Distributed photovoltaic operation and maintenance data intelligent processing method and system

Abstract

The invention relates to the technical field of photovoltaic power generation operation, and discloses a distributed photovoltaic operation and maintenance data intelligent processing method and system, and the method comprises the steps: analyzing common fault types and influence factors of main equipment of a photovoltaic power station; on the equipment level, an intelligent fault diagnosis method is provided for operation characteristics of a transformer and an inverter; on a power station level, an intelligent fault diagnosis process is provided by combining fault analysis of a transformer, an inverter and a photovoltaic array; and constructing a wind power intelligent operation and maintenance system model framework taking the knowledge graph as a core. The beneficial effects of the invention are that through real-time acquisition and wireless transmission of key parameters, accurate monitoring of the operation state of the photovoltaic system is ensured, faults can be quickly found and positioned, the response time is shortened, and the operation efficiency of a power station is improved. Based on a knowledge graph, fault tree analysis and Bayesian reasoning, intelligent fault diagnosis of photovoltaic power station equipment is realized, a complex fault relationship can be automatically deduced, and accurate fault causes and solutions are provided.

Description

Technical Field

[0001] The present invention relates to the technical field of digitalization of photovoltaic power stations, and in particular to a distributed photovoltaic operation and maintenance data intelligent processing method and system. Background Art

[0002] The burning of fossil fuels emits carbon dioxide, which intensifies the greenhouse effect, causing global warming and triggering a climate and environmental crisis. Governments of all countries have formulated relevant carbon emission policies to solve climate problems. Solar energy is an inexhaustible renewable energy source for mankind. It has the advantages of full cleanliness, absolute safety, relative universality, long life and maintenance-free, sufficient resources and potential economy. As the first choice of renewable energy, solar photovoltaic has developed rapidly in recent years.

[0003] At present, although photovoltaic power stations have introduced a variety of information management systems to improve their business level, such as photovoltaic monitoring systems, the operation and maintenance of power stations still need to be handled manually, and it is impossible to fully record the operation and maintenance process of distributed photovoltaic power stations. In addition, due to the large number of types and numbers of power generation equipment and the lack of effective means and measures, operation and maintenance personnel can only deal with faults passively and cannot perform timely operation and maintenance. For a large number of repeated equipment operation and maintenance problems, operation and maintenance personnel will waste a lot of time and energy, resulting in low efficiency of operation and maintenance work. In addition, in order to ensure the safe and reliable operation of photovoltaic power generation equipment, operation and maintenance personnel are often required to be on duty 24 hours a day to deal with emergencies. Coupled with the professional level of operation and maintenance personnel and the geographical distribution of power stations, problems such as high operation and maintenance costs, low efficiency and complicated processes result.

[0004] The current operation and maintenance level and technology of photovoltaic power stations can no longer meet the urgent needs of photovoltaic power stations, which directly affects the operational stability and power generation of photovoltaic power stations, and the operating profit of power stations cannot be reliably guaranteed. Therefore, it is urgent to research and develop intelligent monitoring and intelligent operation and maintenance systems for distributed photovoltaic power stations to achieve centralized, intensive and intelligent operation and maintenance management of many distributed photovoltaic power stations, which is of great significance to promoting the intelligent development of photovoltaic power stations. Summary of the invention

[0005] In view of the above-mentioned problems, the present invention is proposed.

[0006] Therefore, the technical problem solved by the present invention is: how to achieve centralized, intensive and intelligent operation and maintenance management of numerous distributed photovoltaic power stations.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions: a distributed photovoltaic operation and maintenance data intelligent processing method, comprising:

[0008] Collect data on photovoltaic power station equipment;

[0009] Analyze the operating characteristics of the first device and establish a first processing model for prediction;

[0010] Based on the failure analysis of the first device, a second processing flow is established for evaluation;

[0011] Combined with prediction and evaluation results, dynamic monitoring of equipment and intelligent decision support are provided.

[0012] As a preferred solution of the distributed photovoltaic operation and maintenance data intelligent processing method described in the present invention, the collection of photovoltaic power station equipment data includes performing operating condition detection and diagnosis on the equipment to obtain the photovoltaic power station equipment data.

[0013] As a preferred solution of the distributed photovoltaic operation and maintenance data intelligent processing method described in the present invention, wherein: the collection of photovoltaic power station equipment data also includes screening the operation data of the photovoltaic power station equipment, pre-processing the data, and obtaining first equipment data.

[0014] As a preferred solution of the distributed photovoltaic operation and maintenance data intelligent processing method described in the present invention, the analysis of the operating characteristics of the first device includes testing the first device, obtaining test parameters, and processing to obtain first processing model parameters.

[0015] As a preferred solution of the distributed photovoltaic operation and maintenance data intelligent processing method described in the present invention, wherein: the establishment of the first processing model for prediction includes analyzing the operating performance of the photovoltaic system and constructing a photovoltaic output power probability model based on the model parameters of the first processing model.

[0016] As a preferred solution of the distributed photovoltaic operation and maintenance data intelligent processing method described in the present invention, wherein: the second processing flow is evaluated including, using the extreme learning machine as the base learner, establishing an integrated extreme learning machine photovoltaic output power range prediction model; obtaining the efficiency loss model parameters through the fault data of the first device and the photovoltaic power station efficiency loss test, and using the photovoltaic power station power generation performance evaluation tool to evaluate the medium and long-term power generation performance of the power station.

[0017] As a preferred solution of the distributed photovoltaic operation and maintenance data intelligent processing method described in the present invention, the dynamic monitoring of the equipment and intelligent decision support include introducing a correlation coefficient matrix method based on the photovoltaic output power interval prediction model of the integrated extreme learning machine, screening the numerous meteorological factors that affect the photovoltaic output power, retaining the factors with high correlation as input variables, and improving the prediction accuracy while simplifying the model;

[0018] By utilizing the characteristics of online sequential learning machine batch processing data and rolling update output weights, a photovoltaic online interval prediction model is established to achieve dynamic prediction of distributed photovoltaic output power.

[0019] As a preferred solution of the distributed photovoltaic operation and maintenance data intelligent processing system described in the present invention, wherein: the power generation monitoring module collects and monitors the key data of the distributed photovoltaic power generation system in real time;

[0020] Intelligent diagnosis module, a photovoltaic equipment fault diagnosis and maintenance system based on knowledge graph, which conducts intelligent diagnosis of photovoltaic system equipment by collecting equipment operation data, alarm information, and fault tree analysis;

[0021] Energy efficiency evaluation module, providing accurate distributed photovoltaic power generation power prediction and system energy efficiency evaluation;

[0022] The digital management and control platform module realizes digital management and control of the entire process of photovoltaic power stations through the Internet of Things, big data and cloud computing technologies.

[0023] A computer device comprises: a memory and a processor; the memory stores a computer program, comprising: the steps of implementing any one of the methods of the present invention when the processor executes the computer program.

[0024] A computer-readable storage medium stores a computer program, comprising: when the computer program is executed by a processor, the steps of implementing any one of the methods of the present invention are implemented.

[0025] Beneficial effects of the present invention: Through real-time collection and wireless transmission of key parameters, accurate monitoring of the operating status of the photovoltaic system is ensured, faults can be quickly discovered and located, response time can be shortened, and the operating efficiency of the power station can be improved. Based on knowledge graphs, fault tree analysis, and Bayesian reasoning, intelligent fault diagnosis of photovoltaic power station equipment is realized, which can automatically infer complex fault relationships and provide accurate fault causes and solutions. It can better cope with the randomness and uncertainty of photovoltaic power, achieve high-precision power prediction, and help managers optimize resource allocation in advance. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing 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. Among them:

[0027] Figure 1 An overall flow chart of a distributed photovoltaic operation and maintenance data intelligent processing method provided for the first embodiment of the present invention.

[0028] Figure 2This is a diagram showing an example of fault diagnosis of a photovoltaic power station according to a distributed photovoltaic operation and maintenance data intelligent processing method provided in the fourth embodiment of the present invention. DETAILED DESCRIPTION

[0029] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0030] Example 1, reference Figure 1 , is an embodiment of the present invention, and provides a distributed photovoltaic operation and maintenance data intelligent processing method, comprising:

[0031] S1: Collect PV power station equipment data.

[0032] Furthermore, before collecting the PV power station equipment data, the common failure types and influencing factors of the main equipment of the PV power station are analyzed. Based on the analysis of the operating performance of the PV system, the key links of the PV system with problems are tested and collected on-site to obtain the operating data and fault data of the PV power station equipment.

[0033] Furthermore, the operating characteristics of the transformer and inverter include obtaining component efficiency, inconsistency, component attenuation, photovoltaic array attenuation model parameter information and photovoltaic inverter efficiency testing through on-site testing of photovoltaic components and photovoltaic arrays to obtain photovoltaic inverter model parameters.

[0034] Furthermore, the analysis of common fault types and influencing factors of main equipment in a photovoltaic power station may be: at the equipment level, corresponding intelligent fault diagnosis methods are proposed for the operating characteristics of transformers and inverters respectively.

[0035] Furthermore, the analysis of common fault types and influencing factors of the main equipment of the photovoltaic power station can also be, at the power station level, combining the fault analysis of the transformer, inverter and photovoltaic array to propose a complete intelligent fault diagnosis process.

[0036] It should be noted that the efficiency loss model parameters are obtained through the photovoltaic power station efficiency loss test, and the photovoltaic power station power generation performance evaluation tool is used to evaluate the medium and long-term power generation performance of the power station.

[0037] S2: Analyze the operating characteristics of the first device and establish a first processing model for prediction.

[0038] Furthermore, the first equipment includes, but is not limited to, inverters, transformers, photovoltaic modules, combiner boxes, photovoltaic arrays and other key equipment of photovoltaic power stations.

[0039] The first device data may be operating data (such as voltage, current, power, temperature, etc.) and fault data (such as fault code, abnormal log, alarm signal, etc.). Depending on the type of device, the specific monitoring data may vary.

[0040] The analysis of the operating characteristics of the first device includes but is not limited to the following steps: Step 1: Performing device testing: Testing the first device to obtain its key operating parameters. Specifically, the following steps are included:

[0041] Component efficiency test: measure the efficiency of the inverter or transformer, and record data such as input voltage, input current, output voltage, and output current. Calculate the conversion efficiency of the device. Component inconsistency: test the inconsistency of component performance among multiple components, including differences in component current and voltage output, reflecting the unbalanced performance of the system. Component attenuation rate: evaluate the attenuation rate of components through long-term monitoring. Temperature and environmental parameters: collect data such as the operating temperature of the equipment, ambient temperature, and humidity. The collected operating data is transmitted to the central database in real time, and the integrity and real-time nature of the data is ensured through the SCADA system or IoT sensor network.

[0042] Step 2, process model parameterization, collect key parameters through testing and long-term operation data monitoring of the equipment: input voltage V in , input current I in , output voltage V out , output current I out , power output P = V out ×I out , component attenuation rate, device temperature, and ambient temperature.

[0043] Step 3: Establish parameterized equations. Use the collected operating data to establish parameterized equations related to equipment performance. For example, for an inverter, an equation for power conversion efficiency can be established based on the input-output relationship:

[0044]

[0045] Influencing factors such as component attenuation rate and temperature are introduced into the model to more accurately simulate the operating characteristics of the equipment.

[0046] In the embodiment of the present application, the first processing model may be an orthogonal series density estimation model. The step of establishing the first processing model for prediction may include the following specific steps: A1 to A3:

[0047] A1, randomness analysis of equipment output power, establish an orthogonal series density estimation model to analyze the fluctuation characteristics of equipment output power. Assuming that the output power P(t) of the photovoltaic array is a random variable, affected by the external environment, its randomness can be described by a probability model. Use historical operation data to train the orthogonal series density estimation model to estimate the power output distribution under different conditions.

[0048] A2, through the collected equipment operation data, construct a probability model of the equipment output power P(t), taking into account its random fluctuations over time. Assume that the probability density of the output power is f P (p), use the orthogonal series expansion formula to approximate its distribution:

[0049]

[0050] Among them, c n is the expansion coefficient, φ n (p) is an orthogonal polynomial. The coefficient c is fitted by historical data. n , and obtain the probability density distribution model of the device power output.

[0051] A3, using the constructed orthogonal series density estimation model, predict the future operating status of the equipment through real-time collected data. Input the currently collected real-time data into the processing model. Calculate through the orthogonal series density estimation model to predict the future output power fluctuation range. Based on the current conditions, the probability distribution of the output is P(t)~f P (p), from which the expected value and fluctuation range of the output power can be obtained.

[0052] In an optional embodiment, the method of establishing the first processing model for prediction can also be to use mathematical methods to calculate the fluctuation range of each electrical physical quantity under normal operating conditions and various types of faults based on the structure of the photovoltaic power station and the physical model of the solar cell. According to the physical characteristics of solar cells, a relationship model between electrical parameters (such as current, voltage, power, etc.) and environmental factors (such as solar irradiance, temperature) is constructed. Commonly used mathematical models include equivalent circuit models of photovoltaic cells, which can describe the electrical behavior of photovoltaic cells under normal operating conditions and fault conditions. Electrical physical quantity equations: Based on the physical model, mathematical expressions for the current, voltage and power of photovoltaic cells under different conditions are derived. The following classic formula can be used:

[0053]

[0054] Where I is the current, V is the voltage, and I ph is the photocurrent, I0 is the saturation current, R s is the series resistance, R sh is the parallel resistance, n is the diode factor, V tIt is the thermal voltage. By collecting the current, voltage and loss values ​​of the photovoltaic array in real time and observing which range they fall in, the fault type can be determined.

[0055] It should be noted that the intelligent fault diagnosis method is used to establish a photovoltaic output power probability model based on the orthogonal series density estimation theory for the randomness of the output power of the distributed photovoltaic power generation system. The algorithm does not rely on any assumptions about parameter distribution and does not need to establish a physical model of photovoltaic power generation, which reduces the deviation caused by subjectivity. The diagnostic effect of the intelligent diagnosis method depends on multiple factors, including whether the amount of data for model training is large enough, whether the input feature selection of the model is appropriate, and whether the model selection is appropriate. Under the premise of meeting the above conditions, the intelligent diagnosis method can often achieve better diagnostic results and has strong universality in terms of method.

[0056] Embodiment 2: The following is an embodiment of the present invention, which provides a distributed photovoltaic operation and maintenance data intelligent processing method, including:

[0057] S3: Based on the fault analysis of the first device, a second processing flow is established for evaluation.

[0058] In an embodiment of the present application, the second processing flow for evaluation may be established by using an extreme learning machine to directly give a prediction interval of distributed photovoltaic output power, and optimizing the output weight of the extreme learning machine model through a particle swarm algorithm. Using the extreme learning machine as the base learner, an integrated extreme learning machine photovoltaic output power interval prediction model is established. Through the fault data of the first device and the efficiency loss test of the photovoltaic power station, the efficiency loss model parameters are obtained, and the photovoltaic power station power generation performance evaluation tool is used to evaluate the medium and long-term power generation performance of the power station. Specifically including:

[0059] B1, using the extreme learning machine, directly gives the prediction interval of distributed photovoltaic output power, and optimizes the output weight of the extreme learning machine model through the particle swarm algorithm.

[0060] B2, the feature vector of the input layer includes the collected first device fault data, power output, and environmental parameters of the photovoltaic power station (such as irradiance, temperature, etc.). The randomly generated hidden layer weights do not need to be trained, and nonlinear features are mapped through a simple activation function. The goal of the output layer is to predict the output power range of the distributed photovoltaic system, that is, the upper and lower limits of the power output. ELM output formula:

[0061]

[0062] Among them, w i is the weight from the input layer to the hidden layer, β i is the weight from the hidden layer to the output layer, g(·) is the activation function, b iis the bias term, x is the input feature

[0063] B3, the power generation efficiency of photovoltaic power stations will gradually decrease over time, mainly affected by factors such as component attenuation and inverter performance degradation. Through efficiency loss testing, the efficiency loss of each device in the photovoltaic power station is evaluated. Using historical operating data (such as component attenuation, inverter output efficiency, etc.), an efficiency loss model for the equipment is established. The model estimates future power generation efficiency losses based on factors such as operating time, temperature, and irradiance. For example, the efficiency decline trend of components or inverters is predicted through linear regression or exponential decay models.

[0064] B4, using the photovoltaic power station power generation performance evaluation tool, the medium and long-term power generation performance of the power station can be evaluated. The evaluation tool comprehensively considers factors such as equipment operation status, failure rate, efficiency loss, etc., and gives the long-term power generation capacity and performance indicators of the power station. Input the parameters of the efficiency loss model (such as attenuation rate, inverter conversion efficiency, etc.) into the evaluation tool. The long-term operation of the power station is predicted through simulation software, and the power generation and equipment health status in the future period are output.

[0065] In an optional embodiment, establishing the second processing flow for evaluation may also specifically include C1 to C4:

[0066] C1, collect real-time operation data from various devices in the photovoltaic power station, such as inverters, transformers, photovoltaic modules, etc., including: electrical parameters such as voltage, current, power output, and environmental data. Fault log, including historical fault type, fault occurrence time, fault recovery time and other information. Remove noise, outliers and missing data. Use interpolation or mean substitution methods to fill missing values. Normalize or standardize data such as voltage, current, and temperature to eliminate the influence of different dimensions and ensure that the data has the same scale. Split the data into training set, validation set, and test set in chronological order, and split them in a ratio of 70:15:15.

[0067] C2, Model selection and construction, C2.1, Construct a convolutional neural network for fault identification, input multi-dimensional time series data of the equipment operation status. The input dimension may be T×FT\times FT×F, where TTT is the number of time steps and FFF is the number of features. Use two-dimensional convolution (2D-CNN) to extract local features in the time series. The convolution layer learns the features under different fault modes through the convolution kernel. After each convolution layer, average pooling is used to reduce the feature dimension and reduce the computational complexity. After multiple convolutions and pooling, the features are flattened into a one-dimensional vector and input into the fully connected layer for fault category classification. Use the Softmax activation function to map the output of the model to different fault categories. The model will give the probability of each fault.

[0068] C2.2, construct a long short-term memory network (LSTM) for power prediction. The input feature is a multi-dimensional vector of a time series. The LSTM layer learns the state changes of the device over a long time span through the memory and forgetting mechanism. Multiple LSTM layers can be stacked to enhance the model's ability to capture complex time dependencies. The output of the LSTM layer is usually flattened and then passed to the fully connected layer to generate a predicted value. The final output is the power output value of the device for a period of time in the future. Use a linear activation function to output continuous values.

[0069] C3, use the test set to make a final evaluation of the model and calculate the performance indicators of the model. For the CNN fault classification model, use the confusion matrix to evaluate the classification performance of different fault categories, identify the samples with classification errors, and further optimize the model for these samples. For the power prediction results of the LSTM model, analyze the difference between the predicted value and the actual value, and calculate the error distribution. Adjust the model structure or introduce more historical data for training.

[0070] C4, fault diagnosis and evaluation, deploys the trained CNN model to the online fault monitoring system. By inputting the equipment operation data in real time, the model can output the fault category and its confidence in real time. When the confidence of a certain type of fault detected by the model exceeds the set threshold, the system issues a fault warning and generates a fault report indicating the possible cause of the fault. Power output prediction, the trained LSTM model is used for power prediction of photovoltaic power stations. The system inputs the current equipment operation status in real time, and the LSTM model outputs the power prediction value for a period of time in the future. According to the fluctuation of the predicted power, the performance degradation or failure risk of the equipment can be warned in advance.

[0071] It should be noted that the Protégé tool is used to build a fault ontology model based on the fault tree, extracted fault feature parameters and probability importance. The event relationship in the fault tree is represented as the object attribute of the cost entity, and the probability importance of the bottom event is represented as a data attribute. Then, based on the relationship between the facts built in the model, the SWRL rules and the Jess reasoning engine are integrated to reason about the ontology and discover the implicit facts between the faults. And it can automatically match the rules according to the input fault facts to realize reasoning and fault diagnosis decision-making. Carry out case analysis of abnormal fault diagnosis of transformers, inverters and photovoltaic arrays, verify the operation and maintenance auxiliary decision-making performance of the domain knowledge graph built based on the ontology, and improve the sharing and reuse rate of operation and maintenance knowledge.

[0072] Embodiment 3, the following is an embodiment of the present invention, which provides a distributed photovoltaic operation and maintenance data intelligent processing method, including:

[0073] S4: Combine prediction and evaluation results to dynamically monitor equipment and provide intelligent decision support.

[0074] In the embodiment of the present application, dynamic monitoring and intelligent decision support for equipment can be established by establishing an intelligent operation and maintenance system model, including proposing a hierarchical clustering algorithm based on k-means, classifying photovoltaic power generation operation scenarios according to various meteorological factors, and reducing the model error caused by direct classification according to weather types in traditional methods. On the basis of the integrated extreme learning machine photovoltaic output power interval prediction model, the correlation coefficient matrix method is introduced to screen the many meteorological factors that affect the photovoltaic output power, retain the factors with high correlation as input variables, and improve the prediction accuracy while simplifying the model.

[0075] Furthermore, by utilizing the characteristics of online sequential learning machine batch processing data and rolling update of output weights, a photovoltaic online interval prediction model is established to realize the dynamic prediction of distributed photovoltaic output power.

[0076] In an optional embodiment, dynamic monitoring of equipment and intelligent decision support can also be performed by cognitive diagnosis of photovoltaic power station operation and maintenance engineers based on the Bayesian psychometric model to further realize intelligent dispatch. On the basis of the fault knowledge graph, a photovoltaic power station operation and maintenance cognitive diagnosis model is constructed based on the concepts of cognitive diagnosis model and cognitive attributes. The Bayesian psychometric model and MCMC algorithm are used to perform cognitive diagnosis on the operation and maintenance engineers to form a cognitive skill map. Finally, the decision tree algorithm is used to learn the dispatch decision tree from the data of the operation and maintenance engineer's cognitive skill map to realize dispatch prediction.

[0077] Furthermore, we will design and develop an intelligent operation and maintenance analysis system for photovoltaic power stations with the following characteristics: by optimizing the fault alarm model of the monitoring and analysis system, we will improve the fault expert database, workflow management, and spare parts management functions; provide hierarchical data transmission based on the alarm model, realize real-time response to alarms and full-cycle tracking of alarms, establish reasonable workflow management, and provide fault location and intelligent alarm functions; establish a complete power station equipment power generation performance evaluation method, accurately evaluate the power generation loss of each key equipment, and realize the comparison of the equipment's own characteristics with various time and space analysis; based on a scalable distributed database storage architecture, massively store the real-time operation data and environmental data of power station equipment, and find ways to improve equipment performance through analysis and mining of these data.

[0078] It should be noted that a wind power intelligent operation and maintenance system framework with knowledge graph as the core is proposed. The research on intelligent operation and maintenance system includes three main research contents: fault diagnosis decision, cognitive diagnosis, and targeted dispatch. Theoretical analysis is carried out on fault tree, ontology technology, Bayesian reasoning, and decision tree algorithm to lay the foundation for the realization of intelligent operation and maintenance of distributed photovoltaic power stations.

[0079] The following provides a distributed photovoltaic operation and maintenance data intelligent processing system, including:

[0080] Power generation monitoring module, which collects and monitors key data of distributed photovoltaic power generation systems in real time;

[0081] Intelligent diagnosis module, a photovoltaic equipment fault diagnosis and maintenance system based on knowledge graph, which conducts intelligent diagnosis of photovoltaic system equipment by collecting equipment operation data, alarm information, and fault tree analysis;

[0082] Energy efficiency evaluation module, providing accurate distributed photovoltaic power generation power prediction and system energy efficiency evaluation;

[0083] The digital management and control platform module realizes digital management and control of the entire process of photovoltaic power stations through the Internet of Things, big data and cloud computing technologies.

[0084] If the above functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0085] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0086] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0087] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, a plurality of steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0088] Example 4, reference Figure 2 , which is an embodiment of the present invention, provides a distributed photovoltaic operation and maintenance data intelligent processing method. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments.

[0089] Fault diagnosis of photovoltaic power stations is the most important part of intelligent operation and maintenance of photovoltaic power stations, including the detection and diagnosis of the operating conditions of photovoltaic arrays, combiner boxes, inverters, transformers, cables and other components. Among them, because photovoltaic arrays are deployed outdoors for a long time, they are affected by high temperature exposure, wind, rain and various extreme weather conditions. The probability of failure is the highest and the types of failures are the most. They are the focus of fault diagnosis. The types of photovoltaic array failures include battery aging, open circuits and short circuits inside the array, hot plate effects of batteries, dust on the surface of the array, shadows, etc. Figure 2 shown.

[0090] Experimental implementation: At the beginning of the experiment, baseline data collection was performed on the operation of the photovoltaic power station. By testing the efficiency of photovoltaic modules and inverters, the efficiency parameters of the equipment under different conditions were recorded, especially when the equipment was gradually attenuated, to obtain its operating characteristics. Then, transformer overheating faults and inverter output voltage abnormality faults were simulated, and the cause of the faults was analyzed through the knowledge graph model. The system recorded the state parameters of each fault in real time, used the extreme learning machine to predict the photovoltaic output power, and optimized the model weights through the particle swarm algorithm.

[0091] The core of the experiment is to dynamically predict the randomness of photovoltaic power in combination with weather conditions. The k-means clustering algorithm is used to group the photovoltaic system operation data under different meteorological conditions to evaluate the system efficiency and failure rate under various operation scenarios. According to the actual equipment operation status observed on site, the system predicts the photovoltaic output power based on the orthogonal series density estimation theory. The experimental data is batch processed by an online sequential learning machine, and the output weight is gradually optimized in a rolling update manner.

[0092] In order to further verify the advantages of the intelligent fault diagnosis system, the data collected on site is input into the fault tree model, and the knowledge graph reasoning engine is used to analyze equipment failures and output fault diagnosis decisions.

[0093] Table 1 Experimental data comparison table

[0094]

[0095] It can be clearly seen from the above experimental data that the use of intelligent fault diagnosis methods has a significant effect on the monitoring and fault identification of photovoltaic power station equipment. During the operation of the transformer and inverter, the system can accurately capture the output voltage and current fluctuations of the equipment and detect faults in a timely manner. For example, by simulating overheating faults in the experiment, the system can prompt through sound and light alarms as soon as the fault occurs, and use the knowledge graph reasoning engine to quickly diagnose the cause of the fault.

[0096] Compared with the traditional equipment detection method, the present invention can better adapt to the randomness of photovoltaic output power through orthogonal series density estimation theory and extreme learning machine model, avoiding the errors caused by relying on physical models in traditional methods. Especially under complex meteorological conditions, different operating scenarios are classified and processed through k-means clustering algorithm, which greatly reduces the errors caused by direct model division due to weather factors in traditional algorithms. For example, at a high temperature of 35℃ and 950W / m 2 Under high irradiance, the power losses of the transformer and inverter were 0.5% and 1.0% respectively, which were significantly higher than the power losses under low temperature conditions. However, the fault diagnosis system adjusted the operation strategy in time to avoid further expansion of the losses.

[0097] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A distributed photovoltaic operation and maintenance data intelligent processing method, characterized in that: include: Collect data on photovoltaic power station equipment; Analyze the operating characteristics of the first device and establish a first processing model for prediction; Based on the failure analysis of the first device, a second processing flow is established for evaluation; Combined with prediction and evaluation results, dynamic monitoring of equipment and intelligent decision support are provided.

2. The distributed photovoltaic operation and maintenance data intelligent processing method according to claim 1, characterized in that: The collecting of photovoltaic power station equipment data includes performing operating condition detection and diagnosis on the equipment to obtain the photovoltaic power station equipment data.

3. The distributed photovoltaic operation and maintenance data intelligent processing method according to claim 2, characterized in that: The collecting of photovoltaic power station equipment data further includes screening the operation data of the photovoltaic power station equipment and preprocessing the data to obtain first equipment data.

4. The distributed photovoltaic operation and maintenance data intelligent processing method according to claim 3, characterized in that: The analyzing the operating characteristics of the first device includes testing the first device, acquiring test parameters, and processing to obtain model parameters of the first processing model.

5. The distributed photovoltaic operation and maintenance data intelligent processing method according to claim 4, characterized in that: The establishing of the first processing model for prediction includes analyzing the operating performance of the photovoltaic system and constructing the first processing model for probability distribution prediction according to the model parameters of the first processing model.

6. The distributed photovoltaic operation and maintenance data intelligent processing method according to claim 5, characterized in that: The evaluation of the second processing flow includes using the extreme learning machine as the base learner to establish an integrated extreme learning machine photovoltaic output power range prediction model; obtaining efficiency loss model parameters through the fault data of the first device and the photovoltaic power station efficiency loss test, and using the photovoltaic power station power generation performance evaluation tool to evaluate the medium and long-term power generation performance of the power station.

7. The distributed photovoltaic operation and maintenance data intelligent processing method according to claim 6, characterized in that: The dynamic monitoring and intelligent decision support for the equipment includes introducing a correlation coefficient matrix method based on an integrated extreme learning machine photovoltaic output power interval prediction model, screening numerous meteorological factors that affect photovoltaic output power, retaining factors with high correlation as input variables, and improving prediction accuracy while simplifying the model; By utilizing the characteristics of online sequential learning machine batch processing data and rolling update output weights, a photovoltaic online interval prediction model is established to achieve dynamic prediction of distributed photovoltaic output power.

8. A distributed photovoltaic operation and maintenance data intelligent processing system using the method according to any one of claims 1 to 7, characterized in that: Power generation monitoring module, which collects and monitors key data of distributed photovoltaic power generation systems in real time; Intelligent diagnosis module, a photovoltaic equipment fault diagnosis and maintenance system based on knowledge graph, which conducts intelligent diagnosis of photovoltaic system equipment by collecting equipment operation data, alarm information, and fault tree analysis; Energy efficiency evaluation module, providing accurate distributed photovoltaic power generation power prediction and system energy efficiency evaluation; The digital management and control platform module realizes digital management and control of the entire process of photovoltaic power stations through the Internet of Things, big data and cloud computing technologies.

9. A computer device comprising: A memory and a processor; the memory stores a computer program, characterized in that: when the processor executes the computer program, the steps of any one of the distributed photovoltaic operation and maintenance data intelligent processing methods according to claims 1-7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of any one of the distributed photovoltaic operation and maintenance data intelligent processing methods according to claims 1-7 are implemented.

Images