A method and system for early warning and diagnosis of faults in intelligent photovoltaic water heaters

By stripping abnormal weather impact and evaluating photoelectric efficiency and electric heating efficiency, combining data smoothing and clustering analysis, the potential faults of photovoltaic water heaters are identified, and the problem of insufficient accuracy and reliability of fault identification in the existing technology is solved, and more accurate The fault warning and maintenance extends the service life of the equipment.

CN118861649BActive Publication Date: 2024-11-29LIAONING HUAHAI TECH DEV CO LTD +2
View PDF 3 Cites 0 Cited by

Patent Information

Application Number
CN202411329004.4
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-09-24
Publication Date
2024-11-29
Estimated Expiration
2044-09-24

AI Technical Summary

Technical Problem

The existing photovoltaic water heater fault warning technology lacks a unified baseline standard, resulting in insufficient accuracy and reliability of fault identification, and the inability to peel off interference from external factors such as abnormal weather, resulting in high misjudgment rate, making it difficult to accurately identify equipment performance degradation and concealment of faults. , increasing maintenance costs and operating risks.

Method used

Through systematic analysis of the operating data of the photovoltaic water heater, the impact of stripping abnormal weather, calculate and evaluate the photoelectric efficiency and electric heating efficiency, determine its baseline qualifications, use data smoothing and clustering analysis to extract the operating trend items of components, and add these trend items Compare with typical failure modes and calculate similarity to identify potential hidden faults and issue early warning signals.

Benefits of technology

Improve the accuracy and reliability of the operating status evaluation of various parts of the photovoltaic water heater, reduce the occurrence of sudden failures, provide more targeted and predictive maintenance, extend the service life of the equipment, and optimize the overall performance. .

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118861649B_ABST
    Figure CN118861649B_ABST
Patent Text Reader

Abstract

The present invention discloses an intelligent photovoltaic water heater fault early warning diagnosis method and system, which specifically relates to the field of photovoltaic water heater fault monitoring, and is used to solve the problem of inaccurate fault identification, by removing the influence of abnormal weather, calculating and evaluating the photoelectric efficiency and the electric thermal efficiency, and determining the baseline qualification. Based on the baseline conditions, data smoothing and cluster analysis are used to extract and mark the operating trend items of the components, and then these trend items are compared with typical fault modes to identify potential hidden faults and issue early warning signals. Ensure accurate evaluation of the operating status of each component in the long-term trend, eliminate interference from changes in the environment and operating conditions through the determination of the baseline, and make the analysis results more comparable and reliable. Make maintenance work more targeted and predictive, reduce sudden failures, provide reliable early warning diagnosis, improve maintenance efficiency and operational reliability, and ultimately optimize the overall performance and service life of the photovoltaic water heater.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of photovoltaic water heater fault monitoring, and more specifically, to an intelligent photovoltaic water heater fault early warning diagnosis method and system. Background Art

[0002] Photovoltaic water heaters are devices that use solar photovoltaic technology to convert sunlight into electrical energy and heat water through electric heating elements. This system captures solar energy through photovoltaic panels, converts it into electrical energy, stores it in batteries or directly uses it to heat water. This water heater can not only provide hot water for homes and commercial places, but can also be used for heating to a certain extent. Photovoltaic water heaters generate electricity through photovoltaic panels on sunny days and directly use it to heat water. On rainy days or at night, they can be connected to the mains to ensure the continuity and stability of hot water supply.

[0003] The existing photovoltaic water heater fault warning and identification technology has some shortcomings, mainly reflected in the lack of a unified baseline standard, which affects the accuracy and reliability of fault identification. Traditional fault warning methods usually rely on the abnormal detection of a single parameter, such as sudden changes in temperature, voltage or current, but fail to take into account changes in environmental conditions and long-term trends in the operating status of the equipment. In addition, existing technologies are often unable to remove the interference of external factors such as abnormal weather on the data, resulting in a high misjudgment rate, which affects the effectiveness of the early warning system. Due to the lack of a comprehensive evaluation of photovoltaic efficiency and electrothermal efficiency and the determination of baseline standards, existing methods are difficult to accurately identify the performance degradation and hidden faults of equipment, resulting in a lack of predictability and pertinence in maintenance work, which increases the operating risk and maintenance cost of the equipment.

[0004] In order to solve the above problems, a technical solution is now provided. Summary of the invention

[0005] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a smart photovoltaic water heater fault warning diagnosis method and system, which systematically analyzes the operating data of the photovoltaic water heater, removes the influence of abnormal weather, calculates and evaluates the photovoltaic efficiency and electric thermal efficiency, and determines its baseline qualification. Based on the baseline conditions, data smoothing and cluster analysis are used to extract and mark the operating trend items of the components, and then these trend items are compared with typical fault modes, and the similarity is calculated to identify potential hidden faults and issue early warning signals. Ensure the accurate evaluation of the operating status of each component of the photovoltaic water heater in the long-term trend, and ensure that all analysis data are carried out under the same standard through the determination of the baseline, eliminate the interference caused by changes in the environment and operating conditions, and make the analysis results more comparable and reliable. Through the screening of the baseline, we can focus on the changes in the performance of the equipment itself without being affected by external factors, so as to more accurately identify the long-term operating trends and potential performance degradation of the components. The extraction and analysis of trend items can identify abnormal changes and potential failures of components in advance, making maintenance work more targeted and predictive, reducing the occurrence of sudden failures, and providing reliable early warning diagnosis, thereby improving the maintenance efficiency and operational reliability of the equipment, and ultimately optimizing the overall performance and service life of the photovoltaic water heater to solve the problems raised in the above background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] Build an operation database of equipment with daily operation time;

[0008] Stripping away the factors that affect abnormal weather and its delayed effects on photovoltaic power generation;

[0009] Calculate the photoelectric efficiency and electrothermal efficiency data based on the data analysis after stripping;

[0010] The electrical and thermal energy output efficiencies of the photovoltaic efficiency and electrothermal efficiency in different time units were evaluated by data envelopment analysis, and the reproducibility index and durability index were calculated to determine whether the photovoltaic efficiency and electrothermal efficiency were qualified as the baseline;

[0011] If the qualification baseline is obtained, based on the photovoltaic efficiency and electrothermal efficiency baseline conditions, the moving average method and smoothing process are used to extract the mean characteristics of the components, and the trend items of each component are identified and marked through cluster analysis;

[0012] The trend items of each component are compared with typical failure modes and performance degradation indicators, and a matching algorithm is used to calculate the similarity to determine whether there are hidden failures or performance degradation problems and to issue early warning signals.

[0013] In a preferred embodiment, the factors affecting the abnormal weather and its delayed effects on photovoltaic power generation are stripped away, including the following:

[0014] Call the data from the equipment operation database and classify the data into meteorological data and photovoltaic system performance data;

[0015] Use variational autoencoders to extract key features from the equipment operation database;

[0016] The key features extracted by the variational autoencoder are input into the isolation forest algorithm for anomaly detection, from which abnormal data points caused by extreme weather events are identified;

[0017] Based on anomaly detection, causal inference technology is used to analyze the causal relationship between confirmed abnormal weather data and photovoltaic system performance. The impact of extreme weather events on photovoltaic power generation efficiency and its lasting consequences are determined through causal graph technology, and data affected by extreme weather and its delayed effects are removed.

[0018] In a preferred embodiment, the photoelectric efficiency and electrothermal efficiency data are calculated based on the data analysis after the stripping process, including the following contents:

[0019] After removing the data marked as affected by extreme weather, the photovoltaic efficiency is independently calculated and adjusted;

[0020] Collecting operation data related to the photovoltaic system, including output electrical energy and incident light energy, and obtaining first environmental data and first device status data that affect the photoelectric efficiency; calculating the light-to-electricity efficiency, that is, the ratio of output electrical energy to incident light energy, and associating the first environmental data and first device status data variables with the light-to-electricity efficiency by constructing a multivariate regression model, analyzing the specific impact of the first environmental data and the first device status data on the light-to-electricity efficiency, and calculating and adjusting to obtain the photoelectric efficiency;

[0021] First environmental data: environmental factors that directly affect the reception and conversion of solar energy by photovoltaic panels;

[0022] First equipment status data: physical status that affects the performance of the photovoltaic panel;

[0023] Collect the output heat energy and input electric energy data of the hot water system, and collect the second environment data and the second equipment status data that affect the electric heat efficiency, calculate the electric heat conversion efficiency, that is, the ratio of the output heat energy to the input electric energy, analyze the influence of the second environment data and the second equipment status data on the electric heat efficiency by setting up a regression model, and calculate and adjust to obtain the electric heat conversion efficiency;

[0024] Second environmental data: environmental factors that affect the efficiency of the hot water system;

[0025] Second, equipment status data: data reflecting the operation and maintenance status of equipment.

[0026] In a preferred embodiment, the electrical and thermal output efficiencies of the photoelectric efficiency and electrothermal efficiency in different time units are evaluated by data envelopment analysis, and the reproducibility index and durability index are calculated to determine whether they are qualified as a baseline, including the following:

[0027] Collect the electric and thermal energy data calculated from the photoelectric efficiency and electrothermal efficiency in different unit times, and standardize the electric and thermal energy data;

[0028] Each time unit is regarded as a decision unit, and the data of each decision unit includes the electric energy calculated by the photoelectric efficiency and the thermal energy calculated by the electrothermal efficiency;

[0029] In the data envelopment analysis model, photoelectric efficiency and electrothermal efficiency are selected as input variables, and electrical energy output and thermal energy output are selected as output variables. The specific variable definitions are as follows:

[0030] Input variable 1 : Photoelectric efficiency;

[0031] Input variable 2 : Electric heating efficiency;

[0032] Output variable 1 : Power output;

[0033] Output variable 2 : Heat output;

[0034] Using the CCR model and assuming that the scale returns remain constant, the linear programming form of the CCR model is as follows:

[0035] ;

[0036] ;

[0037] ;

[0038] in, For the The efficiency value of a decision-making unit, the goal is to maximize the corresponding efficiency value; Indicates The decision-making unit Outputs; Indicates The decision-making unit Inputs; Indicates The weight of the output; Indicates The weight of the input; Indicates A decision making unit (DMU) is in the The output quantity on the output variable; It means the The decision making unit is The amount of input on the input variables;

[0039] is the number of output variables; The number of input variables; Represents the index of a specific output variable; An index representing a specific decision-making unit; Represents the index of a specific input variable; Represents the index of all decision-making units; Subject to refers to the constraints in linear programming.

[0040] In a preferred embodiment, a linear programming solution method is used to solve the linear programming form of the following data envelopment analysis model to obtain the efficiency value of each decision-making unit;

[0041] Extract the efficiency value of each decision-making unit from the data envelopment analysis model;

[0042] Evaluate the stability and reliability of efficiency values ​​over a long period of time, and calculate the reproducibility index and durability index;

[0043] Divide the entire time period into several smaller time periods, calculate the mean of the efficiency value in each time period, compare the means of adjacent time periods in turn, calculate the difference between the means, average the differences between all means, and obtain the reproducibility index;

[0044] Calculate the overall mean of all efficiency values ​​in the entire time period, calculate the difference between the efficiency value at each time point and the overall mean, calculate the sum of the squares of all differences, find the average, take the square root, and get the durability index;

[0045] The reproducibility index and durability index are respectively greater than or equal to the corresponding thresholds, indicating that the photovoltaic efficiency and electrothermal efficiency are not qualified as the baseline;

[0046] The reproducibility index and durability index are respectively less than the corresponding thresholds, indicating that the photoelectric efficiency and electrothermal efficiency are qualified as baselines.

[0047] In a preferred embodiment, if the qualification baseline is obtained, based on the photovoltaic efficiency and electrothermal efficiency baseline conditions, the mean characteristics of the components are extracted using the moving average method and smoothing processing, and the trend items of each component are identified and marked through cluster analysis, including the following:

[0048] S5-1, if the reproducibility index and the durability index are respectively less than the corresponding thresholds, then the operation data of each component under the same photoelectric efficiency and electrothermal efficiency numerical baseline conditions are collected;

[0049] S5-2, set a rated window size, and for each time point, calculate the average value of all data points in the time point and the windows before and after it. The specific calculation process is as follows:

[0050] For time point , take the window size as , then the moving average is the sum of all data points in the window divided by the window size;

[0051] The calculated moving average is used as the smoothing value at the corresponding time point;

[0052] S5-3, for the smoothed time series data of each component, calculate the overall mean, which represents the average operating status of the corresponding component in the entire time period;

[0053] Add the smoothed values ​​of all time points and divide by the number of time points to get the overall mean;

[0054] The calculated overall mean is used as the mean feature of the corresponding parts;

[0055] S5-4, random selection Initial cluster centers are randomly selected from the data set. mean features;

[0056] According to the mean feature, the data point of each component is assigned to the nearest cluster center, the Euclidean distance between each data point and the cluster center is calculated, and the nearest cluster center is selected for assignment;

[0057] Recalculate the center position of each cluster, that is, the mean of all data points belonging to the cluster;

[0058] Repeat the steps of assigning data points and updating cluster centers until the cluster centers no longer change significantly or the preset number of iterations is reached;

[0059] S5-5, mark the trend items of each component according to the mean of the cluster center.

[0060] In a preferred embodiment, the trend items of each component are compared with typical failure modes and performance degradation indicators, and a matching algorithm is used to calculate the similarity to determine whether there are hidden failures or performance degradation problems, and issue a warning signal, including the following:

[0061] Based on equipment manuals, expert knowledge, historical failure records and literature research, a database of typical failure modes and performance degradation indicators is established. The database includes characteristic descriptions of failure modes and degradation indicators and their value ranges.

[0062] Collect trend items corresponding to each component, compare each trend item with the features in the failure mode and performance degradation index library, and calculate the similarity to determine whether the trend item matches the known failure mode or performance degradation index;

[0063] The similarity is compared with the similarity threshold. If the similarity is greater than or equal to the similarity threshold, it is considered that there is a hidden fault or performance degradation problem, and a warning signal is generated for the corresponding component, indicating that the corresponding component has a hidden problem.

[0064] An intelligent photovoltaic water heater fault early warning and diagnosis system, comprising: an operation database construction module, an abnormal weather stripping module, an efficiency calculation module, a data envelopment processing module, a trend item extraction module, and a fault comparison module;

[0065] The operation database building module builds the operation database of the equipment with the operation time in units of days, and passes it to the abnormal weather stripping module;

[0066] The abnormal weather stripping module processes the equipment operation data in the operation database, strips the factors of abnormal weather and its delayed effects that affect photovoltaic power generation, generates stripped data, and transmits it to the efficiency calculation module;

[0067] The efficiency calculation module calculates the photoelectric efficiency and electrothermal efficiency data based on the stripped data, and transmits these data to the data envelopment analysis module;

[0068] The data envelopment processing module evaluates the output efficiency of electrical energy and thermal energy in different time units, calculates the reproducibility index and durability index, determines whether the photoelectric efficiency and electrothermal efficiency are qualified as the baseline, and transmits the judgment results to the trend item extraction module;

[0069] The trend item extraction module extracts the mean characteristics of components based on the data with baseline qualification by using the moving average method and smoothing processing, identifies and marks the trend items of each component through cluster analysis, and transmits these trend items to the fault comparison module;

[0070] The fault comparison module compares the trend items of each component with typical failure modes and performance degradation indicators, uses a matching algorithm to calculate the similarity, determines whether there are hidden faults or performance degradation problems, and issues a warning signal.

[0071] The technical effects and advantages of the intelligent photovoltaic water heater fault early warning diagnosis method and system of the present invention are as follows:

[0072] The present invention systematically analyzes the operating data of photovoltaic water heaters, removes the influence of abnormal weather, calculates and evaluates the photovoltaic efficiency and electric thermal efficiency, and determines its baseline qualification. Based on the baseline conditions, data smoothing and cluster analysis are used to extract and mark the operating trend items of components, and then these trend items are compared with typical failure modes, and the similarity is calculated to identify potential hidden failures and issue early warning signals. Ensure that the operating status of each component of the photovoltaic water heater is accurately evaluated in the long-term trend, and ensure that all analysis data are carried out under the same standard through the determination of the baseline, eliminate the interference caused by changes in environmental and operating conditions, and make the analysis results more comparable and reliable. Through the screening of the baseline, it is possible to focus on the changes in the performance of the equipment itself without being affected by external factors, so as to more accurately identify the long-term operating trends and potential performance degradation of components. The extraction and analysis of trend items can identify abnormal changes and potential failures of components in advance, making maintenance work more targeted and predictive, reducing the occurrence of sudden failures, and providing reliable early warning diagnosis, thereby improving the maintenance efficiency and operation reliability of the equipment, and ultimately optimizing the overall performance and service life of the photovoltaic water heater. BRIEF DESCRIPTION OF THE DRAWINGS

[0073] Figure 1 A schematic diagram of a process of an intelligent photovoltaic water heater fault early warning diagnosis method of the present invention;

[0074] Figure 2 The present invention is a structural schematic diagram of an intelligent photovoltaic water heater fault early warning and diagnosis system. DETAILED DESCRIPTION

[0075] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0076] Embodiment 1, Figure 1 The present invention provides a fault early warning diagnosis method for an intelligent photovoltaic water heater, comprising:

[0077] Build an operation database of equipment with daily operation time;

[0078] Stripping away the factors that affect abnormal weather and its delayed effects on photovoltaic power generation;

[0079] Calculate the photoelectric efficiency and electrothermal efficiency data based on the data analysis after stripping;

[0080] The electrical and thermal energy output efficiencies of the photovoltaic efficiency and electrothermal efficiency in different time units were evaluated by data envelopment analysis, and the reproducibility index and durability index were calculated to determine whether the photovoltaic efficiency and electrothermal efficiency were qualified as the baseline;

[0081] If the qualification baseline is obtained, based on the photovoltaic efficiency and electrothermal efficiency baseline conditions, the moving average method and smoothing process are used to extract the mean characteristics of the components, and the trend items of each component are identified and marked through cluster analysis;

[0082] The trend items of each component are compared with typical failure modes and performance degradation indicators, and a matching algorithm is used to calculate the similarity to determine whether there are hidden failures or performance degradation problems and to issue early warning signals.

[0083] Build a device operation database with daily operation time, including the following:

[0084] First, all the data of the daily operation of the equipment is collected, including but not limited to temperature, sunshine intensity, current, voltage and equipment operating status. The data source can be directly collected by sensors embedded in the equipment or indirectly obtained through external monitoring systems. The collected data needs to be preprocessed to clean, filter and format to ensure the quality and consistency of the data. The preprocessing steps include removing outliers, filling missing data, normalizing data and unifying time tags. After that, the processed data is stored in the database management system, which should support efficient data retrieval and update functions to facilitate subsequent data analysis and processing. The entire database construction process needs to ensure data security and privacy protection, and reasonably set data access rights and encryption measures to prevent data leakage or unauthorized access.

[0085] The reason for collecting data on a daily basis is that the performance of photovoltaic water heaters is significantly affected by sunlight conditions. The daily changes in sunlight intensity directly affect the photovoltaic conversion efficiency and hot water production. In addition, climate conditions such as temperature and humidity also change on a daily basis, which have a direct impact on equipment performance. By collecting data daily, it is possible to effectively track and analyze the operating performance and energy efficiency changes of the equipment under different weather conditions, and to promptly detect potential performance degradation or failures. The accumulation of data on a daily basis also provides a basis for long-term performance evaluation and maintenance plans, which helps to optimize equipment use and improve energy management efficiency, especially in the analysis of the long-term impact of seasonal changes and different weather conditions on equipment performance.

[0086] It is critical to strip away factors that affect abnormal weather and its delayed effects on photovoltaic power generation, as these factors can seriously distort the true assessment of system performance. In the field of photovoltaic power generation, the energy efficiency and performance of the system are directly affected by environmental conditions, especially meteorological conditions such as temperature, sunshine intensity and precipitation. Abnormal weather events such as extreme heat, heavy rain or continuous cloudy days not only affect the immediate power generation efficiency of photovoltaic panels, but may also cause long-term damage to the equipment, such as accelerated component aging or structural damage, which may continue to affect the performance of the system for a period of time after the event. By identifying and stripping away the data distortions caused by these factors, the true performance of photovoltaic systems under standard operating conditions can be accurately analyzed, thereby effectively optimizing system design, scheduling maintenance and ensuring performance. In addition, accurate data analysis can help operators and maintenance teams predict future performance trends, reasonably arrange maintenance resources, optimize operating costs and ensure return on investment, which is crucial in improving the economic benefits and sustainability of photovoltaic power generation projects.

[0087] The factors affecting abnormal weather and its delayed effects on photovoltaic power generation are stripped away, including the following:

[0088] The data in the equipment operation database is called up and classified into meteorological data and performance data of the photovoltaic system. The meteorological data includes temperature, wind speed, precipitation and cloud coverage.

[0089] The technical details of using variational autoencoders (VAE) to extract key features from the equipment operation database cover data preprocessing, model building, training and optimization. In the data preprocessing stage, the operation data is processed by a standardized method to adapt to the input requirements of the neural network. The architecture of the VAE model is divided into two parts: encoder and decoder. The encoder maps high-dimensional data to a low-dimensional latent space through a multi-layer neural network, while the decoder maps the representation of the latent space back to the original data space. Model training involves using reconstruction loss and KL divergence as loss functions, where the reconstruction loss ensures that the decoded output is as close to the original input as possible, and the KL divergence drives the distribution of the latent variables closer to the normal distribution. By iteratively optimizing these parameters, VAE can effectively capture the key features in the equipment operation data, providing support for further performance evaluation and anomaly detection. This is achieved by converting high-dimensional data into low-dimensional latent representations through feature extraction, thereby providing a concise and information-rich data form for anomaly detection and further causal inference.

[0090] The key features extracted by the variational autoencoder are input into the isolation forest algorithm for anomaly detection. This algorithm is specifically designed to identify abnormal patterns in data and can effectively identify abnormal data points caused by extreme weather events from the features. The detected abnormal data will be used for the next step of causal inference analysis to verify whether these anomalies are really related to weather conditions. The core principle of the isolation forest is that abnormal points are usually sparse and easy to isolate, and each sample point is isolated by building multiple decision trees. In this process, features and the cut values ​​of the features are randomly selected, and decision trees are established to record the average path length required for the sample points to be isolated; the shorter the path, the easier it is to isolate the sample and the higher the probability of anomaly. This average path length is converted into an anomaly score, and a threshold is set. Data points above the threshold are marked as anomalies. Such processing not only ensures the high sensitivity of the algorithm to anomalies, but also adapts to the actual complex data distribution, especially when there is no need to assume the data distribution, it can effectively process large-scale data sets. The labeling of abnormal data points output by the isolation forest model provides an accurately labeled abnormal data set for further causal inference analysis of the photovoltaic system, thereby ensuring the overall accuracy of the data set and the reliability of the analysis results.

[0091] On the basis of anomaly detection, causal inference techniques are used to analyze the causal relationship between confirmed abnormal weather data and photovoltaic system performance. The impact of extreme weather events on photovoltaic power generation efficiency and its ongoing consequences are determined by statistical methods or causal graph techniques. Help determine which data are affected by weather delay effects and should be removed or adjusted. When applying causal inference analysis, by constructing a detailed causal graph model, different meteorological variables such as temperature, wind speed, sunshine, etc. are associated with photovoltaic system performance indicators such as power generation and equipment efficiency, and statistical tests such as conditional independence tests are used to determine the dependencies between variables. In addition, by adjusting the causal model parameters, weather factors that are significantly related to system performance are screened out, and the quantitative analysis of causal relationships is further refined using structural equation models or Bayesian networks, and ultimately determining which performance changes are directly caused by specific meteorological conditions. The implementation of this step ensures that data affected by extreme weather and its delayed effects can be accurately distinguished and removed, providing a solid analytical basis for subsequent data adjustment and system performance evaluation.

[0092] Based on the causal inference results, the data marked as affected by extreme weather are removed.

[0093] The reason why photovoltaic efficiency and electrothermal efficiency can be used as baseline data is that these indicators directly measure the ability of the photovoltaic water heating system to convert solar energy into electrical energy and thermal energy, and accurately reflect the performance of the system under actual operating conditions. These efficiency indicators not only provide direct information about the operating efficiency of the system, but also reveal the responsiveness and stability of the system under different environmental conditions. Therefore, they are key indicators for evaluating the overall health and performance of the system. By continuously monitoring these efficiency data, possible problems in system operation can be identified, such as equipment aging, the impact of environmental changes on system performance, and potential maintenance needs. In addition, photovoltaic efficiency and electrothermal efficiency as baselines can also help optimize system design, improve energy efficiency, and serve as a standard for evaluating the effectiveness of new technologies or improvement measures. Therefore, these two efficiency indicators are of great practical value as baseline data for long-term monitoring of system performance, management decision-making, and technical optimization.

[0094] The photoelectric efficiency and electrothermal efficiency data are calculated based on the data analysis after stripping, including the following:

[0095] After removing the data marked as affected by extreme weather, the photovoltaic efficiency is independently calculated and adjusted.

[0096] Collect the operation data related to the photovoltaic system, including output power and incident light energy. At the same time, it is also necessary to obtain the first environmental data and the first equipment status data that affect the photovoltaic efficiency; calculate the light-to-electricity efficiency, that is, the ratio of output power to incident light energy. By constructing a multivariate regression model, the first environmental data and the first equipment status data variables are associated with the light-to-electricity efficiency, analyze the specific impact of the first environmental data and the first equipment status data on the light-to-electricity efficiency, and calculate and adjust the photovoltaic efficiency, taking into account all significant environmental and equipment factors. Finally, analyze the adjusted photovoltaic efficiency data, identify the efficiency change trend, and prepare a report that details the calculation method, influencing factors, and adjustment results.

[0097] First environmental data: environmental factors that directly affect the photovoltaic panels' reception and conversion of solar energy, such as sunlight intensity and ambient temperature.

[0098] First, equipment status data: physical status that affects the performance of photovoltaic panels, such as the cleanliness of photovoltaic panels, because pollution and dust can significantly reduce the energy consumption of photovoltaic panels.

[0099] Collect the output heat energy and input electric energy data of the hot water system. At the same time, collect the second environmental data and the second equipment status data that may affect the electric heat efficiency, such as water temperature and equipment maintenance records. After cleaning and standardization, these data are used to calculate the electric heat efficiency, that is, the ratio of output heat energy to input electric energy. Subsequently, by setting up a regression model, analyze the impact of the second environmental data and the second equipment status data on the electric heat efficiency, calculate and adjust the electric heat efficiency, and determine the key variables and their specific contribution to the efficiency. According to these regression analysis results, adjust the preliminary calculated electric heat efficiency to ensure that it reflects the performance under actual operating conditions. After completing the efficiency adjustment, conduct a detailed efficiency data analysis, evaluate the effectiveness of the adjustment, and prepare a detailed report to explain the analysis method, adjustment process and final results.

[0100] Second environmental data: environmental factors that affect the efficiency of the hot water system, such as water temperature, because the water temperature will directly affect the energy required for heating.

[0101] Second, equipment status data: data that reflects the operation and maintenance status of equipment, such as equipment maintenance records. Poor maintenance may lead to decreased efficiency.

[0102] It is crucial to analyze and calculate the photovoltaic efficiency and electrothermal efficiency and use them as baseline data, because these efficiency indicators directly reflect the ability of the photovoltaic water heating system to convert solar energy into electrical energy and thermal energy, and are the core parameters for evaluating system performance. Effective photovoltaic efficiency and electrothermal efficiency not only reveal the operating efficiency of the system under the current environmental and technical conditions, but also provide a basis for optimizing system design, improving energy utilization efficiency, and formulating maintenance strategies. By using these efficiency indicators as a baseline, it is possible to more accurately monitor the changes in system performance over time, predict the challenges that the system may face in future operation, and make necessary adjustments and optimizations. At the same time, it is particularly important to combine the photovoltaic efficiency and electrothermal efficiency for analysis, because in the photovoltaic water heating system, the generation processes of electrical energy and thermal energy are interdependent and affected by the same environmental conditions. Analyzing any one efficiency alone may ignore the complex interactions and constraints of the overall system performance, while combining the two can comprehensively evaluate the overall energy efficiency and performance of the system and provide more comprehensive system improvement and decision support information. By analyzing these two parameters in detail, it is possible to more accurately judge the performance of the system under different operating conditions and the issues that need to be focused on during the design and operation and maintenance process.

[0103] Here is a hypothetical example:

[0104] Assume the following data set, which is data from three days, is used to calculate the photovoltaic efficiency and electrothermal efficiency:

[0105] date Output power (kWh) Incident light energy (kWh) Ambient temperature (°C) Sunshine intensity (kW / m²) Photovoltaic panel cleanliness (%) Output heat energy (kWh) Input power (kWh) Water temperature (°C) Equipment maintenance status 2024-06-24 150 500 25 0.5 90 120 100 30 good 2024-06-25 160 550 26 0.6 85 130 110 29 good 2024-06-26 170 600 27 0.7 80 140 120 28 Maintenance required

[0106] The calculation formula of light-to-electricity efficiency: Photoelectric efficiency (%) = (output electrical energy / incident light energy) × 100;

[0107] The calculation formula of electric-to-heat efficiency is: Electric-to-heat efficiency (%) = (output heat energy / input electric energy) × 100;

[0108] date Output power Incident light energy Output heat Input power Photoelectric efficiency (%) Electric heating efficiency (%) 2024-06-24 150 500 120 100 30 120 2024-06-25 160 550 130 110 29.09 118.18 2024-06-26 170 600 140 120 28.33 116.67

[0109] Efficiency adjustment involves making corrections to the base efficiency to take into account environmental factors and equipment status. For example, if the sunlight intensity is lower than the standard or the cleanliness of the photovoltaic panel decreases, the photovoltaic efficiency may need to be adjusted downward.

[0110] The hypothetical regression model shows that every 1% decrease in PV panel cleanliness leads to a 0.5% decrease in photovoltaic efficiency, and every 1°C decrease in water temperature leads to a 1% decrease in electrothermal efficiency.

[0111] Adjusting the photoelectric efficiency:

[0112] 2024-06-24: Photoelectric efficiency adjustment = 30%-(100%-90%)×0.5%=30%-5%=25%

[0113] 2024-06-25: Photoelectric efficiency adjustment = 29.09% - (100% - 85%) × 0.5% = 29.09% - 7.5% = 21.59%

[0114] 2024-06-26: Photoelectric efficiency adjustment = 28.33% - (100% - 80%) × 0.5% = 28.33% - 10% = 18.33%

[0115] Adjusting the electric heating efficiency:

[0116] 2024-06-24: Electric heating efficiency adjustment = 120% - (30°C - 28°C) × 1% = 120% - 2% = 118%

[0117] 2024-06-25: Electric heat efficiency adjustment = 118.18% - (29°C - 28°C) × 1% = 118.18% - 1% = 117.18%

[0118] 2024-06-26: Electric heat efficiency adjustment = 116.67% - (28°C - 28°C) × 1% = 116.67% - 0% = 116.67%

[0119] This process shows how to adjust the calculated photovoltaic and electrothermal efficiencies based on specific data and models. This approach ensures that the efficiencies reflect real operating conditions, allowing for more accurate system assessments and decisions.

[0120] The purpose of evaluating the electrical and thermal output efficiency of photovoltaic efficiency and electrothermal efficiency in different time units by data envelopment analysis, calculating the reproducibility index and durability index, and judging whether it is qualified as a baseline is to ensure the accurate and scientific evaluation of the performance of photovoltaic water heaters. Since photovoltaic water heaters are affected by many factors during operation, such as environmental conditions, component aging, and changes in operating status, relying solely on real-time data for evaluation may lead to deviations and misjudgments. Data envelopment analysis can comprehensively consider a variety of input and output factors, provide a systematic and objective efficiency evaluation method, and ensure that the evaluation results have a high degree of credibility. The calculation of reproducibility index and durability index helps to measure the stability and consistency of efficiency data in different time periods, ensuring that the baseline has long-term stability and reliability. By establishing and applying a reliable baseline, the impact of external environmental changes on the data can be effectively stripped away, focusing on the performance changes of the equipment itself, so as to more accurately identify the long-term operating trends and potential performance degradation of components. This process can not only improve the accuracy and timeliness of fault warning, but also provide a scientific basis for subsequent maintenance and optimization, reduce the occurrence of sudden failures, extend the service life of equipment, and ultimately improve the overall performance and operating efficiency of photovoltaic water heaters. Therefore, evaluating and determining the baseline through this method is a key step in achieving accurate fault warning and scientific equipment management, and has important practical significance and application value.

[0121] The data envelopment analysis method is used to evaluate the electrical and thermal output efficiency of the photoelectric efficiency and electrothermal efficiency in different time units, calculate the reproducibility index and durability index, and determine whether it is qualified as a baseline, including the following:

[0122] Collect the electrical and thermal energy data calculated from the photoelectric efficiency and electrothermal efficiency in different unit times. Standardize the electrical and thermal energy data to eliminate the impact of different dimensions and ensure the data consistency of each indicator.

[0123] Each time unit (such as daily, monthly) is regarded as a decision-making unit (DMU). The data of each decision-making unit includes the electric energy calculated by the photovoltaic efficiency and the thermal energy calculated by the electric thermal efficiency.

[0124] In the data envelopment analysis model, photoelectric efficiency and electrothermal efficiency are selected as input variables, and electrical energy output and thermal energy output are selected as output variables. The specific variable definitions are as follows:

[0125] Input variable 1 : Photoelectric efficiency;

[0126] Input variable 2 : Electric heating efficiency;

[0127] Output variable 1 : Power output;

[0128] Output variable 2 : Heat output;

[0129] The CCR model (Charnes, Cooper and Rhodes) is used, assuming constant returns to scale (CRS). The linear programming form of the CCR model is as follows:

[0130] ;

[0131] ;

[0132] ;

[0133] in, For the The efficiency value of a decision-making unit, the goal is to maximize the corresponding efficiency value; Indicates The decision-making unit Outputs; Indicates The decision-making unit Inputs; Indicates The weight of the output; Indicates The weight of the input; Indicates A decision making unit (DMU) is in the The output quantity on the output variable; It means the The decision making unit is The amount of input on the input variables;

[0134] In the objective function, Indicates The efficiency value of each decision-making unit is evaluated by maximizing this value. The constraint condition ensures that for all decision-making units , its efficiency value does not exceed 1, maintaining the comparison of relative efficiency. and Must be non-negative to ensure model plausibility.

[0135] is the number of output variables. It indicates how many output variables each decision unit has in the DEA model. For example, if a system generates both electrical energy and thermal energy, then the number of output variables is That's 2.

[0136] The number of input variables. Indicates how many input variables each decision unit has in the DEA model. For example, if a system uses photovoltaic efficiency and electrothermal efficiency as input variables, then the number of input variables mmm is 2.

[0137] Represents the index of a specific output variable. For each decision unit , the output variables can be labeled ,in Indicates a specific output variable (from 1 to ).

[0138] Indicates the index of a specific decision unit. In the decision unit model, multiple decision units are analyzed and compared. Represents a specific decision unit for the current analysis. For example, if there are 10 time units, each of which is a decision unit, then Represents one of the time units.

[0139] Represents the index of a specific input variable. For each decision unit , the input variables can be marked as ,in Represents a specific input variable (from 1 to ).

[0140] Represents the index of all decision-making units. In the constraints, represents all decision-making units in the DEA model. For example, if there are 10 time units, Just from 1 to 10, covering all decision-making units.

[0141] Subject to refers to the constraints in linear programming. It specifies the restrictions in the model to ensure that the efficiency value of all decision-making units does not exceed 1. These constraints ensure the basis for comparison of relative efficiency.

[0142] Solve the linear programming form of the following data envelopment analysis (DEA) model using a linear programming solution method (such as the simplex method) to obtain the efficiency value of each decision-making unit (DMU) ;

[0143] Analyze the efficiency values ​​of each decision-making unit. A high efficiency value (close to 1) indicates that the electrical and thermal energy output of the photovoltaic efficiency and electrothermal efficiency in the time unit is high, the resource utilization (input) is relatively small or reasonable, and a high production efficiency is achieved. A low efficiency value (far from 1) indicates that the efficiency performance is poor in the time unit and may need further adjustment or optimization.

[0144] The efficiency values ​​of each time unit (DMU) are extracted from the data envelopment analysis model. These efficiency values ​​will serve as the basic data for subsequent analysis.

[0145] Evaluate the stability and reliability of efficiency values ​​over a long period of time, and calculate the reproducibility index and durability index.

[0146] Divide the entire time period into several smaller time periods, such as by month, quarter or year. Calculate the mean of the efficiency value in each time period, compare the means of adjacent time periods in turn, calculate the difference between the means, average the differences between all means, and obtain the reproducibility index. The smaller the reproducibility index, the smaller the change in the efficiency value in different time periods, the higher the repeatability, and the efficiency value shows strong consistency and stability in different time periods.

[0147] Calculate the overall mean of all efficiency values ​​in the entire time period, and calculate the difference between the efficiency value at each time point and the overall mean. Calculate the sum of the squares of all differences, find the average, take the square root, and get the durability index. The smaller the durability index, the smaller the fluctuation of the efficiency value in the entire time period, the higher the stability, and the efficiency value shows strong stability and reliability over a long period of time.

[0148] The reproducibility index measures the consistency and stability of efficiency values ​​in different time periods by comparing the mean differences in efficiency values ​​in adjacent time periods. The durability index measures the volatility and reliability of efficiency values ​​over a long period of time by calculating the difference between the efficiency value and the overall mean. The two indices comprehensively evaluate the stability and sustainability of photovoltaic efficiency and electrothermal efficiency, providing a scientific basis for determining whether they are qualified as baselines.

[0149] The reproducibility index and durability index are respectively greater than or equal to the corresponding thresholds, indicating that the photoelectric efficiency and electrothermal efficiency vary greatly in different time periods, have low repeatability and consistency, and have high volatility over a long period of time, and poor stability and reliability, indicating that the photoelectric efficiency and electrothermal efficiency may not be qualified as a baseline.

[0150] The reproducibility index and durability index are respectively less than the corresponding thresholds, indicating that the efficiency values ​​show high consistency and stability in different time periods, and have small fluctuations over a long period of time, with strong stability and reliability, which shows that the photoelectric efficiency and electrothermal efficiency are qualified as baselines.

[0151] If the qualification baseline is obtained, based on the photovoltaic efficiency and electrothermal efficiency baseline conditions, the moving average method and smoothing are used to extract the mean characteristics of the components, and the trend items of each component are identified and marked through cluster analysis, in order to achieve more accurate and systematic fault warning and performance evaluation. After the baseline is determined, it can be ensured that all analysis data are carried out under the same standard, eliminating the interference of changes in environmental and operating conditions, so that the analysis results have higher comparability and reliability. The moving average method and smoothing can effectively eliminate short-term fluctuations and highlight long-term trends, thereby more accurately reflecting the actual operating status of components. By extracting the mean characteristics, the long-term operating status of components can be quantified to provide a clear performance indicator. Cluster analysis can classify components with similar operating characteristics and identify typical operating modes and abnormal trends. This method helps to discover potential problems of components in advance. By marking and analyzing trend items, abnormal changes and performance degradation of equipment can be accurately identified, thereby providing early warning signals, making maintenance work more targeted and predictive, reducing the occurrence of sudden failures, and improving the operating reliability and maintenance efficiency of equipment. In addition, the results of cluster analysis can provide data support for equipment optimization and improvement, guide more scientific decision-making, and ultimately optimize the overall performance and service life of photovoltaic water heaters. Therefore, through this systematic analysis method, comprehensive monitoring and accurate evaluation of equipment status can be achieved, providing a solid foundation for fault warning and maintenance.

[0152] If the qualification baseline is obtained, based on the photovoltaic efficiency and electrothermal efficiency baseline conditions, the moving average method and smoothing process are used to extract the mean characteristics of the components, and the trend items of each component are identified and marked through cluster analysis, including the following:

[0153] S5-1, if the reproducibility index and durability index are less than the corresponding thresholds, collect the operating data of each component under the same photoelectric efficiency and electrothermal efficiency numerical baseline conditions. The data includes time series of temperature, pressure, current, voltage, vibration and noise. Ensure that the data is cleaned and standardized to eliminate dimensional differences and ensure the integrity and consistency of the data.

[0154] S5-2, set a nominal window size, such as 7 days or 30 days. The choice of window size depends on the volatility characteristics of the time series to be smoothed and the purpose of analysis.

[0155] For each time point, calculate the average value of all data points in the time point and the window before and after it. The specific calculation process is as follows:

[0156] For time point , take the window size as , then the moving average is the sum of all data points in the window divided by the window size;

[0157] The calculated moving average is used as the smoothing value at the corresponding time point.

[0158] The moving average method eliminates short-term fluctuations and highlights the long-term trend of the time series through smoothing. By averaging each time point, the random noise of the data can be reduced, making the overall trend more obvious.

[0159] S5-3, for the smoothed time series data of each component, calculate the overall mean, which represents the average operating status of the corresponding component in the entire time period.

[0160] Sum the smoothed values ​​for all time points and divide by the number of time points to get the overall mean.

[0161] The calculated overall mean is used as the mean feature of the corresponding component.

[0162] The mean feature reflects the concentration trend of components over a period of time and is a key indicator for measuring the operating status of components. By extracting the mean feature, the long-term operating status of components can be quantitatively analyzed.

[0163] S5-4, select the appropriate number of clusters based on data distribution and analysis requirements The optimal number of clusters can be determined by methods such as the elbow rule or silhouette coefficient.

[0164] Random Selection Initial cluster centers are randomly selected from the data set. Mean features.

[0165] According to the mean feature, the data point of each component is assigned to the cluster center closest to it. The Euclidean distance between each data point and the cluster center is calculated, and the cluster center with the closest distance is selected for assignment.

[0166] Recalculate the center position of each cluster, that is, the mean of all data points belonging to the cluster.

[0167] Repeat the steps of assigning data points and updating cluster centers until the cluster centers no longer change significantly or the preset number of iterations is reached.

[0168] Cluster analysis helps identify and analyze key trend items by grouping components with similar operating characteristics into one category. The mean of the cluster center represents the typical operating status of the component, and these typical statuses are marked as trend items.

[0169] S5-5, calculate the mean of each cluster center as the typical operating state of the corresponding type of parts. The mean of the cluster center can reflect the overall trend and performance of this type of parts.

[0170] According to the mean of the cluster centers, the trend items of each component are marked. For each component, determine which class it belongs to and mark its trend item as the mean of the cluster centers of that class.

[0171] Clustering result analysis can accurately reveal the long-term operating trends of each component by identifying the mean of the cluster center. Marking these trend items can provide a clear reference for subsequent maintenance and optimization.

[0172] Trend items refer to the changing patterns and development directions of the operating status of components within a certain time range, identified by analyzing time series data and statistical feature extraction results. These patterns and directions can reflect the performance changes, potential failures and long-term operating trends of components, providing a reference for maintenance and optimization.

[0173] According to the trend items of each component, the typical failure mode and performance degradation index are compared, and the matching algorithm is used to calculate the similarity to determine whether there are hidden failures or performance degradation problems, and to issue early warning signals, in order to achieve early detection and accurate early warning of potential problems of photovoltaic water heaters. By comparing the actual operation data with the typical failure mode and performance degradation index, the potential abnormal state and failure signs during the operation of the equipment can be identified. The use of matching algorithms can quantify the similarity between trend items and failure modes, and provide an objective judgment standard, thereby improving the accuracy and reliability of fault identification. Hidden failures and performance degradation usually manifest as subtle changes in the early stages, which may be difficult to detect with traditional methods. Through this systematic and data-driven analysis method, these subtle changes can be effectively captured and early warning signals can be issued in time. Doing so can not only prevent small problems from turning into major failures and reduce the unplanned downtime of equipment, but also extend the service life of equipment and reduce maintenance costs. At the same time, the issuance of early warning signals makes maintenance work more targeted and planned, improving maintenance efficiency and resource utilization. In addition, through long-term trend analysis and failure mode comparison, a large amount of operating data and experience can be accumulated, providing important data support and scientific basis for equipment optimization and improvement. Overall, this method can significantly improve the operating reliability and safety of photovoltaic water heaters, ensure that the system operates in the best state, and thus realize intelligent and efficient equipment management.

[0174] The trend items of each component are compared with typical failure modes and performance degradation indicators, and the similarity is calculated using a matching algorithm to determine whether there are hidden failures or performance degradation problems, and to issue early warning signals, including the following:

[0175] Based on equipment manuals, expert knowledge, historical failure records and literature research, a database of typical failure modes and performance degradation indicators is established. The database includes characteristic descriptions of failure modes and degradation indicators and their numerical ranges, such as the rate of temperature rise, the frequency and amplitude of pressure fluctuations, and the range of current anomalies.

[0176] Collect the trend items corresponding to each component and select the appropriate matching algorithm for comparative analysis of trend items, failure modes and performance degradation indicators. Commonly used matching algorithms include:

[0177] Euclidean distance: used to calculate the similarity between trend items and fault modes. The smaller the distance, the higher the similarity.

[0178] Dynamic Time Warping (DTW): It is used to process time series data and can find the best alignment between two series.

[0179] Correlation coefficient: It is used to measure the linear relationship between the trend item and the failure mode. The higher the correlation coefficient, the stronger the correlation.

[0180] The extracted trend items are compared with the features in the failure mode and performance degradation indicator library. The similarity is calculated to determine whether the trend items match the known failure modes or performance degradation indicators. If the match is high, it indicates the presence of potential failure or performance degradation issues.

[0181] The similarity is compared with the similarity threshold. If the similarity is greater than or equal to the similarity threshold, it is considered that there is a hidden fault or performance degradation problem, and a warning signal is generated for the corresponding component, indicating that the corresponding component has a hidden problem.

[0182] The present invention systematically analyzes the operating data of photovoltaic water heaters, removes the influence of abnormal weather, calculates and evaluates the photovoltaic efficiency and electric thermal efficiency, and determines its baseline qualification. Based on the baseline conditions, data smoothing and cluster analysis are used to extract and mark the operating trend items of components, and then these trend items are compared with typical failure modes, and the similarity is calculated to identify potential hidden failures and issue early warning signals. Ensure that the operating status of each component of the photovoltaic water heater is accurately evaluated in the long-term trend, and ensure that all analysis data are carried out under the same standard through the determination of the baseline, eliminate the interference caused by changes in environmental and operating conditions, and make the analysis results more comparable and reliable. Through the screening of the baseline, it is possible to focus on the changes in the performance of the equipment itself without being affected by external factors, so as to more accurately identify the long-term operating trends and potential performance degradation of components. The extraction and analysis of trend items can identify abnormal changes and potential failures of components in advance, making maintenance work more targeted and predictive, reducing the occurrence of sudden failures, and providing reliable early warning diagnosis, thereby improving the maintenance efficiency and operation reliability of the equipment, and ultimately optimizing the overall performance and service life of the photovoltaic water heater.

[0183] Embodiment 2, Figure 2 The present invention provides an intelligent photovoltaic water heater fault early warning and diagnosis system, including: an operation database construction module, an abnormal weather stripping module, an efficiency calculation module, a data envelopment processing module, a trend item extraction module, and a fault comparison module;

[0184] The operation database building module builds the operation database of the equipment with the operation time in units of days, and passes it to the abnormal weather stripping module;

[0185] The abnormal weather stripping module processes the equipment operation data in the operation database, strips the factors of abnormal weather and its delayed effects that affect photovoltaic power generation, generates stripped data, and transmits it to the efficiency calculation module;

[0186] The efficiency calculation module calculates the photoelectric efficiency and electrothermal efficiency data based on the stripped data, and transmits these data to the data envelopment analysis module;

[0187] The data envelopment processing module evaluates the output efficiency of electrical energy and thermal energy in different time units, calculates the reproducibility index and durability index, determines whether the photoelectric efficiency and electrothermal efficiency are qualified as the baseline, and transmits the judgment results to the trend item extraction module;

[0188] The trend item extraction module extracts the mean characteristics of components based on the data with baseline qualification by using the moving average method and smoothing processing, identifies and marks the trend items of each component through cluster analysis, and transmits these trend items to the fault comparison module;

[0189] The fault comparison module compares the trend items of each component with typical failure modes and performance degradation indicators, uses a matching algorithm to calculate the similarity, determines whether there are hidden faults or performance degradation problems, and issues a warning signal.

[0190] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0191] The above description is only by way of illustration of certain exemplary embodiments of the present invention. It is undoubted that those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

[0192] It should be noted that, in this article, if there are relational terms such as first and second, etc., they are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "including a..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0193] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A fault early warning diagnosis method for an intelligent photovoltaic water heater, characterized in that: Build an operation database of equipment with daily operation time; Stripping away the factors that affect abnormal weather and its delayed effects on photovoltaic power generation; Calculate the photoelectric efficiency and electrothermal efficiency data based on the data analysis after stripping; The electrical and thermal energy output efficiencies of the photovoltaic efficiency and electrothermal efficiency in different time units were evaluated by data envelopment analysis, and the reproducibility index and durability index were calculated to determine whether the photovoltaic efficiency and electrothermal efficiency were qualified as the baseline; If the qualification baseline is obtained, based on the photovoltaic efficiency and electrothermal efficiency baseline conditions, the moving average method and smoothing process are used to extract the mean characteristics of the components, and the trend items of each component are identified and marked through cluster analysis; Compare the trend items of each component with typical failure modes and performance degradation indicators, and use a matching algorithm to calculate the similarity to determine whether there are hidden failures or performance degradation problems, and issue early warning signals; The data envelopment analysis method is used to evaluate the electrical and thermal output efficiency of the photoelectric efficiency and electrothermal efficiency in different time units, calculate the reproducibility index and durability index, and determine whether it is qualified as a baseline, including the following: Collect the electric and thermal energy data calculated by the photoelectric efficiency and electrothermal efficiency in different unit time, and standardize the electric and thermal energy data; Each time unit is regarded as a decision unit, and the data of each decision unit includes the electric energy calculated by the photoelectric efficiency and the thermal energy calculated by the electrothermal efficiency; In the data envelopment analysis model, photoelectric efficiency and electrothermal efficiency are selected as input variables, and electrical energy output and thermal energy output are selected as output variables. The specific variable definitions are as follows: Input variable 1 : Photoelectric efficiency; Input variable 2 : Electric heating efficiency; Output variable 1 : Power output; Output variable 2 : Heat output; Using the CCR model and assuming that the scale returns remain constant, the linear programming form of the CCR model is as follows: ; ; ; middle, For the The efficiency value of a decision-making unit, the goal is to maximize the corresponding efficiency value; Indicates The decision-making unit Outputs; Indicates The decision-making unit Inputs; Indicates The weight of the output; Indicates The weight of the input; Indicates A decision making unit (DMU) is in the The output quantity on the output variable; It means the The decision making unit is The amount of input on the input variables; is the number of output variables; The number of input variables; Represents the index of a specific output variable; An index representing a specific decision-making unit; Represents the index of a specific input variable; Represents the index of all decision-making units; Subject to refers to the constraints in linear programming; Use the linear programming method to solve the linear programming form of the following data envelopment analysis model to obtain the efficiency value of each decision-making unit; Extract the efficiency value of each decision-making unit from the data envelopment analysis model; Evaluate the stability and reliability of efficiency values ​​over a long period of time, and calculate the reproducibility index and durability index; Divide the entire time period into several smaller time periods, calculate the mean of the efficiency value in each time period, compare the means of adjacent time periods in turn, calculate the difference between the means, average the differences between all means, and obtain the reproducibility index; Calculate the overall mean of all efficiency values ​​in the entire time period, calculate the difference between the efficiency value at each time point and the overall mean, calculate the sum of the squares of all differences, find the average, take the square root, and get the durability index; The reproducibility index and durability index are respectively greater than or equal to the corresponding thresholds, indicating that the photovoltaic efficiency and electrothermal efficiency are not qualified as the baseline; The reproducibility index and durability index are respectively less than the corresponding thresholds, indicating that the photoelectric efficiency and electrothermal efficiency are qualified as baselines.

2. The intelligent photovoltaic water heater fault early warning diagnosis method according to claim 1 is characterized by: The factors affecting abnormal weather and its delayed effects on photovoltaic power generation are stripped away, including the following: Call the data from the equipment operation database and classify the data into meteorological data and photovoltaic system performance data; Use variational autoencoders to extract key features from the equipment operation database; The key features extracted by the variational autoencoder are input into the isolation forest algorithm for anomaly detection, from which abnormal data points caused by extreme weather events are identified; Based on anomaly detection, causal inference technology is used to analyze the causal relationship between confirmed abnormal weather data and photovoltaic system performance. The impact of extreme weather events on photovoltaic power generation efficiency and its lasting consequences are determined through causal graph technology, and data affected by extreme weather and its delayed effects are removed.

3. The intelligent photovoltaic water heater fault early warning diagnosis method according to claim 1 is characterized by: The photoelectric efficiency and electrothermal efficiency data are calculated based on the data analysis after stripping, including the following: After removing the data marked as affected by extreme weather, the photovoltaic efficiency is independently calculated and adjusted; Collecting operation data related to the photovoltaic system, including output power and incident light energy, and obtaining first environmental data and first device status data that affect photovoltaic efficiency; Calculate the light-to-electricity efficiency, that is, the ratio of output electrical energy to incident light energy, and associate the first environment data and the first device status data variables with the light-to-electricity efficiency by constructing a multivariate regression model, analyze the specific impact of the first environment data and the first device status data on the light-to-electricity efficiency, and calculate and adjust to obtain the photoelectric efficiency; First environmental data: environmental factors that directly affect the reception and conversion of solar energy by photovoltaic panels; First equipment status data: physical status that affects the performance of the photovoltaic panel; Collect the output heat energy and input electric energy data of the hot water system, and collect the second environment data and the second equipment status data that affect the electric heat efficiency, calculate the electric heat conversion efficiency, that is, the ratio of the output heat energy to the input electric energy, analyze the influence of the second environment data and the second equipment status data on the electric heat efficiency by setting up a regression model, and calculate and adjust to obtain the electric heat conversion efficiency; Second environmental data: environmental factors that affect the efficiency of the hot water system; Second, equipment status data: data reflecting the operation and maintenance status of equipment.

4. The intelligent photovoltaic water heater fault early warning diagnosis method according to claim 1 is characterized by: If the qualification baseline is obtained, based on the photovoltaic efficiency and electrothermal efficiency baseline conditions, the moving average method and smoothing process are used to extract the mean characteristics of the components, and the trend items of each component are identified and marked through cluster analysis, including the following: S5-1, if the reproducibility index and the durability index are respectively less than the corresponding thresholds, then the operation data of each component under the same photoelectric efficiency and electrothermal efficiency numerical baseline conditions are collected; S5-2, set a rated window size, and for each time point, calculate the average value of all data points in the time point and the windows before and after it. The specific calculation process is as follows: For time point , take the window size as , then the moving average is the sum of all data points in the window divided by the window size; The calculated moving average is used as the smoothing value at the corresponding time point; S5-3, for the smoothed time series data of each component, calculate the overall mean, which represents the average operating status of the corresponding component in the entire time period; Add the smoothed values ​​of all time points and divide by the number of time points to get the overall mean; The calculated overall mean is used as the mean feature of the corresponding parts; S5-4, random selection Initial cluster centers are randomly selected from the data set. mean features; According to the mean feature, the data point of each component is assigned to the nearest cluster center, the Euclidean distance between each data point and the cluster center is calculated, and the nearest cluster center is selected for assignment; Recalculate the center position of each cluster, that is, the mean of all data points belonging to the corresponding cluster; Repeat the steps of assigning data points and updating cluster centers until the cluster centers no longer change significantly or the preset number of iterations is reached; S5-5, mark the trend items of each component according to the mean of the cluster center.

5. The intelligent photovoltaic water heater fault early warning diagnosis method according to claim 4 is characterized in that: The trend items of each component are compared with typical failure modes and performance degradation indicators, and the similarity is calculated using a matching algorithm to determine whether there are hidden failures or performance degradation problems, and to issue early warning signals, including the following: Based on equipment manuals, expert knowledge, historical failure records and literature research, a database of typical failure modes and performance degradation indicators is established. The database includes characteristic descriptions of failure modes and degradation indicators and their value ranges. Collect trend items corresponding to each component, compare each trend item with the features in the failure mode and performance degradation index library, and calculate the similarity to determine whether the trend item matches the known failure mode or performance degradation index; The similarity is compared with the similarity threshold. If the similarity is greater than or equal to the similarity threshold, it is considered that there is a hidden fault or performance degradation problem, and a warning signal is generated for the corresponding component, indicating that the corresponding component has a hidden problem.

6. An intelligent photovoltaic water heater fault early warning diagnosis system, used to implement an intelligent photovoltaic water heater fault early warning diagnosis method according to any one of claims 1 to 5, comprising: Run the database construction module, abnormal weather stripping module, efficiency calculation module, data envelopment processing module, trend item extraction module, and fault comparison module; The operation database building module builds the operation database of the equipment with the operation time in units of days, and passes it to the abnormal weather stripping module; The abnormal weather stripping module processes the equipment operation data in the operation database, strips the factors of abnormal weather and its delayed effects that affect photovoltaic power generation, generates stripped data, and transmits it to the efficiency calculation module; The efficiency calculation module calculates the photoelectric efficiency and electrothermal efficiency data based on the stripped data, and transmits these data to the data envelopment analysis module; The data envelopment processing module evaluates the output efficiency of electrical energy and thermal energy in different time units, calculates the reproducibility index and durability index, determines whether the photoelectric efficiency and electrothermal efficiency are qualified as the baseline, and transmits the judgment results to the trend item extraction module; The trend item extraction module extracts the mean characteristics of components based on the data with baseline qualification by using the moving average method and smoothing processing, identifies and marks the trend items of each component through cluster analysis, and transmits these trend items to the fault comparison module; The fault comparison module compares the trend items of each component with typical failure modes and performance degradation indicators, uses a matching algorithm to calculate the similarity, determines whether there are hidden faults or performance degradation problems, and issues a warning signal.

Citation Information

Patent Citations

  • Multi-energy complementary microgrid hierarchical control and operation optimization method

    CN111756073A

  • Method and system for establishing data twin based on photovoltaic flexible support

    CN118504433A

  • A Diagnosis Syetem of Photovoltaic Generation Based on IoT

    KR102051402B1