A method and system for fault diagnosis of photovoltaic power plants based on multi-source data fusion

Abstract

This invention relates to the field of photovoltaic power plant fault diagnosis technology, and particularly to a photovoltaic power plant fault diagnosis method and system based on multi-source data fusion. The method includes the following steps: a data acquisition step, periodically collecting multi-source monitoring data from the photovoltaic power plant; a data processing and feature extraction step, preprocessing the multi-source monitoring data; a hierarchical collaborative diagnosis step, constructing a hierarchical diagnostic model from the power plant system level to the component branch level based on the topology of the photovoltaic power generation system; and a fault root cause association and output step, based on the fault determination results, associating and fusing feature evidence from different data sources to perform fault root cause reasoning and analysis. This method and system can improve the comprehensiveness and early warning capability of fault diagnosis, enhance the adaptability and intelligence level of the diagnostic model, deepen fault root cause localization and guide precise operation and maintenance, provide precise maintenance strategy guidance for operation and maintenance personnel, reduce operation and maintenance costs, and improve power generation efficiency.

Description

Technical Field

[0001] This invention relates to the field of photovoltaic power plant fault diagnosis technology, and in particular to a photovoltaic power plant fault diagnosis method and system based on multi-source data fusion. Background Technology

[0002] Photovoltaic power generation, as an important form of clean energy, has been widely applied in recent years. With the rapid growth of installed capacity of photovoltaic power plants and the increasing complexity of deployment environments (such as unattended scenarios in deserts and hilly areas), the reliable operation and efficient maintenance of power plant equipment face severe challenges. For example, an effective hierarchical statistical diagnostic method provided in existing technology (publication number CN106100579A) may encounter the following technical problems in practical applications: (1) Single data source: Diagnosis mainly relies on real-time operating power and current data, as well as basic meteorological data. For potential faults or slow performance degradation caused by internal equipment degradation (such as capacitor aging, abnormal IGBT junction temperature), micro-environmental changes (such as local dust accumulation, hot spot effect) or grid-side disturbances (such as voltage sag, harmonics), the detection sensitivity is insufficient, making it difficult to achieve early warning and root cause differentiation.

[0003] (2) Static diagnostic model: The confidence interval is calculated based on the statistical characteristics (mean and standard deviation) of the data in the current sampling period, which is essentially a static threshold diagnosis. Its performance is limited by the representativeness and quality of the selected historical data. It cannot dynamically learn and adaptively adjust according to the historical evolution trend of the equipment health status and the horizontal comparison of similar equipment, which may lead to false alarms or missed alarms under complex and variable operating conditions.

[0004] (3) Limited depth of fault root cause location: The diagnostic process ends at the conclusion that the branch current is too low. Although possible causes are listed (such as component damage or obstruction), there is a lack of steps to further identify and verify these possible causes by using more dimensions of data (such as infrared thermal imaging temperature, insulation resistance, and component IV curve characteristics). Maintenance personnel still need to go to the site to investigate. There is room for improvement in the intelligence and accuracy of the diagnosis.

[0005] Therefore, there is an urgent need for an intelligent photovoltaic power plant fault diagnosis method that can integrate multi-source information, possess self-learning and self-adaptive capabilities, and achieve accurate fault location and root cause analysis, so as to improve the safety and economy of power plant operation. Summary of the Invention

[0006] The main objective of this invention is to overcome the shortcomings of the prior art and provide a method and system for diagnosing photovoltaic power plant faults based on multi-source data fusion.

[0007] The technical solution adopted by this invention to achieve its technical objective is: a photovoltaic power plant fault diagnosis method based on multi-source data fusion, comprising the following steps: S1: Data acquisition step, periodically collecting multi-source monitoring data of the photovoltaic power station; the multi-source monitoring data includes at least: real-time operating electrical data of the photovoltaic power generation system, status parameter data of key photovoltaic equipment, power station environmental monitoring data, and power quality data at the grid connection point; S2: Data processing and feature extraction step, preprocessing the multi-source monitoring data and extracting time-series features and state features for fault diagnosis; S3: Layered collaborative diagnostic steps. Based on the topology of the photovoltaic power generation system, a layered diagnostic model is constructed from the power station system level to the component branch level. Using the time-series features and state features, and integrating the pre-trained power station health baseline model, anomaly detection and fault determination are performed layer by layer. The power station health baseline model is dynamically generated based on historical normal data and incorporates horizontal comparison information of similar equipment. S4: Fault Root Cause Association and Output Step. Based on the fault determination result of step S3, the feature evidence from different data sources is associated and integrated to perform fault root cause reasoning and analysis, and output a diagnostic report containing fault location, type and possible causes.

[0008] Preferably, in step S1: The real-time operating electrical data includes at least the inverter output power, combiner box current, branch current, and voltage; The status parameter data of the key equipment includes at least one or more of the following: temperature of key components inside the inverter, infrared thermal imaging temperature of the components, and insulation resistance value. The power plant environmental monitoring data shall include at least the following data: irradiance, ambient temperature, component backsheet temperature, and dust concentration. The power quality data at the grid connection point includes at least voltage, frequency, and harmonic distortion rate data.

[0009] Preferably, the fusion of the pre-trained power plant health baseline model in step S3 specifically includes: For key parameters such as inverter output power and branch current, a dynamic prediction model for their changes with environmental factors such as irradiance and temperature is established using time series analysis based on their historical normal data as the first baseline. For a group of similar equipment, the distribution of its performance parameters under similar operating conditions is statistically analyzed to form a group performance distribution model as a second baseline. The deviation analysis is performed between the current device parameters and the predicted values ​​of the first baseline, and the group distribution is compared with that of the second baseline. When both deviations exceed the adaptive threshold, it is determined to be abnormal.

[0010] Preferably, the reasoning and analysis of the root cause of the fault in step S4 specifically includes: When a low current is diagnosed in a certain branch, the infrared thermal imaging temperature distribution characteristics, historical IV curve characteristics and insulation resistance values ​​of the corresponding components in that branch are further retrieved. If a local hot spot is detected accompanied by a step-like feature on the IV curve, the associated diagnosis is that the component's bypass diode has failed. If a uniform temperature increase is detected but is below the hot spot threshold, and the dust concentration data is high, the associated diagnosis is that dust accumulation on the component surface is causing performance degradation.

[0011] Preferably, the method further includes: S5: Diagnostic feedback and model optimization steps, feeding back the final fault cause and handling results confirmed by on-site operation and maintenance to the diagnostic system, which is used to correct the adaptive threshold of the power plant health baseline model and optimize the fault root cause association rules.

[0012] This invention also provides a photovoltaic power plant fault diagnosis system based on multi-source data fusion, comprising: The data acquisition module is configured to periodically collect multi-source monitoring data from photovoltaic power plants. The data processing and feature library module is configured to preprocess and extract features from multi-source monitoring data, and store historical feature data. The health baseline modeling module is configured to dynamically establish and maintain health baseline models for all levels of equipment in the power plant based on historical normal characteristic data and by incorporating population comparison information. The hierarchical collaborative diagnostic engine is configured to invoke the health baseline model and, based on the system topology, perform layer-by-layer anomaly detection and fault determination from the system level to the component branch level. The multi-source fusion root cause analysis module is configured to perform fault root cause inference by fusing multi-source feature evidence based on the preliminary judgment results of the diagnostic engine. The human-computer interaction and report output module is configured to display diagnostic results, alarm information and detailed diagnostic analysis reports, and receive operation and maintenance feedback information.

[0013] Preferably, the data acquisition module includes an electrical quantity acquisition unit, a status quantity acquisition unit, an environmental quantity acquisition unit, and a power quality acquisition unit. Each unit is connected to the data processing and feature library module through the power plant monitoring network.

[0014] Preferably, the health baseline modeling module uses machine learning algorithms to model derived features such as inverter efficiency and component performance degradation rate, and has the function of updating model parameters online based on new feedback data.

[0015] The working principle of the photovoltaic power plant fault diagnosis method and system based on multi-source data fusion is as follows: by integrating multi-source heterogeneous data such as electrical quantities, equipment status quantities, environmental quantities and power quality, and constructing a dynamic health baseline model that integrates historical longitudinal trends and equipment horizontal comparisons, intelligent and precise fault diagnosis of photovoltaic power plants from the system to the components is achieved.

[0016] Compared with the prior art, the beneficial effects of the present invention are: This photovoltaic power plant fault diagnosis method and system based on multi-source data fusion can improve the comprehensiveness of fault diagnosis and early warning capability: by integrating electrical quantities, equipment status quantities, environmental quantities and power grid quality data, a more three-dimensional equipment health profile can be constructed, which can capture hidden faults and performance degradation trends that are difficult to reflect by a single data source, and realize the transformation from post-fault diagnosis to pre-fault warning.

[0017] This photovoltaic power plant fault diagnosis method and system based on multi-source data fusion enhances the adaptability and intelligence of the diagnostic model: by introducing time series analysis, machine learning and other methods, the diagnostic model can learn the normal behavior pattern of equipment based on historical data and dynamically update the diagnostic benchmark. Then, through cross-equipment and cross-power plant group comparison, it can identify individual anomalies, reduce misjudgments caused by common factors such as weather and seasons, and improve diagnostic accuracy.

[0018] This photovoltaic power plant fault diagnosis method and system based on multi-source data fusion can deepen the root cause location of faults and guide precise operation and maintenance: on the basis of hierarchical diagnosis, a root cause analysis based on multi-source fusion data is added to provide operation and maintenance personnel with precise maintenance strategy guidance, reduce operation and maintenance costs, and improve power generation efficiency. Attached Figure Description

[0019] Figure 1 This is a flowchart of the steps in a photovoltaic power plant fault diagnosis method based on multi-source data fusion.

[0020] Figure 2 This is a system framework diagram of a photovoltaic power plant fault diagnosis system based on multi-source data fusion. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. However, it should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of the invention. Furthermore, descriptions of well-known structures and technologies are omitted in the following description to avoid unnecessarily obscuring the concept of the invention.

[0022] In the description of this invention, it should be noted that when an element is referred to as being "fixed to" or "set on" another element, it can be directly on or indirectly on the other element. When an element is referred to as being "connected to" another element, it can be directly connected to or indirectly connected to the other element.

[0023] In the description of this invention, it should be noted that the terms "center," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, or the orientation or positional relationship commonly used when the product of this invention is in use. They are used only for the convenience of describing the invention and for 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 the invention. Furthermore, the terms "first," "second," and "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified. "Several" means one or more, unless otherwise explicitly specified.

[0024] In the description of this invention, it should also be noted that, unless otherwise explicitly specified and limited, the terms "set," "install," "connect," and "link" 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 of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances. Example 1:

[0025] Please see Figures 1-2 A method for fault diagnosis of photovoltaic power plants based on multi-source data fusion includes the following steps: S1: Data acquisition step, periodically collecting multi-source monitoring data from the photovoltaic power station; the multi-source monitoring data includes at least: real-time operating electrical data of the photovoltaic power generation system, status parameter data of key photovoltaic equipment, power station environmental monitoring data, and power quality data at the grid connection point. In this step: The real-time operating electrical data includes at least the inverter output power, combiner box current, branch current, and voltage; The status parameter data of the key equipment includes at least one or more of the following: temperature of key components inside the inverter, infrared thermal imaging temperature of the components, and insulation resistance value. The power plant environmental monitoring data shall include at least the following data: irradiance, ambient temperature, component backsheet temperature, and dust concentration. The power quality data at the grid connection point includes at least voltage, frequency, and harmonic distortion rate data.

[0026] S2: Data processing and feature extraction step, preprocessing the multi-source monitoring data and extracting time-series features and state features for fault diagnosis.

[0027] S3: Layered collaborative diagnostic steps, based on the topology of the photovoltaic power generation system, to construct a layered diagnostic model from the power station system level to the component branch level; By utilizing the aforementioned time-series and state features, and integrating them with a pre-trained power plant health baseline model, anomaly detection and fault determination are performed layer by layer. The power plant health baseline model is dynamically generated based on historical normal data and incorporates horizontal comparison information of similar equipment.

[0028] In this step, the fusion of the pre-trained power plant health baseline model specifically includes: For key parameters such as inverter output power and branch current, a dynamic prediction model for their changes with environmental factors such as irradiance and temperature is established using time series analysis based on their historical normal data as the first baseline. For a group of similar equipment, the distribution of its performance parameters under similar operating conditions is statistically analyzed to form a group performance distribution model as a second baseline. The deviation analysis is performed between the current device parameters and the predicted values ​​of the first baseline, and the group distribution is compared with that of the second baseline. When both deviations exceed the adaptive threshold, it is determined to be abnormal.

[0029] S4: Fault Root Cause Association and Output Step. Based on the fault determination result of step S3, the feature evidence from different data sources is associated and integrated to perform fault root cause reasoning and analysis, and output a diagnostic report containing fault location, type and possible causes.

[0030] In this step, the reasoning and analysis of the root cause of the failure specifically includes: When a low current is diagnosed in a certain branch, the infrared thermal imaging temperature distribution characteristics, historical IV curve characteristics and insulation resistance values ​​of the corresponding components in that branch are further retrieved. If a local hot spot is detected accompanied by a step-like feature on the IV curve, the associated diagnosis is that the component's bypass diode has failed. If a uniform temperature increase is detected but is below the hot spot threshold, and the dust concentration data is high, the associated diagnosis is that dust accumulation on the component surface is causing performance degradation.

[0031] S5: Diagnostic feedback and model optimization steps, feeding back the final fault cause and handling results confirmed by on-site operation and maintenance to the diagnostic system, which is used to correct the adaptive threshold of the power plant health baseline model and optimize the fault root cause association rules.

[0032] Specifically, in practice, the system first periodically collects and preprocesses multi-dimensional data such as inverter power, combiner box current, component temperature, irradiance, dust concentration, and grid harmonics, extracting key features. Second, based on the power station's physical topology, a hierarchical collaborative diagnosis is initiated. Each layer compares real-time equipment data with a dynamic health baseline (including individual prediction models based on time series and distribution models based on population statistics). When the deviation exceeds an adaptive threshold, an anomaly is identified, thus converging the fault range layer by layer from top to bottom (system level → inverter level → combiner box level → branch level). Subsequently, for the initially located abnormal branches, a multi-source fusion root cause analysis is initiated, comprehensively retrieving characteristic evidence such as infrared thermal imaging, IV curves, and insulation resistance for correlation reasoning to distinguish between specific causes such as "bypass diode failure" and "surface dust accumulation." Finally, the system outputs a diagnostic report containing precise location and root cause inference, and uses the operation and maintenance feedback results to continuously optimize the health baseline model and diagnostic rules, forming a self-learning diagnostic optimization closed loop. Example 2:

[0033] Please see Figures 1-2 Based on the above embodiments, this invention also provides a photovoltaic power plant fault diagnosis system based on multi-source data fusion, including a data acquisition module, a data processing and feature library module, a health baseline modeling module, a hierarchical collaborative diagnosis engine, a multi-source fusion root cause analysis module, and a human-computer interaction and report output module.

[0034] In this implementation, the data acquisition module is configured to periodically collect multi-source monitoring data from the photovoltaic power station.

[0035] The data acquisition module includes an electrical quantity acquisition unit, a status quantity acquisition unit, an environmental quantity acquisition unit, and a power quality acquisition unit. Each unit is connected to the data processing and feature library module through the power plant monitoring network.

[0036] In this implementation, the data processing and feature library module is configured to preprocess and extract features from multi-source monitoring data, and store historical feature data.

[0037] In this implementation, the health baseline modeling module is configured to dynamically establish and maintain health baseline models for all levels of equipment in the power plant based on historical normal characteristic data and by incorporating population comparison information.

[0038] The health baseline modeling module uses machine learning algorithms to model derived features such as inverter efficiency and component performance degradation rate, and has the function of updating model parameters online based on new feedback data.

[0039] In this implementation, the hierarchical collaborative diagnostic engine is configured to invoke the health baseline model and, based on the system topology, perform layer-by-layer anomaly detection and fault determination from the system level to the component branch level.

[0040] In this implementation, the multi-source fusion root cause analysis module is configured to perform fault root cause inference by fusing multi-source feature evidence based on the preliminary judgment results of the diagnostic engine.

[0041] In this implementation, the human-computer interaction and report output module is configured to display diagnostic results, alarm information, and detailed diagnostic analysis reports, and to receive operation and maintenance feedback information.

[0042] The solution in this embodiment can be selectively combined with solutions in other embodiments.

[0043] The core of this photovoltaic power plant fault diagnosis method and system based on multi-source data fusion lies in constructing a three-layer collaborative intelligent diagnosis architecture of "data layer - model layer - application layer": At the data layer, the system integrates and collects four-dimensional data streams from photovoltaic power plants through the Internet of Things sensor network, including real-time operating electrical data reflecting energy conversion efficiency (inverter power, branch current and voltage), key parameter data characterizing the health status of equipment (IGBT temperature, module infrared thermal temperature, insulation resistance), environmental monitoring data affecting power generation performance (irradiance, backsheet temperature, dust concentration), and power quality data reflecting grid connection safety (voltage harmonics, frequency deviation). At the model layer, the system uses dual baseline modeling technology to establish a dynamic prediction model for individual equipment performance based on time series analysis on the one hand, and to construct a performance reference model for a group of similar equipment based on statistical distribution on the other hand, forming an adaptive health benchmark with environmental adaptability and horizontal comparability. At the application layer, the system achieves an intelligent closed loop from anomaly detection to fault location through a topology-driven hierarchical diagnostic engine and an evidence fusion-based root cause reasoning mechanism.

[0044] Specifically, it manifests as a six-step intelligent diagnostic and optimization closed loop: The first step (data aggregation) involves the acquisition units distributed at various nodes of the power plant periodically uploading multi-source monitoring data to the central processing platform.

[0045] The second step (feature construction) involves the platform cleaning, aligning, and performing feature engineering on the raw data to extract key temporal and statistical features that reflect the equipment status.

[0046] The third step (layered screening) involves the diagnostic engine initiating a three-level progressive diagnosis from top to bottom based on the physical connection topology of the photovoltaic system. First, at the inverter layer, the actual output power of each inverter is compared with its dynamic prediction baseline based on meteorological conditions and the distribution of inverters in the same area to identify units with abnormal power. Next, for the abnormal inverter, a similar dual-benchmark analysis is performed on the output current of its subordinate combiner boxes to locate the abnormal combiner box. Finally, a detailed comparison is performed on the current of each branch under the abnormal combiner box to pinpoint the specific faulty branch.

[0047] The fourth step (root cause analysis) involves the system initiating multi-evidence correlation analysis on the located faulty branch. It automatically retrieves data such as the infrared thermal image of the corresponding component (to detect hot spot distribution), historical IV curves (to identify characteristic distortion), and insulation resistance change trends. Evidence is fused through a rule engine or a lightweight machine learning model to distinguish specific fault types such as "bypass diode breakdown," "partial component shading," or "dust accumulation and performance degradation."

[0048] Step 5 (Decision Output): The system generates a structured diagnostic report, which displays the fault location map, root cause analysis conclusions, and maintenance priority suggestions through a visual interface.

[0049] Step 6 (Feedback Optimization): After on-site maintenance personnel verify and repair the diagnostic results, they feed back the final confirmed fault causes and handling effects to the system. This is used to dynamically calibrate the threshold parameters of the health baseline model and optimize the root cause inference rule base, enabling the diagnostic system to self-evolve and improve accuracy during continuous operation. The entire process achieves end-to-end intelligentization from "massive data collection" to "precise operation and maintenance decision-making," significantly improving the availability and operation and maintenance efficiency of the power plant.

[0050] It should be noted that although the above embodiments have been described herein, this does not limit the scope of patent protection of this invention. Therefore, any changes and modifications made to the embodiments described herein based on the innovative concept of this invention, or equivalent structural, procedural, or functional transformations made using the description and drawings of this invention, directly or indirectly applying the above technical solutions to other related technical fields, are all included within the scope of protection of this invention.

Claims

1. A photovoltaic power station fault diagnosis method based on multi-source data fusion, characterized in that, The method comprises the following steps: S1: a data acquisition step, periodically acquiring multi-source monitoring data of the photovoltaic power station; The multi-source monitoring data at least includes real-time operation electrical data of the photovoltaic power generation system, state parameter data of photovoltaic key equipment, power station environment monitoring data, and grid-connected point power quality data; S2: a data processing and feature extraction step, preprocessing the multi-source monitoring data, and extracting time sequence features and state features for fault diagnosis therefrom; S3: a hierarchical collaborative diagnosis step, based on the topology of the photovoltaic power generation system, constructing a hierarchical diagnosis model from the power station system level to the component branch level; using the time sequence features and state features, and fusing a pre-trained power station health baseline model, performing abnormality detection and fault determination layer by layer; the power station health baseline model is dynamically generated based on historical normal data and by introducing horizontal comparison information of similar equipment; S4: a fault root cause correlation and output step, based on the fault determination result of step S3, correlating and fusing feature evidences from different data sources, reasoning and analyzing the fault root cause, and outputting a diagnosis report containing the fault location, type and possible cause. 2.The photovoltaic power station fault diagnosis method based on multi-source data fusion according to claim 1, characterized in that, In step S1: The real-time operation electrical data at least includes inverter output power, busbar current, branch current and voltage; The state parameter data of the key equipment at least includes one or more of inverter internal key component temperature, component infrared thermal imaging temperature, and insulation resistance value; The power station environment monitoring data at least includes irradiance, environmental temperature, component backboard temperature, and dust concentration data; The grid-connected point power quality data at least includes voltage, frequency, and harmonic distortion rate data. 3.The photovoltaic power station fault diagnosis method based on multi-source data fusion according to claim 1, characterized in that, In step S3, the fusion of the pre-trained power station health baseline model specifically includes: For key parameters such as inverter output power and branch current, based on their historical normal data, a dynamic prediction model of the parameters changing with environmental factors such as irradiance and temperature is established as a first baseline by using time series analysis method; For a group of similar equipment, the performance parameter distribution of the equipment under similar working conditions is counted to form a group performance distribution model as a second baseline; Deviation analysis is performed on the current equipment parameters and the predicted values of the first baseline, and the group distribution of the second baseline is compared, and when the deviations of both exceed the adaptive threshold, it is determined to be abnormal. 4.The photovoltaic power station fault diagnosis method based on multi-source data fusion according to claim 1, characterized in that, In step S4, the reasoning and analysis of the fault root cause specifically includes: When a branch current is diagnosed to be low, the component infrared thermal imaging temperature distribution characteristics, historical IV curve characteristic change trend, and insulation resistance value corresponding to the branch are further retrieved; If local hot spots are detected with IV curve step features, the correlation diagnosis is bypass diode failure of the component; If the temperature uniformly rises but is lower than the hot spot threshold, and the dust concentration data is high, the correlation diagnosis is that the surface area of the component is covered with dust, causing performance degradation. 5.The photovoltaic power station fault diagnosis method based on multi-source data fusion according to claim 1, characterized in that, The method further comprises: S5: a diagnosis feedback and model optimization step, feeding back the final fault cause and treatment result confirmed by on-site operation and maintenance to the diagnosis system, for correcting the adaptive threshold of the power station health baseline model and optimizing the fault root cause correlation rule. 6.A photovoltaic power station fault diagnosis system based on multi-source data fusion for implementing the method of any one of claims 1 to 5, characterized in that, It comprises: The data acquisition module is configured to periodically acquire multi-source monitoring data of the photovoltaic power station. The data processing and feature library module is configured to pre-process and extract features of the multi-source monitoring data, and store historical feature data. The health baseline modeling module is configured to dynamically establish and maintain a health baseline model of each level of equipment of the power station based on historical normal feature data and by introducing group comparison information. The hierarchical collaborative diagnosis engine is configured to call the health baseline model and perform layer-by-layer abnormality detection and fault determination from the system level to the component branch level according to the system topology. The multi-source fusion root cause analysis module is configured to fuse multi-source feature evidence based on the preliminary determination result of the diagnosis engine to perform fault root cause reasoning. The human-computer interaction and report output module is configured to display the diagnosis result, alarm information and detailed diagnosis analysis report, and receive operation and maintenance feedback information.

7. The photovoltaic power station fault diagnosis system based on multi-source data fusion according to claim 6, characterized in that, The data acquisition module includes an electrical quantity acquisition unit, a state quantity acquisition unit, an environmental quantity acquisition unit and a power quality acquisition unit, and each unit is in communication connection with the data processing and feature library module through a power station monitoring network. 8.The photovoltaic power station fault diagnosis system based on multi-source data fusion of claim 6, characterized in that, The health baseline modeling module adopts a machine learning algorithm to model derived features such as inverter efficiency and component performance attenuation rate, and has the function of online updating model parameters based on new feedback data.

Images