Intelligent detection method and device for photovoltaic equipment

Abstract

The present invention discloses an intelligent detection method and device for photovoltaic equipment, and relates to the technical field of photovoltaic equipment. The method includes: taking photoelectric effect and non-invasive detection as the benchmark, performing data-driven modeling, and building a cascade detection module; continuously monitoring the photoelectric effect data of the target photovoltaic power station and transmitting it back to the continuous detection unit to determine the first detection data; making an over-limit judgment on the first detection data and generating a trigger detection instruction; reading the non-invasive detection data, combining the trigger detection unit to perform detection positioning, and determining the second detection data; determining the redundant configuration of the target photovoltaic power station, performing fault-tolerant detection, and determining the third detection data; based on the first detection data, the second detection data and the third detection data, as the equipment detection result. The technical problem of low detection efficiency and accuracy of photovoltaic equipment in the prior art is solved, and the technical effect of improving the detection efficiency and accuracy of photovoltaic equipment is achieved by building a cascade detection module.

Description

Technical Field

[0001] The present invention relates to the technical field of photovoltaic equipment, and in particular to an intelligent detection method and device for photovoltaic equipment. Background Art

[0002] In the rapidly developing field of renewable energy, photovoltaic technology, as an important pillar, is gradually becoming a key force in the transformation of the global energy structure. With the widespread deployment and expansion of photovoltaic power stations, how to ensure the continuous and efficient operation of these key facilities has become the focus of industry attention. However, photovoltaic equipment will inevitably encounter various performance degradation or failure problems in the complex and changeable natural environment and long-term operation. If these problems are not discovered and solved in time, they will directly affect the power generation efficiency of photovoltaic power stations. Traditional photovoltaic equipment detection methods often rely on manual inspections and regular maintenance. This method is not only inefficient, but also difficult to achieve real-time and accurate monitoring of the status of photovoltaic equipment. In addition, the location and diagnosis of faults often rely on experience judgment, lack of scientific and effective means, resulting in long maintenance cycles and high costs. Summary of the invention

[0003] The present application provides an intelligent detection method and device for photovoltaic equipment, which solves the technical problem of low detection efficiency and accuracy of photovoltaic equipment in the prior art.

[0004] In view of the above problems, the present application provides an intelligent detection method and device for photovoltaic equipment.

[0005] In a first aspect of the present application, a smart detection method for a photovoltaic device is provided, the method comprising:

[0006] Based on the photoelectric effect and non-invasive detection, data-driven modeling is carried out to build a cascade detection module, which includes a continuous detection unit and a trigger detection unit; the photoelectric effect data of the target photovoltaic power station is continuously monitored and transmitted back to the continuous detection unit, and the analysis based on the photoelectric conversion efficiency is carried out by combining dynamic and static methods to determine the first detection data; the first detection data is judged as out of limit, and a trigger detection instruction is generated, and the trigger detection instruction is a start instruction for non-invasive detection; the non-invasive detection data is read, and the data is multi-scale decomposition and topological exclusion detection positioning are carried out in combination with the trigger detection unit to determine the second detection data, wherein the detection positioning is carried out based on the direct response data and the linear function relationship; the redundant configuration of the target photovoltaic power station is determined, and the fault-tolerant detection of the photovoltaic equipment is carried out to determine the third detection data, wherein the redundant configuration includes structural redundancy and system redundancy; the first detection data, the second detection data and the third detection data are used as the equipment detection result.

[0007] A second aspect of the present application provides an intelligent detection device for a photovoltaic device, the device comprising:

[0008] A cascade detection module building component, the cascade detection module building component is used to perform data-driven modeling based on photoelectric effect and non-invasive detection, and build a cascade detection module, the cascade detection module includes a continuous detection unit and a trigger detection unit; a data analysis component, the data analysis component is used to continuously monitor the photoelectric effect data of the target photovoltaic power station and transmit it back to the continuous detection unit, and perform a dynamic and static analysis based on the photoelectric conversion efficiency to determine the first detection data; an over-limit judgment component, the over-limit judgment component is used to perform an over-limit judgment on the first detection data and generate a trigger detection instruction, the trigger detection instruction is a start instruction for non-invasive detection; detection A positioning component, the detection and positioning component is used to read non-intrusive detection data, and perform multi-scale data decomposition and topology exclusion detection and positioning in combination with the trigger detection unit to determine second detection data, wherein the detection and positioning is performed based on the direct response data and the linear function relationship; a fault-tolerant detection component, the fault-tolerant detection component is used to determine the redundant configuration of the target photovoltaic power station, perform fault-tolerant detection of photovoltaic equipment, and determine third detection data, wherein the redundant configuration includes structural redundancy and system redundancy; a detection result generation component, the detection result generation component is used to use the first detection data, the second detection data and the third detection data as the device detection result.

[0009] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0010] First, based on the photoelectric effect and non-invasive detection, data-driven modeling is carried out to build a cascade detection module, which includes a continuous detection unit and a trigger detection unit. Then, the photoelectric effect data of the target photovoltaic power station is continuously monitored and transmitted back to the continuous detection unit, and the analysis based on the photoelectric conversion efficiency is carried out by combining dynamic and static methods to determine the first detection data. Then, the first detection data is judged to be over-limit, and a trigger detection instruction is generated, which is the start instruction for non-invasive detection. Next, the non-invasive detection data is read, and the data is multi-scale decomposition and topological exclusion detection positioning are carried out in combination with the trigger detection unit to determine the second detection data, wherein the detection positioning is based on the direct response data and the linear function relationship. At the same time, the redundant configuration of the target photovoltaic power station is determined, and the fault-tolerant detection of the photovoltaic equipment is carried out to determine the third detection data, wherein the redundant configuration includes structural redundancy and system redundancy. Finally, the first detection data, the second detection data and the third detection data are used as the equipment detection result. The technical problem of low detection efficiency and accuracy of photovoltaic equipment in the prior art is solved, and the technical effect of improving the detection efficiency and accuracy of photovoltaic equipment is achieved by building a cascade detection module. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0012] Figure 1 A schematic diagram of a flow chart of an intelligent detection method for photovoltaic equipment provided in an embodiment of the present application;

[0013] Figure 2 A schematic diagram of the structure of an intelligent detection device for photovoltaic equipment provided in an embodiment of the present application.

[0014] Explanation of the accompanying drawings: cascade detection module building component 11, data analysis component 12, over-limit determination component 13, detection and positioning component 14, fault-tolerant detection component 15, detection result generation component 16. DETAILED DESCRIPTION

[0015] The present application solves the technical problem of low detection efficiency and accuracy of photovoltaic equipment in the prior art by providing an intelligent detection method and device for photovoltaic equipment.

[0016] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0017] It should be noted that the terms "including" and "having" are intended to cover non-exclusive inclusions. For example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules that are not explicitly listed or are inherent to these processes, methods, products or devices.

[0018] Embodiment 1, as Figure 1 As shown, the present application provides an intelligent detection method for photovoltaic equipment, wherein the method comprises:

[0019] Based on the photoelectric effect and non-invasive detection, data-driven modeling is carried out to build a cascade detection module, which includes a continuous detection unit and a trigger detection unit.

[0020] Based on the principle of photoelectric effect and non-invasive detection technology, a data-driven detection model was constructed, and a cascade detection module was built accordingly to improve the detection efficiency and accuracy of photovoltaic equipment. Among them, the cascade detection module consists of a continuous detection unit and a trigger detection unit, which work together to achieve comprehensive equipment monitoring. The continuous detection unit is responsible for continuously and uninterruptedly collecting the photoelectric effect data of the photovoltaic power station, including key parameters such as current, voltage, and light intensity, which directly reflect the operating status and photoelectric conversion efficiency of the photovoltaic equipment; the trigger detection unit is responsible for receiving the first detection data generated by the continuous detection unit and making an over-limit judgment. The over-limit judgment is based on the preset thresholds. When the detection data exceeds these thresholds, it indicates that the equipment may be abnormal or the performance has degraded.

[0021] Furthermore, the construction of the cascade detection module includes:

[0022] Obtain the photoelectric conversion efficiency of the target photovoltaic power station; build a photovoltaic decision tree based on the photoelectric conversion efficiency and in combination with historical photovoltaic record data, wherein the photovoltaic decision tree executes photovoltaic effect decisions based on photovoltaic scenarios; construct a continuous detection unit based on the photovoltaic decision tree; parallelize the continuous detection unit and the trigger detection unit, and perform data-driven training to generate the cascade detection module.

[0023] By collecting the real-time operation data of the target photovoltaic power station, including key parameters such as light intensity, temperature, photovoltaic panel output voltage and current, the photoelectric conversion efficiency of the target photovoltaic power station can be calculated using these data. Photoelectric conversion efficiency is an important indicator for evaluating the performance of photovoltaic systems, reflecting the ability of photovoltaic systems to convert solar energy into electrical energy. Based on the obtained photoelectric conversion efficiency, combined with historical photovoltaic record data (including the performance of photovoltaic power stations under different weather conditions, seasonal changes, equipment aging and other factors), machine learning algorithms (such as decision tree algorithms) can be used to build photovoltaic decision trees. Photovoltaic decision trees can make decisions on the optimal photovoltaic operation strategy based on input characteristics such as photoelectric conversion efficiency and weather conditions. These strategies may include adjusting the angle of photovoltaic panels, cleaning photovoltaic panels, replacing aging components, etc., to improve the photovoltaic effect and power station efficiency; after the photovoltaic decision tree is built, a continuous detection unit can be built based on its decision logic and prediction results. The continuous detection unit is responsible for collecting the operation data of the photovoltaic power station in real time and using the photovoltaic decision tree for analysis and judgment. By running the continuous detection unit and the trigger detection unit in parallel, the photovoltaic power station can be fully monitored. The continuous detection unit is responsible for continuous monitoring and preliminary analysis, while the trigger detection unit is responsible for initiating an in-depth non-intrusive detection process when receiving a specific trigger signal (such as an abnormality found by the continuous detection unit). Furthermore, through a data-driven training process, including using a large amount of historical data and real-time data to train models, adjust parameters, verify performance, etc., the continuous detection unit, the trigger detection unit, and the interaction logic between them can be optimized and adjusted to generate an efficient cascade detection module. In summary, by obtaining the photoelectric conversion efficiency, building a photovoltaic decision tree, building a continuous detection unit, running the continuous detection unit and the trigger detection unit in parallel, and data-driven training, an efficient cascade detection module can be generated to monitor and manage the operating status of the photovoltaic power station.

[0024] Furthermore, the trigger detection unit is constructed, including:

[0025] The equipment installed capacity of the target photovoltaic power station is read to construct a photovoltaic equipment topology; a data acquisition method for photovoltaic detection is determined, data tags are decomposed at multiple scales, and a multi-scale feature pyramid is constructed; a trigger detection unit is constructed by combining the photovoltaic equipment topology with the multi-scale feature pyramid.

[0026] By reading the equipment installed capacity of the target photovoltaic power station, a detailed equipment list of the photovoltaic power station is obtained, including but not limited to the model, quantity and installation location of photovoltaic modules, inverters, junction boxes, transformers and other equipment; based on the obtained equipment information. Use graphical tools (such as AutoCAD, etc.) to build a topological diagram of the photovoltaic power station. The topological diagram can clearly show the connection relationship between the equipment, the flow of current and voltage, and the location distribution of each device in the power station. By measuring the output power, current, voltage and other parameters of the photovoltaic module, photovoltaic detection data can be obtained; according to the requirements of the detection task, define appropriate data labels, such as power generation, equipment failure type, etc., and decompose the data labels according to the time scale (such as hours, days, months, years) or spatial scale (such as a single photovoltaic panel, photovoltaic panel array, the entire photovoltaic power station); use feature maps (such as image pyramids, feature pyramids DEF, etc.) to build a multi-scale feature pyramid, extract key features at different scales, and perform feature fusion to obtain a more comprehensive description of the system status. The device information in the photovoltaic device topology is integrated with the feature information in the multi-scale feature pyramid to construct a trigger detection unit. The trigger detection unit can monitor and evaluate the operating status of the photovoltaic power station in real time according to preset thresholds or conditions. For example, fault detection is triggered when the power generation is lower than the set threshold; warning is triggered when the equipment temperature is abnormal, etc.

[0027] The photoelectric effect data of the target photovoltaic power station is continuously monitored and transmitted back to the continuous detection unit, and the analysis based on the photoelectric conversion efficiency is performed in a dynamic and static manner to determine the first detection data.

[0028] Continuously monitor the photoelectric effect data of the target photovoltaic power station, such as the power, voltage, current, temperature and other parameters of key components such as solar panels, battery packs, inverters, etc., and transmit these photoelectric effect data to the continuous detection unit. Through the dynamic and static analysis method, combined with the change trend of photoelectric conversion efficiency, the equipment status is evaluated and the first detection data is generated.

[0029] Furthermore, the determining the first detection data includes:

[0030] Acquire the real-time photovoltaic environment, and determine the first photoelectric effect data based on the photovoltaic conversion efficiency in combination with the photovoltaic decision tree; transmit back the second photoelectric effect data of the target photovoltaic power station, calibrate the first photoelectric effect data, and determine the static detection data; measure the photovoltaic effect trend with photovoltaic environment variables, and determine the dynamic detection data by calibrating the photovoltaic effect trend; determine the first detection data based on the static detection data and the dynamic detection data.

[0031] Preferably, the photovoltaic power station's sensors are used to monitor photovoltaic environmental parameters such as solar irradiance, temperature, humidity, and wind speed in real time to obtain the real-time photovoltaic environment; the real-time photovoltaic environmental data is input into the photovoltaic decision tree to obtain the first photoelectric effect data, including the expected photoelectric conversion efficiency, power generation forecast, etc.; the second photoelectric effect data refers to the data directly obtained from the actual operation of the photovoltaic power station, including the actual power generation, voltage, current and other parameters, the second photoelectric effect data is collated with the first photoelectric effect data to obtain static detection data, static monitoring data refers to stable and repeatable measurement data of the operating status of the photovoltaic power station at a specific time point or time period, these data reflect the basic performance of the photovoltaic power station under the current environmental conditions, that is, the first photoelectric effect data is calibrated according to the second photoelectric effect data, after calibration, the first photoelectric effect data is combined with the stable and repeatable parts of the second photoelectric effect data to determine the static detection data. Through time series analysis, regression analysis and other methods, the photovoltaic environmental variables (such as solar irradiance, temperature, etc.) in historical data and real-time data are used to measure the trend of the photovoltaic effect. By analyzing the relationship between photovoltaic environmental variables and the photoelectric effect, the changing trend of the photovoltaic effect is determined, such as seasonal changes, weather changes, etc.; the photovoltaic effect trend obtained by analysis is checked with the actual situation, and based on the checked photovoltaic effect trend, the dynamic detection data of the photovoltaic power station is determined. The dynamic detection data refers to the data reflecting the changes in the power generation performance of the photovoltaic power station over time, such as real-time power generation data, etc. By integrating the static detection data with the dynamic detection data, the first detection data can be determined.

[0032] An over-limit determination is performed on the first detection data to generate a trigger detection instruction, where the trigger detection instruction is a start instruction for performing non-intrusive detection.

[0033] The thresholds of key parameters of the photovoltaic power station (such as power generation, photoelectric conversion efficiency, battery temperature, voltage fluctuation, etc.) are comprehensively determined based on factors such as the design specifications of the photovoltaic power station, historical operating data, and industry standards; the first detection data is compared with the set threshold, and based on the comparison result, it is determined whether one or more parameters of the photovoltaic power station exceed the set threshold; if a parameter exceeds the threshold, it is deemed to be out of limit; when the judgment result is out of limit, a trigger detection instruction will be generated, and the trigger detection instruction is a start instruction for non-invasive detection. The non-invasive detection method can obtain in-depth internal status information of the equipment without interrupting the normal operation of the photovoltaic equipment.

[0034] The non-intrusive detection data is read, and the data is multi-scale decomposition and topology exclusion detection positioning are performed in combination with the trigger detection unit to determine the second detection data, wherein the detection positioning is performed based on the direct response data and the linear function relationship.

[0035] Based on the trigger detection instruction, the photovoltaic equipment is non-invasively detected. During the non-invasive detection process, the non-invasive detection data is read, including the temperature distribution, voltage and current waveforms, power output, etc. of the photovoltaic components, which reflect the operating status of the photovoltaic power station at different times and under different conditions. The non-invasive detection data is synchronized with the data of the trigger detection unit, and the data is multi-scale decomposed through methods such as wavelet transform and Fourier transform. The data is decomposed into components of different frequencies or different time periods, so as to analyze the operating status of the photovoltaic power station from different time scales and spatial scales. Topological exclusion detection and positioning is a detection method based on the topological structure and data correlation of the photovoltaic power station. It uses the electrical and physical connection relationship between the various components in the photovoltaic power station to gradually narrow the scope of the problem through the elimination method, and finally locates the specific fault point or abnormal area. Based on the direct response data, that is, those data directly related to the trigger detection unit, a mathematical model is established in combination with the linear function relationship between different parameters of the photovoltaic equipment (such as voltage and current, power and irradiance, etc.), and the detection data is further analyzed and predicted. By comparing the differences between the actual data and the model predicted data, abnormal points or areas can be identified; combining the multi-scale decomposition and the analysis results of the direct response data, the topological structure information is used to gradually exclude normal areas and finally locate the problem area, that is, the second detection data. The second detection data is an accurate description of the operating status and problem location of the photovoltaic power station.

[0036] Furthermore, topological exclusion detection and positioning is performed, including:

[0037] Determine a multi-scale decomposition feature, wherein the multi-scale decomposition feature has a device tag identifier and a time phase identifier; traverse the multi-scale decomposition feature to determine a first exclusion feature group; based on the first exclusion feature group, determine a vector value of a feature response, and determine a linear function relationship based on a time phase feature trend; map the vector value and the linear function relationship and make a judgment to determine a step of detection data.

[0038] Preferably, the non-intrusive detection data is decomposed at multiple scales to extract features at different time scales and frequency scales, which can reflect the operating status and performance changes of the photovoltaic power station at different time periods and frequencies; in the multi-scale decomposition process, each feature is assigned a device label identifier (such as photovoltaic module number, inverter model, etc.) and a phase identifier (such as date, time, season, etc.); all features obtained by multi-scale decomposition are traversed, and the attributes and correlations of each feature are analyzed; based on the properties of the features (such as outliers, redundancy, irrelevant to the problem, etc.), a preliminary excluded feature group, i.e., the first excluded feature group, is determined, which may not have a positive impact on problem location in subsequent analysis , so they can be excluded first; for the features that are not excluded, calculate their response values, which can be direct measurement values, organize the feature response values ​​into vector form to obtain vector response values, and the vector response values ​​correspond to the performance of the features at specific time points and conditions; analyze the feature change trend corresponding to the phase identifier, identify the linear relationship related to time, and establish a linear function relationship based on the phase feature trend; map the vector value of the feature response with the linear function relationship based on the phase feature trend, and determine whether there is an abnormality or problem based on the mapping result and the predetermined judgment criteria (such as threshold value, range, etc.); based on the judgment result, generate one-step detection data, and the one-step detection data includes the identification of the problem area.

[0039] Further, mapping the relationship between the vector value and the linear function and making a determination includes:

[0040] Perform an over-limit judgment on the vector value, perform a trend deviation judgment on the linear function relationship, and determine one judgment data; determine whether the vector value conforms to the linear function relationship, and determine two judgment data; based on the one judgment data and the two judgment data, determine the one-step detection data.

[0041] Preferably, each vector value is compared with its corresponding threshold value, which is set based on historical data, industry standards or technical specifications, and is used to determine whether the current vector value is within a normal range; based on the result of the over-limit judgment, if the vector value exceeds or falls below the threshold value, the vector value is considered abnormal and is marked as a judgment data; it is determined whether the current actual linear relationship conforms to the standard linear relationship, and at the same time, it is determined whether the actual linear relationship is consistent with the actual response, otherwise, there may be a fault that makes the response inconsistent with the expectations of the actual operation, such as frequent fluctuations in data and too discrete data in the same photovoltaic scenario, that is, unstable control. Based on the result of the trend bias judgment, if the trend of the linear function relationship obviously deviates from the expected or normal range, it is considered that there is a trend abnormality and it is used as two judgment data; one judgment data and two judgment data are integrated to form one-step detection data.

[0042] Determine the redundant configuration of the target photovoltaic power station, perform fault tolerance detection of photovoltaic equipment, and determine third detection data, wherein the redundant configuration includes structural redundancy and system redundancy.

[0043] Redundant configuration aims to enhance the fault tolerance of photovoltaic power stations by adding additional equipment or systems, so that a certain power generation capacity and operating efficiency can be maintained when some equipment fails. By determining the redundant configuration of the target photovoltaic power station, the fault tolerance detection of photovoltaic equipment is carried out to determine the third detection data. Redundant configuration includes structural redundancy and system redundancy. Among them, structural redundancy includes organizational redundancy, cable redundancy, etc. Component redundancy refers to the installation of additional photovoltaic components in the photovoltaic array so that when some components fail due to shading, damage or aging, other components can continue to generate electricity; cable redundancy refers to the laying of additional cable lines so that when the main line fails, it can quickly switch to the backup line to reduce the power outage time; system redundancy includes inverter redundancy, energy storage system redundancy, etc. Inverter redundancy refers to the configuration of multiple inverters and the realization of load distribution and fault switching between them to ensure that the power generation capacity of the entire photovoltaic power station will not be affected when a single inverter fails; energy storage system redundancy refers to the installation of additional energy storage equipment (such as battery packs) to provide additional power support during grid failures or peak demand.

[0044] Furthermore, the fault-tolerant detection of the photovoltaic device includes:

[0045] Based on the redundant configuration, in combination with the mapped device failure of the photovoltaic equipment, redundant docking points are determined, wherein the redundant docking points are multiple sequences characterized by redundant configuration features-device failure points; historical redundant compensation records of a preset time zone are read to determine the frequency of failure occurrence and the frequency of redundant configuration, and multiple groups of redundant fault tolerance rates are determined, wherein the multiple groups of redundant fault tolerance rates correspond one-to-one to the redundant docking points, and the frequency ratio is used as a fault tolerance calculation method; the multiple groups of redundant fault tolerance rates are integrated, and the overall fault tolerance rate is weightedly calculated to determine the third detection data.

[0046] Based on the determined redundant configuration, combined with the mapped device failures of the photovoltaic equipment, the redundant docking points can be determined. These docking points are actually the association sequence between the redundant configuration features and the device failure points. They characterize which redundant configurations can play a role in compensating for the losses caused by the failure in the case of a specific device failure. For example, if a photovoltaic module fails, the associated redundant docking points may include spare photovoltaic modules, additional inverter capacity, or supplementary power supply from the energy storage system. In order to evaluate the actual effect of the redundant configuration, it is necessary to read the historical redundant compensation records within a preset time zone (such as the past year, quarter, or month). These records should record in detail how the redundant configuration is triggered and compensates for the impact of each device failure. By analyzing these records, the frequency of failures (i.e., the number of times a specific device or system fails within a certain period of time) and the frequency of redundant configurations (i.e., the number of times the redundant configuration is triggered to compensate for the impact of the failure) can be determined. Based on the frequency of failures and the frequency of redundant configurations, the redundant fault tolerance rate can be calculated for each redundant docking point. The redundant fault tolerance rate reflects the proportion of the redundant configuration that successfully compensates for the impact of a specific device failure. The redundant fault tolerance rate can be calculated using the frequency ratio method, that is, redundant fault tolerance rate = redundant configuration frequency / number of failures. For each redundant docking point, a corresponding redundant fault tolerance rate can be calculated to form multiple sets of redundant fault tolerance rate data. Since there may be multiple redundant docking points in a photovoltaic power station, it is necessary to integrate these multiple sets of redundant fault tolerance rates and obtain the overall redundant fault tolerance rate through weighted calculation, that is, the third detection data.

[0047] The first detection data, the second detection data, and the third detection data are used as a device detection result.

[0048] The first detection data, the second detection data and the third detection data are integrated, and the integrated data is used as the device detection result.

[0049] In summary, the embodiments of the present application have at least the following technical effects:

[0050] First, based on the photoelectric effect and non-invasive detection, data-driven modeling is carried out to build a cascade detection module, which includes a continuous detection unit and a trigger detection unit. Then, the photoelectric effect data of the target photovoltaic power station is continuously monitored and transmitted back to the continuous detection unit, and the analysis based on the photoelectric conversion efficiency is carried out by combining dynamic and static methods to determine the first detection data. Then, the first detection data is judged to be over-limit, and a trigger detection instruction is generated, which is the start instruction for non-invasive detection. Next, the non-invasive detection data is read, and the data is multi-scale decomposition and topological exclusion detection positioning are carried out in combination with the trigger detection unit to determine the second detection data, wherein the detection positioning is based on the direct response data and the linear function relationship. At the same time, the redundant configuration of the target photovoltaic power station is determined, and the fault-tolerant detection of the photovoltaic equipment is carried out to determine the third detection data, wherein the redundant configuration includes structural redundancy and system redundancy. Finally, the first detection data, the second detection data and the third detection data are used as the equipment detection result. The technical problem of low detection efficiency and accuracy of photovoltaic equipment in the prior art is solved, and the technical effect of improving the detection efficiency and accuracy of photovoltaic equipment is achieved by building a cascade detection module.

[0051] Embodiment 2 is based on the same inventive concept as the intelligent detection method for photovoltaic equipment in the above embodiments. Figure 2 As shown, the present application provides an intelligent detection device for photovoltaic equipment, wherein the device includes:

[0052] A cascade detection module building component 11, the cascade detection module building component 11 is used to perform data-driven modeling based on photoelectric effect and non-invasive detection, and build a cascade detection module, the cascade detection module includes a continuous detection unit and a trigger detection unit; a data analysis component 12, the data analysis component 12 is used to continuously monitor the photoelectric effect data of the target photovoltaic power station and transmit it back to the continuous detection unit, and perform a dynamic and static analysis based on the photoelectric conversion efficiency to determine the first detection data; an over-limit judgment component 13, the over-limit judgment component 13 is used to make an over-limit judgment on the first detection data and generate a trigger detection instruction, the trigger detection instruction is a start instruction for non-invasive detection; detection A positioning component 14, wherein the detection and positioning component 14 is used to read non-invasive detection data, and to perform multi-scale decomposition of data and topological exclusion detection and positioning in combination with the trigger detection unit to determine second detection data, wherein detection and positioning are performed based on the direct response data and the linear function relationship; a fault-tolerant detection component 15, wherein the fault-tolerant detection component 15 is used to determine the redundant configuration of the target photovoltaic power station, perform fault-tolerant detection of photovoltaic equipment, and determine third detection data, wherein the redundant configuration includes structural redundancy and system redundancy; a detection result generation component 16, wherein the detection result generation component 16 is used to use the first detection data, the second detection data and the third detection data as the device detection result.

[0053] Furthermore, the cascade detection module building component 11 is used to perform the following method:

[0054] Obtain the photoelectric conversion efficiency of the target photovoltaic power station; build a photovoltaic decision tree based on the photoelectric conversion efficiency and in combination with historical photovoltaic record data, wherein the photovoltaic decision tree executes photovoltaic effect decisions based on photovoltaic scenarios; construct a continuous detection unit based on the photovoltaic decision tree; parallelize the continuous detection unit and the trigger detection unit, and perform data-driven training to generate the cascade detection module.

[0055] Furthermore, the cascade detection module building component 11 is used to perform the following method:

[0056] The equipment installed capacity of the target photovoltaic power station is read to construct a photovoltaic equipment topology; a data acquisition method for photovoltaic detection is determined, data tags are decomposed at multiple scales, and a multi-scale feature pyramid is constructed; a trigger detection unit is constructed by combining the photovoltaic equipment topology with the multi-scale feature pyramid.

[0057] Furthermore, the data analysis component 12 is used to perform the following method:

[0058] Acquire the real-time photovoltaic environment, and determine the first photoelectric effect data based on the photovoltaic conversion efficiency in combination with the photovoltaic decision tree; transmit back the second photoelectric effect data of the target photovoltaic power station, calibrate the first photoelectric effect data, and determine the static detection data; measure the photovoltaic effect trend with photovoltaic environment variables, and determine the dynamic detection data by calibrating the photovoltaic effect trend; determine the first detection data based on the static detection data and the dynamic detection data.

[0059] Furthermore, the detection and positioning component 14 is used to perform the following method:

[0060] Determine a multi-scale decomposition feature, wherein the multi-scale decomposition feature has a device tag identifier and a time phase identifier; traverse the multi-scale decomposition feature to determine a first exclusion feature group; based on the first exclusion feature group, determine a vector value of a feature response, and determine a linear function relationship based on a time phase feature trend; map the vector value and the linear function relationship and make a judgment to determine a step of detection data.

[0061] Furthermore, the detection and positioning component 14 is used to perform the following method:

[0062] Perform an over-limit judgment on the vector value, perform a trend deviation judgment on the linear function relationship, and determine one judgment data; determine whether the vector value conforms to the linear function relationship, and determine two judgment data; based on the one judgment data and the two judgment data, determine the one-step detection data.

[0063] Furthermore, the fault-tolerance detection component 15 is used to perform the following method:

[0064] Based on the redundant configuration, in combination with the mapped device failure of the photovoltaic equipment, redundant docking points are determined, wherein the redundant docking points are multiple sequences characterized by redundant configuration features-device failure points; historical redundant compensation records of a preset time zone are read to determine the frequency of failure occurrence and the frequency of redundant configuration, and multiple groups of redundant fault tolerance rates are determined, wherein the multiple groups of redundant fault tolerance rates correspond one-to-one to the redundant docking points, and the frequency ratio is used as a fault tolerance calculation method; the multiple groups of redundant fault tolerance rates are integrated, and the overall fault tolerance rate is weightedly calculated to determine the third detection data.

[0065] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. And the above-mentioned specific embodiments of this specification are described. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0066] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

[0067] This specification and drawings are merely exemplary illustrations of the present application and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application intends to include these modifications and variations.

Claims

1. An intelligent detection method for photovoltaic equipment, characterized in that: The method comprises: Based on the photoelectric effect and non-invasive detection, data-driven modeling is carried out to build a cascade detection module, which includes a continuous detection unit and a trigger detection unit; Continuously monitor the photoelectric effect data of the target photovoltaic power station and transmit it back to the continuous detection unit, and perform a dynamic and static analysis based on the photoelectric conversion efficiency to determine the first detection data; Performing an over-limit determination on the first detection data and generating a trigger detection instruction, wherein the trigger detection instruction is a start instruction for performing non-intrusive detection; Reading non-intrusive detection data, combining the trigger detection unit to perform multi-scale data decomposition and topology exclusion detection positioning, and determining second detection data, wherein the detection positioning is performed based on the direct response data and the linear function relationship; Determine the redundant configuration of the target photovoltaic power station, perform fault tolerance detection of photovoltaic equipment, and determine third detection data, wherein the redundant configuration includes structural redundancy and system redundancy; Based on the first detection data, the second detection data and the third detection data, as a device detection result; The cascade detection module is constructed, comprising: Obtaining the photoelectric conversion efficiency of the target photovoltaic power station; Based on the photoelectric conversion efficiency and in combination with historical photovoltaic record data, a photovoltaic decision tree is constructed, wherein the photovoltaic decision tree performs photovoltaic effect decisions based on photovoltaic scenarios; based on the photovoltaic decision tree, a continuous detection unit is constructed; The continuous detection unit and the trigger detection unit are operated in parallel, and data-driven training is performed to generate the cascade detection module; The determining of the first detection data comprises: Acquire the real-time photovoltaic environment, and determine the first photoelectric effect data based on the photovoltaic conversion efficiency in combination with the photovoltaic decision tree; Transmitting back the second photoelectric effect data of the target photovoltaic power station, calibrating the first photoelectric effect data, and determining static detection data; The photovoltaic effect trend is measured by photovoltaic environmental variables, and the dynamic detection data is determined by calibrating the photovoltaic effect trend; Determine the first detection data based on the static detection data and the dynamic detection data; Perform topological exclusion detection and positioning, including: Determine a multi-scale decomposition feature, wherein the multi-scale decomposition feature includes a device tag identifier and a time phase identifier; Traversing the multi-scale decomposition features to determine a first exclusion feature group; Based on the first excluded feature group, determining a vector value of a feature response, and determining a linear function relationship based on a temporal feature trend; Mapping the vector value and the linear function relationship and making a judgment to determine a step of detection data; The mapping the relationship between the vector value and the linear function and making a determination includes: Performing an over-limit determination on the vector value, performing a trend deviation determination on the linear function relationship, and determining a determination data; Determine whether the vector value conforms to the linear function relationship, and determine two determination data; Determine the one-step detection data based on the one determination data and the two determination data; The fault-tolerant detection of photovoltaic equipment comprises: Determining redundant docking points based on the redundant configuration and in combination with mapped device failures of the photovoltaic device, wherein the redundant docking points are multiple sequences characterized as redundant configuration features-device failure points; Read the historical redundancy compensation record of the preset time zone, determine the frequency of fault occurrence and the frequency of redundant configuration, and determine multiple groups of redundant fault tolerance rates, wherein the multiple groups of redundant fault tolerance rates correspond to the redundant docking points one by one, and the frequency ratio is used as the fault tolerance calculation method; The multiple groups of redundant fault tolerance rates are integrated, and the overall fault tolerance rate is weightedly calculated to determine the third detection data.

2. The intelligent detection method for photovoltaic equipment according to claim 1, characterized in that: Constructing the trigger detection unit includes: Read the equipment installed capacity of the target photovoltaic power station and construct a photovoltaic equipment topology; Determine the data acquisition method for photovoltaic detection, perform multi-scale decomposition of data labels, and build a multi-scale feature pyramid; The photovoltaic device topology and the multi-scale feature pyramid are combined to construct a trigger detection unit.

3. An intelligent detection device for photovoltaic equipment, characterized in that: For implementing the intelligent detection method for photovoltaic equipment according to any one of claims 1 to 2, the device comprises: A cascade detection module building component, wherein the cascade detection module building component is used to perform data-driven modeling based on photoelectric effect and non-invasive detection to build a cascade detection module, wherein the cascade detection module includes a continuous detection unit and a trigger detection unit; A data analysis component, the data analysis component is used to continuously monitor the photoelectric effect data of the target photovoltaic power station and transmit it back to the continuous detection unit, and perform a dynamic and static analysis based on the photoelectric conversion efficiency to determine the first detection data; an over-limit determination component, the over-limit determination component being used to perform an over-limit determination on the first detection data and generate a trigger detection instruction, the trigger detection instruction being a start instruction for performing non-intrusive detection; A detection and positioning component, the detection and positioning component is used to read non-intrusive detection data, and perform multi-scale data decomposition and topological exclusion detection and positioning in combination with the trigger detection unit to determine second detection data, wherein the detection and positioning is performed based on a direct response data and a linear function relationship; A fault-tolerant detection component, the fault-tolerant detection component is used to determine the redundant configuration of the target photovoltaic power station, perform fault-tolerant detection of photovoltaic equipment, and determine third detection data, wherein the redundant configuration includes structural redundancy and system redundancy; A detection result generating component is used to generate a device detection result based on the first detection data, the second detection data and the third detection data.

Images