that conforms to the above array format.

[0105] Specifically, the aging detection period is a time period. After determining the aging degree at a certain point in time at a certain peak, if the aging detection period is h, then only the fault data during the time period h after the moment when the aging degree is calculated is considered, so as to avoid the problem that the fault data of too long a time period has low reference value.

[0106] Specifically, S1 represents a series of failures at this peak during the aging detection period after the aging detection period has been determined.

[0107] Specifically, s1 represents the number of faults in this series of faults.

[0108] Specifically, P(ij|q) represents the probability that vertex j will fail due to a failure of vertex i when vertex i has an aging degree of q.

[0109] Specifically, if a fault occurs at vertex i, and due to the correlation of faults, vertex j also causes a fault to occur, then the fault caused by vertex j is called the resultant fault of the fault caused by vertex i. Conversely, the fault caused by i is called the cause fault of the fault caused by j.

[0110] In some embodiments of this application, vertex d is set i The aging period T corresponding to different aging degrees q 老 ,include:

[0111] Construct vertex d i Fault-aging image;

[0112] The fault-aging image is divided into Class I and Class II regions;

[0113] When q is located in a certain region, T 老 =a1*c i *c 距i Where a1 is a type of coefficient; c i c is the total number of times a fault occurs in the region containing vertex i and q. 距i The time interval for the fault occurring in the region where q is located at vertex i;

[0114] When q is in the second type region, T 老 =a2*c i *c 距i ; where a2 is a type II coefficient.

[0115] Specifically, the fault-aging graph consists of a series of points and a two-dimensional xy coordinate axis. The horizontal axis of these points represents the occurrence time of each fault, and the vertical axis represents the fault intensity. The horizontal axis represents time, and the vertical axis represents fault intensity. Based on the aging data, the time corresponding to different aging degrees can be obtained on the horizontal axis. The time interval formed by the time points of each pair of consecutive faults is called the fault interval.

[0116] Specifically, a type I region is defined as a region where the difference between adjacent intervals is less than 0.1 times the length of the previous fault interval; a type II region is defined as a region where the difference between adjacent intervals is greater than 0.1 times the length of the previous fault interval.

[0117] Specifically, a1 is preferably 0.3; a2 is preferably 0.8.

[0118] Specifically, c 距i This represents the average time interval between failures in the area.

[0119] In some embodiments of this application, the number of fruit faults of vertex i during the aging detection period corresponding to aging degree q includes:

[0120] Get the vertex d corresponding to S1 i The lower-level fault;

[0121] Construct a lower-level fault sequence J, J = (j1, j2, ... jj) m …j n), where n is the total number of lower-level faults corresponding to S1, j m This is the m-th lower-level fault corresponding to S1;

[0122] When j m When the number of faults in the parent level is 1, j will be... m The fault is set to S1;

[0123] When j m When the number of superior faults is greater than 1, set j m The fault intensity sequence L for each superior fault, L = (l1, l2, ... l q …l n ), where l m For j m The fault intensity corresponding to the q-th superior fault; n is the fault intensity of j. m The number of all superior faults;

[0124] Calculate j m The severity of the failure;

[0125] When the fault intensity corresponding to the fault in S1 is equal to j m When the fault intensity is closest, j m The fault is set to S1.

[0126] Specifically, a lower-level fault is as follows: if a fault k at vertex j is affected by a fault f at a higher point i during the fault period of j, then k is called a lower-level fault of f; otherwise, f is a higher-level fault of k.

[0127] A subordinate fault is different from a resultant fault. A fault may exist simultaneously within the fault-affecting period of multiple faults, and is a subordinate fault of multiple faults; however, only the superior fault that caused the fault can be the cause of the fault.

[0128] Specifically, the fault impact period is determined by the duration of the impact of the fault on the power station. If the impact of the fault is short-term, the fault impact period is the time from the occurrence of the fault until it is repaired. If the fault has a permanent impact on the equipment, such as the hot spot effect caused by sand and dust, which deforms the surface of the photovoltaic panel and permanently reduces the power generation of the photovoltaic panel, the fault impact period is set until the photovoltaic panel is replaced or the photovoltaic panel reaches the rated life of the photovoltaic power station and is replaced or overhauled on a large scale.

[0129] Specifically, j m The fault set as S1 is j. m It was caused by a fault in S1.

[0130] In some embodiments of this application, the fault intensity includes:

[0131] l m =P q后 / P q前

[0132] Among them, P q后 For j m The average power of a vertex during the period from the occurrence of the q-th superior fault until the next fault occurs at the vertex containing the superior fault; P q前 For j m The average power of a vertex during the period from before the occurrence of the qth superior fault to after the occurrence of the previous fault of the vertex where the superior fault is located.

[0133] Specifically, fault intensity indicates the severity of the power reduction caused by a fault at a point in the equipment.

[0134] In some embodiments of this application, the third processing module is further configured to:

[0135] Real-time acquisition of power data sequences from various device points;

[0136] Determine whether the power data of each device point meets the fault conditions;

[0137] If the conditions are met, the device is determined to have experienced a real-time fault; if the conditions are not met, the device is determined not to have experienced a real-time fault.

[0138] The fault condition is that the power data is continuously lower than the fault threshold for more than the first time window.

[0139] Specifically, the fault threshold is the average power of the device during its normal operating period in historical operating data; normal operation is the operating state of the device without any faults.

[0140] Specifically, the first time window is preferably 0.01*t. 平 ;t 平 This refers to the average time interval between device malfunctions.

[0141] In some embodiments of this application, the fourth processing module is further configured to:

[0142] The device point where a real-time fault occurs is set as the fault source vertex D. 源 D 源 =(d 源1 d 源2 , ...d 源m …d 源n ), where d 源m Let be the i-th fault source vertex; n is the total number of fault source vertices;

[0143] Traversing the fault-related directed graph d mGiven all the edges pointed to, obtain the probability of each edge, where the probability represents the probability of a failure caused by vertex i;

[0144] Obtain the fault severity of all fault source vertices that cause resultant faults.

[0145] Specifically, the fault source vertex is the equipment point that has failed, detected by sensors through real-time operational data of the system.

[0146] Specifically, the severity of all resulting faults from fault source vertices is obtained. Fault source vertices can trigger other faults based on fault correlations, and these faults have varying probabilities of occurrence. The probability of the resulting fault is then converted from real-time fault data into expected fault data.

[0147] In some embodiments of this application, the maintenance module is also used for:

[0148] Set up the points to be inspected;

[0149] Implement maintenance strategies for the maintenance points;

[0150] The points to be inspected are the equipment points that meet the conditions for inspection.

[0151] Specifically, once the points to be repaired are identified, the necessary equipment for repairing each point is brought along. If a repair cannot be completed in one go, the line repair is performed by testing the faults detected by the sensors. The faults are then sorted from highest to lowest probability of occurrence and tested.

[0152] In some embodiments of this application, the maintenance conditions include:

[0153] The maintenance conditions include: the first maintenance condition and the second maintenance condition;

[0154] The first maintenance condition is that a fault occurs at the equipment point;

[0155] The second maintenance condition is that the probability of a fault occurring at the equipment point is greater than the first threshold or the severity of the fault is greater than the second threshold.

[0156] Specifically, the first threshold is preferably 0.7; the second threshold is preferably 0.5.

[0157] The above are merely preferred embodiments of this application. It should be noted that, for those skilled in the art, several improvements and substitutions can be made without departing from the technical principles of this application, and these improvements and substitutions should also be considered within the scope of protection of this application.

Claims

1. A remote monitoring system for photovoltaic power plants based on sensor networks, characterized in that, include: The monitoring unit is used to set up multiple equipment points, install sensor modules on the equipment points, and collect real-time operating data of each equipment point; The central control unit is used to set up a fault association directed model and generate expected fault data based on real-time operating data and the fault association directed model. The central control unit includes: The first processing module is used to pre-define the fault association directed model; The second processing module is used to calculate the aging degree of each device point and extract the corresponding fault association directed graph from the fault association directed model based on the aging degree. The third processing module is used to generate real-time fault data for equipment points based on real-time operating data. The fourth processing module is used to generate expected fault data based on the fault association directed graph and real-time fault data; The maintenance unit is used to develop maintenance strategies based on expected fault data. The first processing module is also used for: Each device point is set as a vertex of the fault association directed model, and the fault association relationship between each device point is set as the edge between each vertex of the fault association directed model. The fault association relationship is constructed based on a preset fault cascade table; Establish a comparison database and obtain comparison datasets for each device point; The probability of each fault association at different aging degrees is calculated based on the fault data in the comparison dataset. Generate a directed fault association model; The calculation of the probability of various fault correlations at different aging degrees includes: Let the vertex sequence D be. =( , ,… … … ), n is the total number of vertices; For the i-th vertex; For the j-th vertex; Get Vertex Compare the fault data with the aging data in the dataset; Set vertex The aging period T corresponding to different aging degrees q 老 ; Vertex The fault setting within the aging detection period corresponding to aging degree q is as follows: ; Calculate the probability of edge ij when the aging degree is q. ; ； in, As vertex The number of fruit defects during the aging period corresponding to aging degree q; for Quantity; The set vertex The aging period T corresponding to different aging degrees q 老 ,include: Constructing vertices Fault-aging image; The fault-aging image is divided into Class I and Class II regions; When q is located in the first type of region, T 老 = * * ;in, It is a type of coefficient; This represents the total number of failures that occurred in the region where vertex i is located (q). The time interval for the fault occurring in the region where q is located at vertex i; When q is located in the second type of region, T 老 = * * ;in, These are Class II coefficients.

2. The remote monitoring system for photovoltaic power plants based on sensor networks as described in claim 1, characterized in that, The number of faults at vertex i within the aging detection period corresponding to aging degree q includes: Obtain the Corresponding vertex The lower-level fault; Construct a lower-level fault sequence J, J = (j1, j2, ... jj) m …j n ), where n is The corresponding total number of subordinate faults, j m for The corresponding m-th lower-level fault; When j m When the number of faults in the parent level is 1, j will be... m Set as The result was faulty; When j m When the number of superior faults is greater than 1, set j m The fault intensity series of each superior fault , , For j m The fault intensity corresponding to the q-th superior fault; n is the fault intensity of j. m The number of all superior faults; Calculate j m The severity of the failure; when The fault intensity corresponding to the fault in j m When the fault intensity is closest, j m Set as The result was faulty.

3. The remote monitoring system for photovoltaic power plants based on sensor networks as described in claim 2, characterized in that, The severity of the fault includes: ； in, For j m The average power of the vertex after the qth superior fault occurs until the next fault of the vertex where the superior fault occurred; For j m The average power of the vertex during the period from before the occurrence of the qth superior fault to after the occurrence of the previous fault of the vertex where the superior fault is located.

4. The remote monitoring system for photovoltaic power plants based on sensor networks as described in claim 1, characterized in that, The third processing module is also used for: Real-time acquisition of power data sequences from various device points; Determine whether the power data of each device point meets the fault conditions; If the conditions are met, it is determined that the device has experienced a real-time fault; if the conditions are not met, it is determined that the device has not experienced a real-time fault. The fault condition is that the power data is continuously lower than the fault threshold for more than the first time window.

5. The remote monitoring system for photovoltaic power plants based on sensor networks as described in claim 4, characterized in that, The fourth processing module is also used for: Set the device point where the real-time fault occurs as the fault source vertex. , ,in, Let be the i-th fault source vertex; n is the total number of fault source vertices; Traversing the fault-related directed graph Given all the edges pointed to, obtain the probability of each edge, where the probability represents the probability of a failure caused by vertex i; Obtain the fault severity of all fault source vertices that cause resultant faults.

6. The remote monitoring system for photovoltaic power plants based on sensor networks as described in claim 5, characterized in that, The maintenance unit is also used for: Set up the points to be inspected; Implement maintenance strategies for the points to be inspected; The points to be inspected are equipment points that meet the inspection conditions.

7. The remote monitoring system for photovoltaic power plants based on sensor networks as described in claim 6, characterized in that, The maintenance conditions include: The maintenance conditions include: the first maintenance condition and the second maintenance condition; The first maintenance condition is that a fault occurs at the equipment point; The second maintenance condition is that the probability of a fault occurring at the equipment point is greater than the first threshold or the severity of the fault is greater than the second threshold.

Images