A method for analyzing and evaluating the photovoltaic power generation efficiency

Abstract

The present application relates to a method for analyzing and evaluating photovoltaic power generation efficiency. The method includes obtaining historical operation data, corresponding historical meteorological data, and historical power grid data of a photovoltaic power station; performing causal inference analysis on the historical operation data, historical meteorological data, and historical power grid data to obtain influencing factors affecting the power generation efficiency of the photovoltaic power station; combining the fuzzy analytic hierarchy process to comprehensively evaluate and assign weights to the influencing factors to obtain weighted influencing factors; constructing a dynamic DEA model based on time series analysis according to the weighted influencing factors, and performing efficiency evaluation based on the dynamic DEA model to obtain an efficiency evaluation result, where the efficiency evaluation result includes an efficiency change trend and true efficiency values at different time points. By combining causal inference, fuzzy analytic hierarchy process, and dynamic DEA model, this method can comprehensively and dynamically evaluate photovoltaic power generation efficiency, provide a scientific basis for operation and maintenance management, optimize resource allocation and improve power generation efficiency, and ensure the long-term stable operation of the power station in a complex environment.

Description

Technical Field

[0001] The present invention belongs to the fields of computer and communication technologies, and particularly relates to a method for analyzing and evaluating photovoltaic power generation efficiency. Background Art

[0002] With the continuous development of technologies in the field of photovoltaic power generation, the evaluation of photovoltaic power generation efficiency is the key to improving the operation efficiency of power stations and enhancing market competitiveness. However, traditional efficiency calculation models often rely on single environmental parameters such as solar irradiance and temperature for efficiency evaluation. Although these methods have certain practicality, they often ignore the interaction of multiple factors and the complex dynamic changes within the system, resulting in large errors in the evaluation results, especially in different climate conditions or complex environments. Moreover, most existing methods are based on static data analysis, making it difficult to capture the reasons and trends of efficiency fluctuations and unable to provide forward-looking decision support for power station operation and maintenance management. Summary of the Invention

[0003] Based on this, it is necessary to provide a method for analyzing and evaluating photovoltaic power generation efficiency to solve the above technical problems, so as to improve the accuracy of photovoltaic power generation efficiency evaluation and provide forward-looking decision support for the operation and maintenance management of power stations.

[0004] In a first aspect, the present application provides a method for analyzing and evaluating photovoltaic power generation efficiency, the method comprising:

[0005] Obtain the historical operation data, corresponding historical meteorological data, and historical grid data of a photovoltaic power station, where the historical operation data includes historical power generation data, power station operation and maintenance data, and photovoltaic panel cleanliness data;

[0006] Perform causal inference analysis on the historical operation data, historical meteorological data, and historical grid data to obtain influencing factors affecting the power generation efficiency of the photovoltaic power station;

[0007] Combine the fuzzy analytic hierarchy process to comprehensively evaluate and assign weights to the influencing factors to obtain weighted influencing factors;

[0008] Construct a dynamic DEA model based on time series analysis according to the weighted influencing factors, and perform efficiency evaluation based on the dynamic DEA model to obtain an efficiency evaluation result, where the efficiency evaluation result includes the efficiency change trend and the true efficiency values at different time points, and the efficiency evaluation result is used for the operation and maintenance management of the photovoltaic power station.

[0009] In one embodiment, constructing a dynamic DEA model based on time series analysis according to the weighted influencing factors includes:

[0010] Divide the weighted influencing factors into input indicators and output indicators, where the input indicators include equipment failure rate and operation and maintenance cost, and the output indicators include power generation and power generation efficiency;

[0011] Determine the time lag relationship between input indicators and output indicators through the mutual information method, and construct the time lag parameter;

[0012] Based on the time lag parameter, align the time series data of input indicators and output indicators in time lag, and construct the time lag data set;

[0013] Adopt the sliding window mechanism to dynamically divide the time lag data set, and dynamically adjust the window size according to the data coefficient of variation to obtain the sliding window;

[0014] Based on the time lag data set and the sliding window, construct the objective function of the dynamic DEA model, and obtain the efficiency value sequence at different time points by solving the objective function;

[0015] Based on the efficiency value sequence, use the Malmquist index to analyze the efficiency change, and obtain the efficiency change trend, which includes comprehensive technical efficiency, pure technical efficiency and scale efficiency;

[0016] Through the efficiency prediction model based on the weighted influence factor, generate the efficiency prediction value, and add the efficiency prediction value as a virtual decision-making unit to the DEA evaluation. Combining the efficiency value sequence and the efficiency change trend, obtain the dynamic DEA model.

[0017] In one embodiment, the objective function is:

[0018] ;

[0019] Wherein, and are the input indicator and output indicator at time t respectively, and are the input-output data in the time lag period τ respectively, λ is the weight vector, Λ is the weight constraint set, is the time lag parameter.

[0020] In one embodiment, the calculation formula of the Malmquist index is:

[0021] ;

[0022] Wherein, represents the Malmquist index, represents the input at time t and output to the distance of the efficiency frontier, represents the input at time t+1 and output to the distance of the efficiency frontier, represents the input at time t+1 and output The efficiency frontier distance at time t represents the input at time t and the output at time t + 1. The efficiency frontier distance

[0023] In one embodiment, causal inference analysis is performed on historical operation data, historical meteorological data, and historical power grid data to obtain the influencing factors affecting the power generation efficiency of the photovoltaic power station, including:

[0024] Data cleaning and data alignment are performed on historical operation data, historical meteorological data, and historical power grid data. Data cleaning includes missing value filling, outlier handling, and data standardization. Data alignment includes alignment by timestamp and resampling of non-uniformly sampled data;

[0025] Based on the Granger causality test, the causal relationships between variables in historical operation data, historical meteorological data, and historical power grid data are identified. A significance level is set, and direct influencing factors are screened out;

[0026] Based on the Bayesian network, a causal relationship model is constructed. The conditional probability distribution between variables is learned using historical operation data, historical meteorological data, and historical power grid data, and the indirect causal relationships between direct influencing factors are identified;

[0027] According to the direct influencing factors and the causal relationship model, a structural equation model is constructed to quantify the direct and indirect effects of each direct influencing factor on the power generation efficiency. The maximum likelihood estimation method is used to solve the model parameters, and the causal effect values of each direct influencing factor are obtained;

[0028] Based on the causal effect values of each direct influencing factor, a causal effect threshold is set, and the influencing factors with causal effect values greater than the causal effect threshold are screened out. These influencing factors are used as the final influencing factors affecting the power generation efficiency of the photovoltaic power station.

[0029] In one embodiment, the influencing factors are comprehensively evaluated and weighted using the fuzzy analytic hierarchy process to obtain weighted influencing factors, including:

[0030] A hierarchical structure model is constructed. The hierarchical structure model includes an objective layer, a criterion layer, and a scheme layer. The objective layer is the photovoltaic power generation efficiency. The criterion layer includes meteorological factors, equipment status, and operation and maintenance level. The scheme layer includes solar irradiance, temperature, equipment failure rate, and cleanliness;

[0031] The 1-9 scale method is used to compare the factors in the criterion layer and the scheme layer to obtain the evaluation scoring results, and the evaluation scoring results are converted into triangular fuzzy numbers to construct a fuzzy judgment matrix;

[0032] Based on the fuzzy judgment matrix, the fuzzy analytic hierarchy process is used to calculate the weights of the influencing factors, and the weight values are obtained through defuzzification;

[0033] Multiply the causal effect value of the impact factor by the corresponding weight value to obtain a weighted impact factor.

[0034] In one embodiment, the method further includes:

[0035] Based on the efficiency evaluation result, identify the time periods with abnormal efficiency, and in combination with the contribution analysis of the impact factor, determine the key impact factors that cause abnormal efficiency;

[0036] Generate an optimization instruction according to the key impact factor, and the optimization instruction is used to execute the corresponding operation and maintenance strategy of the photovoltaic power station.

[0037] In a second aspect, the present application also provides a photovoltaic power generation efficiency analysis and evaluation system, which includes:

[0038] A data acquisition module, configured to acquire historical operation data of the photovoltaic power station and corresponding historical meteorological data and historical power grid data, and the historical operation data includes historical power generation data, power station operation and maintenance data, and photovoltaic panel cleanliness data;

[0039] A factor acquisition module, configured to perform causal inference analysis on the historical operation data, historical meteorological data, and historical power grid data to obtain impact factors that affect the power generation efficiency of the photovoltaic power station;

[0040] A factor processing module, configured to comprehensively evaluate and assign weights to the impact factors by combining the fuzzy analytic hierarchy process to obtain weighted impact factors;

[0041] An efficiency evaluation module, configured to construct a dynamic DEA model based on time series analysis according to the weighted impact factor, and perform efficiency evaluation based on the dynamic DEA model to obtain an efficiency evaluation result. The efficiency evaluation result includes the efficiency change trend and the true efficiency values at different time points, and the efficiency evaluation result is used for the operation and maintenance management of the photovoltaic power station.

[0042] In a third aspect, the present application also provides a computer device, including a memory and a processor, where the memory stores a computer program, and when the processor executes the computer program, it implements the steps of the method in any one of the first aspects.

[0043] In a fourth aspect, the present application also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the steps of the method in any one of the first aspects.

[0044] In the above-mentioned method for analyzing and evaluating photovoltaic power generation efficiency, first, by obtaining the historical operation data, historical meteorological data, and historical power grid data of the photovoltaic power station, a comprehensive data basis is provided for subsequent power generation efficiency analysis. The historical operation data includes power generation data, power station operation and maintenance data, and photovoltaic panel cleanliness data, which can reflect the actual operation status of the power station. Through causal inference analysis of these data, the influence of different factors on photovoltaic power generation efficiency can be analyzed from multiple perspectives, and then the factors affecting the power generation efficiency of the photovoltaic power station can be identified, providing a scientific basis for subsequent evaluation and optimization. Secondly, combined with the fuzzy analytic hierarchy process, a comprehensive evaluation and weight assignment are carried out on these influencing factors to obtain weighted influencing factors. This step can effectively quantify the influence degree of each influencing factor on power generation efficiency and ensure the accuracy and rationality of the evaluation results. Finally, a dynamic DEA model is constructed based on the weighted influencing factors, and the power generation efficiency of the photovoltaic power station is dynamically evaluated through time series analysis to capture the trend of the change of photovoltaic power generation efficiency over time, and an efficiency evaluation result is obtained. This efficiency evaluation result not only includes the efficiency change trend but also provides the true efficiency values at different time points, providing an important reference basis for the operation and maintenance management of the photovoltaic power station.

[0045] Compared with the traditional method for analyzing photovoltaic power generation efficiency, this method combines causal inference, fuzzy analytic hierarchy process, and dynamic DEA model, which can not only more comprehensively identify the factors affecting power generation efficiency but also dynamically reflect the change trend of efficiency, helping operation and management personnel to timely discover potential problems and make optimization adjustments, and thus enabling the operation and maintenance management of the photovoltaic power station to be more accurate and scientific, effectively improving the long-term stability and power generation efficiency of the power station. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following will briefly introduce the drawings required for use in the description of the embodiments or related technologies. Obviously, the following drawings are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0047] Figure 1 Flowchart of a method for analyzing and evaluating photovoltaic power generation efficiency provided by an exemplary embodiment of the present invention;

[0048] Figure 2 Schematic structural diagram of a system for analyzing and evaluating photovoltaic power generation efficiency provided by an exemplary embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0049] In order to make the objectives, technical solutions and advantages of the present application more clear and understandable, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0050] In one embodiment, as Figure 1 shown, a method for analyzing and evaluating the photovoltaic power generation efficiency is provided. In this embodiment, taking the application to a terminal as an example, it can be understood that this can also be applied to a server, and can also be applied to a system including a terminal and a server, and is implemented through the interaction between the terminal and the server. In this embodiment, the method includes the following steps:

[0051] S101: Obtain the historical operation data of the photovoltaic power station, the corresponding historical meteorological data, and the historical power grid data. The historical operation data includes historical power generation data, power station operation and maintenance data, and photovoltaic panel cleanliness data.

[0052] The historical operation data is a key information source for evaluating the performance of the photovoltaic power station. Among them, the historical power generation data contains the power generation information of the photovoltaic power station in different past time periods, including power generation efficiency and power generation amount, etc., and can further reflect the actual power generation capacity of the power station under different conditions. The power station operation and maintenance data can include information such as equipment maintenance records, maintenance costs, and repair times. Through these information, it can be further understood whether the power generation efficiency has been affected due to equipment maintenance problems during the past operation of the power station. The photovoltaic panel cleanliness data can reflect the impact of the cleanliness of the photovoltaic panel on the power generation efficiency. If the photovoltaic panel has not been cleaned for a long time, it may be covered by dust, debris, etc., affecting its ability to absorb solar energy and thus affecting the power generation efficiency. The historical meteorological data can include past weather information of the area where the photovoltaic power station is located, such as light intensity, temperature, humidity, wind speed, rainfall, etc. Schematically, different meteorological conditions will directly affect the power generation efficiency of the photovoltaic power station. For example, the light intensity directly affects the amount of solar energy obtained, and too high or too low temperature may affect the performance of the photovoltaic panel. The historical power grid data can include information related to the connection of the photovoltaic power station to the power grid, such as the voltage stability of the power grid and the losses during power transmission. These information can understand whether the interaction between the power station and the power grid has an impact on the power generation efficiency. For example, in the case of unstable power grid voltage, it may impose certain restrictions on the output power of the photovoltaic power station.

[0053] S102: Perform causal inference analysis on the historical operation data, historical meteorological data, and historical power grid data to obtain the influencing factors that affect the power generation efficiency of the photovoltaic power station.

[0054] Specifically, causal inference analysis can identify the causal relationships between different data, and then determine which factors affect the power generation efficiency of a photovoltaic power station and the degree and direction of the influence between different factors. For example, by analyzing the relationship between the number of equipment repairs and the power generation efficiency in historical operation data, it can be identified that when the number of equipment repairs increases, the power generation efficiency will decrease. Then, the equipment repair situation can be judged as an influencing factor affecting the power generation efficiency. For historical meteorological data, by observing the power generation efficiency data under different light intensities and combining time series analysis, the causal impact of light intensity on power generation efficiency can be determined. Exemplarily, after the light intensity reaches a certain level, the power generation efficiency may no longer increase linearly with the increase of light intensity, and may even decrease due to the increase in temperature (the temperature effect of the photovoltaic panel). Then, through causal inference, the light intensity can be judged as an influencing factor, and the specific influence mode of different light intensity ranges on the power generation efficiency can be determined. This step helps to screen out the factors that truly have an important impact on the power generation efficiency from complex data, exclude those irrelevant or less relevant factors, and provide a more accurate basis for subsequent evaluation and decision-making.

[0055] S103: Combine the fuzzy analytic hierarchy process to comprehensively evaluate and assign weights to the influencing factors to obtain weighted influencing factors.

[0056] The fuzzy analytic hierarchy process is a multi-criteria decision-making analysis method that combines the analytic hierarchy process and fuzzy set theory, and can comprehensively evaluate the influencing factors by considering multiple aspects of each influencing factor. For example, for the equipment operation and maintenance factor, not only the number of repairs is considered, but also multiple dimensions such as the difficulty of repair and the repair cost are considered. Based on fuzzy set theory, the influence degree of each influencing factor on the power generation efficiency is quantified, and the corresponding weights are assigned through the analytic hierarchy process to reflect the relative importance of different influencing factors in affecting the power generation efficiency. For example, if the equipment operation and maintenance factor is given a higher weight, while the photovoltaic panel cleanliness factor is given a relatively lower weight, this means that in the evaluation of power generation efficiency, the equipment operation and maintenance situation will have a greater impact on the final result than the photovoltaic panel cleanliness.

[0057] S104: Construct a dynamic DEA model based on time series analysis according to the weighted influencing factors, and conduct efficiency evaluation based on the dynamic DEA model to obtain the efficiency evaluation result. The efficiency evaluation result includes the efficiency change trend and the true efficiency values at different time points. The efficiency evaluation result is used for the operation and maintenance management of the photovoltaic power station.

[0058] Specifically, DEA (Data Envelopment Analysis) is a method used to evaluate the efficiency of multi-input and multi-output decision-making units. By constructing a dynamic DEA model based on time series analysis with weighted impact factors, the characteristics that the power generation efficiency of a photovoltaic power station changes over time can be captured, and it can better reflect the operating status and performance changes of the photovoltaic power station at different time points. For example, in different seasons and different years, the power generation efficiency of a photovoltaic power station varies due to factors such as sunshine duration, equipment aging, and maintenance conditions. Among them, the weighted impact factors can include various types of data, such as meteorological data (light intensity, temperature, etc.), operation data (equipment operation and maintenance, cleanliness of photovoltaic panels, etc.), and grid data (voltage, power factor, etc.). And different weighted impact factors are used as inputs and outputs at different time points respectively to construct a dynamic input-output relationship.

[0059] Based on the dynamic DEA model, the input-output relationship at different time points can be evaluated through mechanisms such as time series data and sliding windows to obtain the efficiency evaluation results. For example, by setting a sliding window, the relative efficiency of the photovoltaic power station in different time periods can be observed, and then it can be judged whether it reaches the optimal input-output state in different time periods and whether the resource utilization is effective. Among them, the efficiency change trend reflects the efficiency change of the photovoltaic power station over a long time span. By analyzing the efficiency change trend, it can be understood whether the power station is in a state of efficiency improvement, decline, or relative stability. The true efficiency values at different time points can be used to monitor and evaluate the operation of the power station in detail.

[0060] In the above method for analyzing and evaluating the power generation efficiency of photovoltaic power, by obtaining the historical operation data, historical meteorological data, and historical grid data of the photovoltaic power station, it provides basic data support for evaluating the power generation efficiency of the photovoltaic power station. Among them, the historical operation data includes power generation data, operation and maintenance data, and cleanliness data of photovoltaic panels, and these data can comprehensively understand the operation of the photovoltaic power station. Through causal inference analysis of these data, the mutual relationships between various factors can be identified, and the impact factors that have a significant impact on the power generation efficiency of the photovoltaic power station can be screened out. Through this process, it can be clearly identified which factors have a positive or negative impact on the power station efficiency, thus providing a basis for further analysis. After obtaining the impact factors, the fuzzy analytic hierarchy process is used to comprehensively evaluate them, and the weights of each factor are reasonably allocated to obtain the weighted impact factors. This step can effectively quantify the impact degree of each factor on the power generation efficiency and ensure the accuracy and rationality of the evaluation results.

[0061] Finally, based on these weighted impact factors, a dynamic DEA model based on time series analysis can be constructed. By considering the factors of time change, this dynamic DEA model can truly reflect the operating status of the power station at different time points, and then can dynamically evaluate the efficiency change of the photovoltaic power station, provide real efficiency values for each time point, and obtain the efficiency evaluation results. According to the efficiency evaluation results, the operation and maintenance strategies can be optimized, the power generation plan can be adjusted, and the equipment management level can be improved, so as to improve the overall power generation efficiency of the photovoltaic power station and ensure good benefits in long-term operation.

[0062] Compared with the traditional power generation efficiency analysis method, this method combines causal inference, fuzzy hierarchical analysis and dynamic DEA model, which can not only identify the factors affecting power generation efficiency more comprehensively, but also dynamically reflect the change trend of efficiency. Through scientific weight allocation of impact factors and time series analysis, this method optimizes the operation and maintenance management strategy of the photovoltaic power station, improves the evaluation accuracy of power generation efficiency, and provides strong support for the long-term stable operation of the power station. In addition, by combining advanced causal inference technology and dynamic DEA model, this method can effectively cope with complex meteorological and grid conditions, ensure the efficient operation of the photovoltaic power station in various environments, and provide a scientific basis for the fault prevention and performance optimization of the photovoltaic power station.

[0063] In one exemplary embodiment, constructing a dynamic DEA model based on time series analysis according to the weighted impact factors includes:

[0064] Dividing the weighted impact factors into input indicators and output indicators, where the input indicators include equipment failure rate and operation and maintenance cost, and the output indicators include power generation and power generation efficiency;

[0065] Determining the time lag relationship between the input indicators and output indicators through the mutual information method, and constructing time lag parameters;

[0066] Based on the time lag parameters, aligning the time series data of the input indicators and output indicators in time lag, and constructing a time lag data set;

[0067] Adopting a sliding window mechanism to dynamically divide the time lag data set, and dynamically adjusting the window size according to the data coefficient of variation to obtain a sliding window;

[0068] Based on the time lag data set and the sliding window, constructing the objective function of the dynamic DEA model, and obtaining the efficiency value sequence at different time points by solving the objective function;

[0069] Based on the efficiency value sequence, using the Malmquist index to conduct efficiency change analysis to obtain the efficiency change trend, and the efficiency change trend includes comprehensive technical efficiency, pure technical efficiency and scale efficiency;

[0070] Generate an efficiency prediction value through an efficiency prediction model based on a weighted influence factor, and use the efficiency prediction value as a virtual decision-making unit to participate in DEA evaluation. Combine the efficiency value sequence and the efficiency change trend to obtain a dynamic DEA model.

[0071] The mutual information method is a statistical method used to measure the correlation and dependence between two variables, which can determine whether there is a time delay relationship between input indicators and output indicators, and then construct a time delay parameter. For example, the failure of a device may not immediately lead to a decrease in power generation and power generation efficiency, and there may be a certain time delay. Through the mutual information method, the length of this delay, that is, the time delay relationship, can be determined. This time delay parameter can reflect the time difference between the change of input indicators and the change of output indicators, which helps to more accurately understand and analyze the dynamic relationship between various factors in a photovoltaic power station, and avoid errors that may be brought by simple simultaneity analysis. Considering the time delay relationship between input indicators and output indicators, adjust their time series data according to the time delay parameter to make the input indicators and output indicators consistent in time, and obtain a time delay data set.

[0072] The sliding window mechanism is a method of dividing time series data into multiple subsequences, which can perform local analysis on data in different time periods, observe the input-output relationship in different time ranges, and thus better capture the dynamic characteristics of the data. The data coefficient of variation reflects the degree of dispersion of the data. By dynamically adjusting the window size according to the data coefficient of variation, the window division can be made more stable and representative. Based on the time delay data set and the sliding window, construct an objective function for calculating the efficiency value of each time window. Solve this objective function through mathematical optimization methods to obtain the efficiency value sequence at different time points. These efficiency value sequences reflect the relative efficiency of the photovoltaic power station in different time windows, providing data support for subsequent analysis. The Malmquist index is a tool used to analyze the change of production efficiency. By taking the efficiency value sequences at different time points as input, the efficiency change trend can be obtained. Among them, the comprehensive technical efficiency can reflect the overall technical efficiency of the entire photovoltaic power station, the pure technical efficiency can reflect the impact of the main technical factors on the efficiency, and the scale efficiency measures the impact of the change in scale on the efficiency, which can reflect the scale economy effect of the power station, that is, whether there is an increase or decrease in efficiency due to the change in scale.

[0073] Schematically, an efficiency prediction model can be constructed based on deep learning models such as multi-layer perceptrons or long short-term memory networks, combined with weighted influence factors, to generate efficiency prediction values. By taking the efficiency prediction value as a virtual decision-making unit and adding it to the DEA evaluation, the power generation situation in a future period can be considered, making the DEA model more forward-looking. Through this process, the final dynamic DEA model fully considers time factors, the dynamic changes of data, the time-lag relationship between inputs and outputs, and future predictions, and can conduct a comprehensive and dynamic analysis of the efficiency evaluation and management of photovoltaic power plants, further improving the operation efficiency of the power plants and the scientific nature of decision-making.

[0074] In one exemplary embodiment, the objective function is:

[0075] ;

[0076] Wherein, and are the input index and output index at time t respectively, and are the input-output data of the time lag τ period respectively, λ is the weight vector, Λ is the weight constraint set, is the time lag parameter.

[0077] Specifically, the input index may cover various elements such as equipment failure rate and operation and maintenance costs, and the output index includes power generation amount, power generation efficiency, etc. and are the input-output data of the time lag τ period, which indicates that there is a time delay between input and output. For example, equipment failure may not immediately affect the power generation amount, but take effect after τ time. λ is the weight vector, which is used to weight different input and output indexes. Through reasonable weighting, the importance degree of each index in the efficiency evaluation can be more accurately reflected. Λ is the weight constraint set, which restricts the value range and conditions of the weight vector λ to ensure that the weight values are reasonable and stable. For example, it is required that the weights are non-negative or the sum of the weights is 1, etc. Under the above constraints, solving this objective function makes reach the minimum value, and this minimum value is the efficiency value of the decision-making unit at time t, which is used for subsequent work such as efficiency evaluation and analysis, and helps to understand the operation efficiency of the system at different times.

[0078] In one exemplary embodiment, the calculation formula of the Malmquist index is:

[0079] ;

[0080] Wherein, denotes the Malmquist index, represents the input at time t, and the output distance to the efficiency frontier, represents the input at time t+1 and the output distance to the efficiency frontier, represents the input at time t+1 and the output distance to the efficiency frontier at time t, represents the input at time t and the output distance to the efficiency frontier at time t+1.

[0081] Schematically, in the above formula, the efficiency frontier can be understood as the optimal output boundary that can be achieved under the given input. The distance to the efficiency frontier reflects the gap between the current input-output combination and the optimal state. This formula is obtained by calculating the product of the ratios of the distances of the input-output combinations at different times to the efficiency frontiers at different times and taking the square root . This index can reflect the change in production efficiency from time t to time t+1, including the comprehensive impact of various factors such as technological progress and technical efficiency change on production efficiency, which helps to deeply understand the dynamic evolution of efficiency in the production process and provides a scientific basis for the decision-making of photovoltaic power stations on the reasons for efficiency improvement or decline.

[0082] In one exemplary embodiment, causal inference analysis is performed on historical operation data, historical meteorological data, and historical grid data to obtain the influencing factors that affect the power generation efficiency of a photovoltaic power station, including:

[0083] Perform data cleaning and data alignment on historical operation data, historical meteorological data, and historical grid data. Data cleaning includes missing value filling, outlier handling, and data standardization. Data alignment includes alignment by timestamp and resampling of non-uniformly sampled data;

[0084] Based on the Granger causality test, identify the causal relationships between variables in historical operation data, historical meteorological data, and historical grid data, set the significance level, and screen out the direct influencing factors;

[0085] Based on the Bayesian network, construct a causal relationship model, use historical operation data, historical meteorological data, and historical grid data to learn the conditional probability distribution between variables, and identify the indirect causal relationships between direct influencing factors;

[0086] According to the direct influencing factors and the causal relationship model, construct a structural equation model, quantify the direct and indirect effects of each direct influencing factor on power generation efficiency, use the maximum likelihood estimation method to solve the model parameters, and obtain the causal effect values of each direct influencing factor;

[0087] Based on the causal effect values of each direct influencing factor, a causal effect threshold is set, and the influencing factors with causal effect values greater than the causal effect threshold are screened out, and these influencing factors are used as the final influencing factors for the power generation efficiency of the photovoltaic power station.

[0088] Illustratively, missing values can be supplemented by mean filling, median filling, or model-based prediction filling, etc., to ensure the integrity of the data. And outliers in the data can be identified and processed by statistical methods or specific algorithms such as using box plots to judge and handle outliers, to avoid their adverse effects on subsequent analysis. Finally, data normalization is performed by using methods such as Z-score normalization or Min-Max normalization to eliminate the influence of dimensions and make the data comparable. In addition, to ensure the consistency and comparability of the data, various types of data can be aligned according to the timestamp, and they can be resampled into uniformly spaced time interval data by resampling non-uniformly sampled data.

[0089] Granger causality test is a statistical method that can judge the causal relationship between two variables by testing whether the lagged values of one variable can significantly affect the current value of another variable. By setting a significance level, if the lagged value of the test result is less than this significance level, it is considered that there is a Granger causal relationship between the two variables, thus screening out the factors that directly affect the power generation efficiency of the photovoltaic power station. In addition, Bayesian network is a probabilistic graphical model that can intuitively represent the causal relationship and conditional probability distribution between variables. By using historical operation data, historical meteorological data, and historical power grid data to learn the conditional probability distribution between variables, the indirect causal relationship between direct influencing factors can be identified.

[0090] Structural equation model is a comprehensive statistical analysis method that can consider both direct and indirect relationships between multiple variables. Therefore, a structural equation model can be constructed according to the direct influencing factors and the causal relationship model. By using the maximum likelihood estimation method to solve the model parameters, the direct and indirect effects of each direct influencing factor on the power generation efficiency can be quantified, and the causal effect values of each direct influencing factor can be obtained. For example, through the structural equation model, the direct influence degree of light intensity on power generation and the indirect influence degree through other intermediate variables can be calculated. By setting a causal effect threshold, the influencing factors with causal effect values greater than this causal effect threshold are screened out, and these influencing factors are used as the final influencing factors for the power generation efficiency of the photovoltaic power station, then the factors that have a significant impact on the power generation efficiency can be highlighted, providing a key variable basis for subsequent power generation efficiency analysis and optimization.

[0091] In one exemplary embodiment, the influencing factors are comprehensively evaluated and weighted by combining the fuzzy analytic hierarchy process to obtain weighted influencing factors, including:

[0092] Construct a hierarchical structure model, which includes an objective layer, a criterion layer, and a scheme layer. The objective layer is the photovoltaic power generation efficiency. The criterion layer includes meteorological factors, equipment status, and operation and maintenance level. The scheme layer includes solar irradiance, temperature, equipment failure rate, and cleanliness;

[0093] Use the 1-9 scale method to compare the factors in the criterion layer and the scheme layer, obtain the evaluation score results, and convert the evaluation score results into triangular fuzzy numbers to construct a fuzzy judgment matrix;

[0094] Based on the fuzzy judgment matrix, use the fuzzy analytic hierarchy process to calculate the weights of the influencing factors, and obtain the weight values through defuzzification;

[0095] Multiply the causal effect value of the influencing factor by the corresponding weight value to obtain the weighted influencing factor.

[0096] Specifically, the objective layer is the photovoltaic power generation efficiency, which is the object of the final analysis. The criterion layer is a classification of the factors affecting the photovoltaic power generation efficiency from a macroscopic perspective and is the intermediate layer connecting the objective layer and the scheme layer. The scheme layer includes specific factors such as solar irradiance, temperature, equipment failure rate, and cleanliness. By decomposing complex problems into multiple levels through the hierarchical structure model, it is convenient for systematic analysis and evaluation. The 1-9 scale method is a quantitative method of subjective judgment that can compare the factors in the criterion layer and the scheme layer pairwise to evaluate their relative importance. Converting the evaluation score results into triangular fuzzy numbers based on fuzzy theory can better handle the uncertainty and fuzziness in the evaluation. And based on the triangular fuzzy numbers, constructing the fuzzy judgment matrix of the criterion layer and the scheme layer can represent the fuzzy relationship of the relative importance between different factors. Based on the fuzzy judgment matrix, use the relevant calculation steps of the fuzzy analytic hierarchy process to calculate the weights of the influencing factors, such as the multiplication and addition of fuzzy numbers, and defuzzification operations can be performed through the centroid method to convert the obtained fuzzy weights into clear weight values. Finally, multiplying the causal effect value of the influencing factor by the corresponding weight value can obtain the weighted influencing factor. Among them, the causal effect value reflects the actual influence degree of the influencing factor on the photovoltaic power generation efficiency, and the weight value reflects the relative importance of the influencing factor in the comprehensive evaluation. Through the combination of the two, the final weighted influencing factor takes into account both the actual influence of the factor and its importance in the overall evaluation, providing a more scientific and reasonable quantitative basis for subsequent work such as constructing a dynamic DEA model based on time series analysis.

[0097] In one exemplary embodiment, the method further includes:

[0098] Based on the efficiency evaluation results, identify the time periods with abnormal efficiency, and combine with the contribution degree analysis of the influencing factors to determine the key influencing factors leading to abnormal efficiency;

[0099] Generate optimization instructions according to key influencing factors, and the optimization instructions are used to execute corresponding operation and maintenance strategies for the photovoltaic power station.

[0100] Schematically, through the efficiency evaluation results obtained based on the dynamic DEA model, which include the efficiency change trend and the true efficiency values at different time points. These data can be compared with reference standards such as the preset efficiency range or historical average efficiency under normal conditions, so as to identify the time periods with abnormal efficiency. For example, if the power generation efficiency value during a certain period is significantly lower than the historical average level in the same period, or the efficiency change trend shows a sudden sharp decline, then this time period can be determined as an efficiency abnormal time period, and the contribution degrees of various influencing factors such as solar irradiance and equipment failure rate obtained from the previous analysis are combined for analysis. Among them, the contribution degree can be measured by means of the direct and indirect effects of each direct influencing factor on the power generation efficiency obtained through the structural equation model. By analyzing which influencing factor changes contribute more to the efficiency decline during the efficiency abnormal time period. For example, during a certain efficiency abnormal time period, it is found that the equipment failure rate suddenly increases, and through the contribution degree analysis, it is known that the contribution of the equipment failure rate to the power generation efficiency decline accounts for a relatively large proportion, then the equipment failure rate is the key influencing factor leading to this efficiency abnormality. And by generating corresponding optimization instructions, it is used to indicate the execution of the corresponding operation and maintenance strategies for the photovoltaic power station, so as to improve the power generation efficiency of the photovoltaic power station by specifically solving the key problems leading to the efficiency abnormality, making it return to a normal or better operating state, thereby ensuring the stable and efficient operation of the photovoltaic power station and improving the energy utilization efficiency and economic benefits.

[0101] Based on the same inventive concept, as Figure 2 shown, the embodiment of the present application also provides a photovoltaic power generation efficiency analysis and evaluation system 200, and the system includes:

[0102] A data acquisition module 201, configured to acquire historical operation data, corresponding historical meteorological data, and historical grid data of the photovoltaic power station, and the historical operation data includes historical power generation data, power station operation and maintenance data, and photovoltaic panel cleanliness data;

[0103] A factor acquisition module 202, configured to perform causal inference analysis on the historical operation data, historical meteorological data, and historical grid data to obtain influencing factors affecting the power generation efficiency of the photovoltaic power station;

[0104] A factor processing module 203, configured to comprehensively evaluate and assign weights to the influencing factors by combining the fuzzy analytic hierarchy process to obtain weighted influencing factors;

[0105] An efficiency evaluation module 204 is used to construct a dynamic DEA model based on time series analysis according to the weighted impact factors, and conduct efficiency evaluation based on the dynamic DEA model to obtain an efficiency evaluation result. The efficiency evaluation result includes the efficiency change trend and the true efficiency values at different time points, and the efficiency evaluation result is used for the operation and maintenance management of the photovoltaic power station.

[0106] The system is divided into four main modules. The data acquisition module 201 can acquire the historical operation data of the photovoltaic power station, as well as the corresponding historical meteorological data and historical grid data. The historical operation data covers historical power generation data, power station operation and maintenance data, and photovoltaic panel cleanliness data. The accurate data acquisition of this module provides a comprehensive and reliable information basis for subsequent analysis work, ensuring that the analysis process of the entire system has solid data support. The factor acquisition module 202 conducts causal inference analysis on the data provided by the data acquisition module 201, and can accurately obtain the impact factors that affect the power generation efficiency of the photovoltaic power station, and then clarify which factors play a key role in the power generation efficiency among many factors, pointing the direction for subsequent research and optimization. The factor processing module 203 can comprehensively evaluate and assign weights to the impact factors obtained by the factor acquisition module in combination with the fuzzy analytic hierarchy process. Through this scientific processing method, the factor processing module 203 can quantify the importance of different impact factors to obtain weighted impact factors, thus providing strong support for constructing a reasonable evaluation model. In the efficiency evaluation module 204, a dynamic DEA model based on time series analysis can be constructed according to the weighted impact factors, and efficiency evaluation is conducted based on this model. Finally, an efficiency evaluation result including the efficiency change trend and the true efficiency values at different time points is obtained. This efficiency evaluation result will be used for the operation and maintenance management of the photovoltaic power station, providing a key basis for the scientific operation and maintenance of the power station.

[0107] Through the collaborative work of these four modules, the system can accurately evaluate the power generation efficiency of the photovoltaic power station, which helps operation and maintenance management personnel optimize the management of the photovoltaic power station under different environments and operating conditions, thereby improving the power generation efficiency and overall performance, and ensuring the long-term stable and efficient operation of the photovoltaic power station.

[0108] Furthermore, the efficiency evaluation module 204 includes a model construction sub-unit, which is used for:

[0109] Dividing the weighted impact factors into input indicators and output indicators. The input indicators include equipment failure rate and operation and maintenance cost, and the output indicators include power generation and power generation efficiency;

[0110] Determining the time lag relationship between the input indicators and output indicators through the mutual information method to construct time lag parameters;

[0111] Based on the time lag parameters, aligning the time series data of the input indicators and output indicators in time lag to construct a time lag data set;

[0112] A sliding window mechanism is adopted to dynamically divide the time-delay data set, and the window size is dynamically adjusted according to the data coefficient of variation to obtain a sliding window;

[0113] Based on the time-delay data set and the sliding window, the objective function of the dynamic DEA model is constructed, and by solving the objective function, a sequence of efficiency values at different time points is obtained;

[0114] Based on the sequence of efficiency values, the Malmquist index is used for efficiency change analysis to obtain the efficiency change trend, and the efficiency change trend includes comprehensive technical efficiency, pure technical efficiency, and scale efficiency;

[0115] Through an efficiency prediction model based on a weighted influence factor, an efficiency prediction value is generated, and the efficiency prediction value is added as a virtual decision-making unit to the DEA evaluation. Combining the sequence of efficiency values and the efficiency change trend, a dynamic DEA model is obtained.

[0116] Exemplarily, the objective function is:

[0117] ;

[0118] Wherein, and are the input index and output index at time t respectively, and are the input-output data in the time-delay period τ respectively, λ is the weight vector, Λ is the weight constraint set, is the time-delay parameter.

[0119] Exemplarily, the calculation formula of the Malmquist index is:

[0120] ;

[0121] Wherein, represents the Malmquist index, represents the input at time t and output to the distance of the efficiency frontier, represents the input at time t + 1 and output to the distance of the efficiency frontier, represents the input at time t + 1 and output at the distance of the efficiency frontier at time t, represents the input at time t and output at the distance of the efficiency frontier at time t + 1.

[0122] Furthermore, the factor acquisition module 202 includes:

[0123] A data preprocessing subunit for performing data cleaning and data alignment on historical operation data, historical meteorological data, and historical power grid data. Data cleaning includes missing value filling, outlier handling, and data standardization. Data alignment includes alignment by timestamp and resampling of non-uniformly sampled data;

[0124] A factor screening subunit for:

[0125] Identifying the causal relationships between variables in historical operation data, historical meteorological data, and historical power grid data based on Granger causality test, setting a significance level, and screening out direct influencing factors;

[0126] Constructing a causal relationship model based on a Bayesian network, learning the conditional probability distribution between variables using historical operation data, historical meteorological data, and historical power grid data, and identifying the indirect causal relationships between direct influencing factors;

[0127] Constructing a structural equation model according to the direct influencing factors and the causal relationship model, quantifying the direct and indirect effects of each direct influencing factor on power generation efficiency, using the maximum likelihood estimation method to solve the model parameters, and obtaining the causal effect values of each direct influencing factor;

[0128] Based on the causal effect values of each direct influencing factor, setting a causal effect threshold, screening out the influencing factors with causal effect values greater than the causal effect threshold, and taking the influencing factors as the final influencing factors affecting the power generation efficiency of the photovoltaic power station.

[0129] Further, the factor processing module 203 includes:

[0130] A hierarchical structure model construction subunit for constructing a hierarchical structure model. The hierarchical structure model includes an objective layer, a criterion layer, and a solution layer. The objective layer is the photovoltaic power generation efficiency. The criterion layer includes meteorological factors, equipment status, and operation and maintenance level. The solution layer includes solar irradiance, temperature, equipment failure rate, and cleanliness;

[0131] A fuzzy evaluation subunit for comparing the factors in the criterion layer and the solution layer using the 1-9 scale method to obtain an evaluation scoring result, converting the evaluation scoring result into a triangular fuzzy number, and constructing a fuzzy judgment matrix; for calculating the weights of the influencing factors using the fuzzy analytic hierarchy process based on the fuzzy judgment matrix and obtaining the weight values through defuzzification;

[0132] A weighting subunit for multiplying the causal effect values of the influencing factors by the corresponding weight values to obtain weighted influencing factors.

[0133] Further, the system further includes an operation and maintenance implementation module for:

[0134] Based on the efficiency evaluation results, identify the time periods with abnormal efficiency, and determine the key influencing factors causing the abnormal efficiency by combining the contribution analysis of the influencing factors.

[0135] Generate an optimization instruction according to the key influencing factor, and the optimization instruction is used to execute the corresponding operation and maintenance strategy of the photovoltaic power station.

[0136] In an exemplary embodiment, the present invention also provides a computer device, including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the steps of a method for analyzing and evaluating the photovoltaic power generation efficiency of the present application are implemented. A multi-core processor is preferably used to improve the parallel processing ability of the system. Memory: Provide sufficient temporary storage space to support the operation of the program and the processing of data. The memory capacity should be large enough to accommodate a large amount of supply information and computing tasks.

[0137] In an exemplary embodiment, the present invention also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of a method for analyzing and evaluating the photovoltaic power generation efficiency of the present application are implemented. The computer-readable storage medium may include: read-only memory (ROM, Read Only Memory), random access memory (RAM, Random Access Memory), solid-state drive (SSD, Solid State Drives), or optical disc, etc. Among them, the random access memory may include resistive random access memory (ReRAM, Resistance Random Access Memory) and dynamic random access memory (DRAM, Dynamic Random Access Memory).

[0138] The above embodiments only represent several implementation manners of the embodiments of the present application, and the description thereof is relatively specific and detailed, but it should not be construed as a limitation on the scope of the patent of the embodiments of the application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the embodiments of the present application, several modifications and improvements can still be made, and these all belong to the protection scope of the embodiments of the present application.

Claims

1. A method for analyzing and evaluating the photovoltaic power generation efficiency, characterized in that, The method includes: Obtaining historical operation data of the photovoltaic power station, corresponding historical meteorological data, and historical grid data, where the historical operation data includes historical power generation data, power station operation and maintenance data, and photovoltaic panel cleanliness data; Performing causal inference analysis on the historical operation data, the historical meteorological data, and the historical grid data to obtain influencing factors affecting the power generation efficiency of the photovoltaic power station; Combining the fuzzy analytic hierarchy process to comprehensively evaluate and assign weights to the influencing factors to obtain weighted influencing factors; Constructing a dynamic DEA model based on time series analysis according to the weighted influencing factors, and performing efficiency evaluation based on the dynamic DEA model to obtain an efficiency evaluation result, where the efficiency evaluation result includes an efficiency change trend and true efficiency values at different time points, and the efficiency evaluation result is used for the operation and maintenance management of the photovoltaic power station; Among them, constructing a dynamic DEA model based on time series analysis according to the weighted influencing factors includes: Dividing the weighted influencing factors into input indicators and output indicators, where the input indicators include equipment failure rate and operation and maintenance cost, and the output indicators include power generation and power generation efficiency; Determining the time lag relationship between the input indicators and the output indicators through the mutual information method to construct time lag parameters; Based on the time lag parameters, performing time lag alignment on the time series data of the input indicators and the output indicators to construct a time lag data set; Using a sliding window mechanism to dynamically divide the time lag data set, and dynamically adjusting the window size according to the data coefficient of variation to obtain a sliding window; Based on the time lag data set and the sliding window, constructing an objective function of the dynamic DEA model, and by solving the objective function, obtaining a sequence of efficiency values at different time points; Performing efficiency change analysis based on the sequence of efficiency values using the Malmquist index to obtain an efficiency change trend, where the efficiency change trend includes comprehensive technical efficiency, pure technical efficiency, and scale efficiency; Generating an efficiency prediction value through an efficiency prediction model based on the weighted influencing factors, adding the efficiency prediction value as a virtual decision-making unit to the DEA evaluation, and combining the sequence of efficiency values and the efficiency change trend to obtain the dynamic DEA model.

2. The method according to claim 1, wherein The objective function is: ； Among them, and are the input index and the output index at time t respectively, and are the input-output data of the time lag τ period respectively, λ is the weight vector, and Λ is the weight constraint set, is the time lag parameter.

3. The method according to claim 1, characterized in that, The calculation formula of the Malmquist index is: ； Among them, represents the Malmquist index, represents the input at time t and output to the distance of the efficiency frontier, represents the input at time t + 1 and output to the distance of the efficiency frontier, represents the input at time t + 1 and output to the distance of the efficiency frontier at time t, represents the input at time t and output to the distance of the efficiency frontier at time t + 1.

4. The method according to claim 1, characterized in that, Performing causal inference analysis on the historical operation data, the historical meteorological data, and the historical grid data to obtain influencing factors affecting the power generation efficiency of the photovoltaic power station includes: Performing data cleaning and data alignment on the historical operation data, the historical meteorological data, and the historical grid data, where the data cleaning includes missing value filling, outlier processing, and data standardization, and the data alignment includes alignment by timestamp and resampling of non-uniformly sampled data; Identifying the causal relationship between variables in the historical operation data, the historical meteorological data, and the historical grid data based on Granger causality test, setting a significance level, and screening out direct influencing factors; Construct a causal relationship model based on a Bayesian network, and use the historical operation data, the historical meteorological data, and the historical power grid data to learn the conditional probability distribution between the variables, and identify the indirect causal relationships between the direct influencing factors; Construct a structural equation model according to the direct influencing factors and the causal relationship model, quantify the direct and indirect effects of each direct influencing factor on the power generation efficiency, and use the maximum likelihood estimation method to solve the model parameters to obtain the causal effect values of each direct influencing factor; Based on the causal effect values of each direct influencing factor, set a causal effect threshold, screen out the direct influencing factors whose causal effect values are greater than the causal effect threshold, and use the direct influencing factors whose causal effect values are greater than the causal effect threshold as the influencing factors that ultimately affect the power generation efficiency of the photovoltaic power station.

5. The method according to claim 4, wherein The method of combining the fuzzy analytic hierarchy process to comprehensively evaluate and assign weights to the influencing factors to obtain weighted influencing factors includes: Construct a hierarchical structure model, the hierarchical structure model includes a target layer, a criterion layer, and a scheme layer. The target layer is the photovoltaic power generation efficiency, the criterion layer includes meteorological factors, equipment status, and operation and maintenance level, and the scheme layer includes solar irradiance, temperature, equipment failure rate, and cleanliness; Use the 1-9 scale method to compare the factors in the criterion layer and the scheme layer to obtain an evaluation scoring result, and convert the evaluation scoring result into a triangular fuzzy number to construct a fuzzy judgment matrix; Based on the fuzzy judgment matrix, use the fuzzy analytic hierarchy process to calculate the weights of the influencing factors, and obtain the weight values through defuzzification; Multiply the causal effect value of the influencing factor by the corresponding weight value to obtain the weighted influencing factor.

6. The method according to claim 1, characterized in that, The method further includes: Based on the efficiency evaluation result, identify the time periods with abnormal efficiency, and combine the contribution degree analysis of the influencing factors to determine the key influencing factors that cause the abnormal efficiency; Generate an optimization instruction according to the key influencing factor, and the optimization instruction is used to execute the corresponding operation and maintenance strategy of the photovoltaic power station.

7. A photovoltaic power generation efficiency analysis and evaluation system, characterized in that, The system includes: A data acquisition module for acquiring the historical operation data, the corresponding historical meteorological data, and the historical power grid data of the photovoltaic power station. The historical operation data includes historical power generation data, power station operation and maintenance data, and photovoltaic panel cleanliness data; A factor acquisition module for performing causal inference analysis on the historical operation data, the historical meteorological data, and the historical power grid data to obtain the influencing factors that affect the power generation efficiency of the photovoltaic power station; A factor processing module for comprehensively evaluating and assigning weights to the influencing factors by combining the fuzzy analytic hierarchy process to obtain weighted influencing factors; An efficiency evaluation module for constructing a dynamic DEA model based on time series analysis according to the weighted influencing factors, and performing efficiency evaluation based on the dynamic DEA model to obtain an efficiency evaluation result. The efficiency evaluation result includes the efficiency change trend and the true efficiency values at different time points, and the efficiency evaluation result is used for the operation and maintenance management of the photovoltaic power station; Among them, the efficiency evaluation module includes a model construction subunit, and the model construction subunit is used for: Divide the weighted impact factor into input indicators and output indicators. The input indicators include equipment failure rate and operation and maintenance cost, and the output indicators include power generation and power generation efficiency; Determine the time lag relationship between the input indicators and the output indicators through the mutual information method, and construct time lag parameters; Based on the time lag parameters, perform time lag alignment on the time series data of the input indicators and the output indicators to construct a time lag data set; Adopt a sliding window mechanism to dynamically divide the time lag data set, and dynamically adjust the window size according to the data variation coefficient to obtain a sliding window; Based on the time lag data set and the sliding window, construct the objective function of the dynamic DEA model, and obtain a sequence of efficiency values at different time points by solving the objective function; Conduct efficiency change analysis using the Malmquist index based on the sequence of efficiency values to obtain the efficiency change trend, and the efficiency change trend includes comprehensive technical efficiency, pure technical efficiency and scale efficiency; Generate an efficiency prediction value through an efficiency prediction model based on the weighted impact factor, add the efficiency prediction value as a virtual decision-making unit to the DEA evaluation, and combine the sequence of efficiency values and the efficiency change trend to obtain the dynamic DEA model.

8. A computer device, comprising a memory and a processor, the memory storing a computer program, characterized in that, When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images