Photovoltaic power station power prediction method and device, computer equipment and storage medium

Abstract

The invention relates to the field of photovoltaic technology, and discloses a photovoltaic power station power prediction method and device, computer equipment and a storage medium, and the method comprises the steps: obtaining the operation data and environment data of a photovoltaic power station; calculating a target attenuation coefficient of the photovoltaic power station according to the operation data and the environment data; and correcting an initial power predicted value of the photovoltaic power station based on the target attenuation coefficient to obtain a target power predicted value, the initial power predicted value being determined according to meteorological data of the photovoltaic power station. The method solves the problems that in the prior art, the interaction effect between photovoltaic panel aging and dust deposition is not considered, and due to the fact that a complex model or newly-added hardware is depended on, the power prediction precision of an old photovoltaic power station is low, and the applicability is poor.

Description

Technical Field

[0001] This invention relates to the field of photovoltaic technology, specifically to photovoltaic power plant power prediction methods, devices, computer equipment, and storage media. Background Technology

[0002] Early-operated photovoltaic power plants are gradually entering the aging stage, facing the dual challenges of component performance degradation and surface dust accumulation. Component aging (such as yellowing of encapsulation materials and corrosion of cell grid lines) leads to a continuous decline in their photoelectric conversion efficiency. At the same time, during long-term operation, dust and particulate matter adhere to the component surface, forming a dust layer that blocks incident sunlight, causing additional power loss. These two degradation factors are not independent but rather mutually influential and exacerbate each other: the hydrophilicity of the aged component surface increases, making it more susceptible to dust adhesion and difficult to remove by natural rainfall; dust accumulation also affects component heat dissipation, and localized temperature rise accelerates the material aging process. This interactive effect between aging and dust accumulation significantly reduces the applicability of traditional forecasting methods based on the linear relationship between weather and power in older power plants, increasing prediction bias.

[0003] Currently, power prediction for aging photovoltaic power plants mainly relies on single attenuation correction, deep learning fitting based on large amounts of data, or hardware-assisted solutions such as adding monitoring equipment. However, these existing methods have significant limitations: they typically only quantify the impact of one factor, either aging or dust accumulation, without fully considering and modeling the synergistic enhancement effect between the two, leading to insufficient estimation of actual power loss. Furthermore, deep learning methods depend on high-quality, multi-dimensional historical data and strong computing power, while hardware upgrades involve additional equipment investment and maintenance costs. Both are difficult to adapt to the actual operating conditions of aging power plants with rudimentary monitoring conditions, limited data resources, and insufficient computing power. Therefore, existing technologies lack a solution that can accurately quantify the interaction effect of aging and dust accumulation, and achieve efficient, lightweight, and high-precision power prediction using only existing power plant data and publicly available meteorological information. Summary of the Invention

[0004] In view of this, embodiments of the present invention provide a photovoltaic power plant power prediction method, apparatus, computer equipment and storage medium to solve the problems of low power prediction accuracy and poor applicability of old photovoltaic power plants caused by the failure to consider the interaction effect between photovoltaic panel aging and dust accumulation, and the reliance on complex models or new hardware in the prior art.

[0005] In a first aspect, embodiments of the present invention provide a method for predicting the power output of a photovoltaic power plant, the method comprising:

[0006] The process involves acquiring operational and environmental data of a photovoltaic (PV) power station; calculating the aging degradation coefficient of the PV power station based on the operational data; calculating the dust accumulation degradation coefficient of the PV power station based on the environmental data; calculating the interaction effect coefficient between aging characteristics and dust accumulation characteristics in the PV power station based on the operational and environmental data; fusing the aging degradation coefficient, the dust accumulation degradation coefficient, and the interaction effect coefficient to obtain the target degradation coefficient of the PV power station; and correcting the initial power prediction value of the PV power station based on the target degradation coefficient to obtain the target power prediction value, wherein the initial power prediction value is determined based on meteorological data of the PV power station.

[0007] Furthermore, the calculation of the aging degradation coefficient of the photovoltaic power station based on the operating data includes:

[0008] Extract the operating years of the power station and the average annual degradation coefficient from the operating data; obtain the attribute parameters of the photovoltaic panels in the photovoltaic power station, wherein the attribute parameters include material characteristic parameters and shape characteristic parameters; calculate the aging degradation coefficient of the photovoltaic power station based on the operating years of the power station, the average annual degradation coefficient, the material characteristic parameters and the shape characteristic parameters.

[0009] Furthermore, the calculation of the dust accumulation attenuation coefficient of the photovoltaic power station based on the environmental data includes:

[0010] The average daily rainfall within the first time period is extracted from the environmental data; the lower limit of the coefficient is determined according to the value range corresponding to the ash attenuation coefficient, and the maximum ash attenuation rate corresponding to the lower limit of the coefficient is obtained; the ash attenuation coefficient of the photovoltaic power station is calculated based on the average daily rainfall, the lower limit of the coefficient, and the maximum ash attenuation rate.

[0011] Furthermore, the step of calculating the interaction effect coefficient between aging characteristics and dust accumulation characteristics in the photovoltaic power station based on the operating data and the environmental data includes:

[0012] Obtain the range of the improvement ratio of the dust adhesion rate of photovoltaic modules in the photovoltaic power station; determine the upper limit and lower limit of the ratio based on the improvement ratio range; calculate the interaction effect coefficient based on the upper limit, the lower limit, the operating years of the power station in the operating data, and the daily average rainfall in the environmental data.

[0013] Furthermore, the calculation of the interaction effect coefficient based on the upper limit of the ratio, the lower limit of the ratio, the operating years of the power plant in the operating data, and the daily average rainfall in the environmental data includes:

[0014] The interaction effect coefficients are obtained by pre-setting an interaction effect coefficient model based on the upper limit of the ratio, the lower limit of the ratio, the operating years of the power plant in the operating data, and the daily average rainfall in the environmental data; wherein, the pre-set interaction effect coefficient model has initial coefficients, which are calibrated using regression analysis based on the upper limit of the ratio, the lower limit of the ratio, the average rainfall, the dust type, and the operating years.

[0015] Furthermore, the step of correcting the initial power prediction value of the photovoltaic power station based on the target attenuation coefficient to obtain the target power prediction value includes:

[0016] Meteorological data of the photovoltaic power station during a second time period is obtained; the initial power prediction value of the photovoltaic power station is calculated based on the meteorological data; a dynamic weighting factor is determined based on the operating years of the power station in the operating data; the initial power prediction value is corrected using the dynamic weighting factor and the target attenuation coefficient to obtain the target power prediction value.

[0017] Furthermore, the method also includes:

[0018] Obtain the measured power value of the photovoltaic power station in the third time period; obtain the predicted power value of the photovoltaic power station in the third time period after correction based on the target attenuation coefficient, and compare the measured power value with the predicted power value to obtain a comparison result; iteratively optimize the annual average attenuation coefficient in the operating data based on the comparison result to obtain the optimized annual average attenuation coefficient; update the target attenuation coefficient using the optimized annual average attenuation coefficient to obtain the updated target attenuation coefficient.

[0019] Secondly, embodiments of the present invention provide a photovoltaic power plant power prediction device, the device comprising:

[0020] The system includes: an acquisition module for acquiring operational data and environmental data of a photovoltaic power station; a first coefficient calculation module for calculating the aging degradation coefficient of the photovoltaic power station based on the operational data; a second coefficient calculation module for calculating the dust accumulation degradation coefficient of the photovoltaic power station based on the environmental data; a third coefficient calculation module for calculating the interaction effect coefficient between aging characteristics and dust accumulation characteristics in the photovoltaic power station based on the operational data and the environmental data; a fourth coefficient calculation module for fusing the aging degradation coefficient, the dust accumulation degradation coefficient, and the interaction effect coefficient to obtain the target degradation coefficient of the photovoltaic power station; and a correction module for correcting the initial power prediction value of the photovoltaic power station based on the target degradation coefficient to obtain the target power prediction value, wherein the initial power prediction value is determined based on the meteorological data of the photovoltaic power station.

[0021] Thirdly, embodiments of the present invention provide a computer device, including: a memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, and the processor executing the computer instructions to perform the method described in the first aspect or any corresponding embodiment thereof.

[0022] Fourthly, embodiments of the present invention provide a computer-readable storage medium storing computer instructions that cause a computer to perform the method described in the first aspect or any of its corresponding embodiments.

[0023] The method provided in this application has the following beneficial effects:

[0024] The method provided in this application, by acquiring operational and environmental data from photovoltaic power plants, can fully utilize existing and publicly available data sources to comprehensively collect key parameters affecting the degradation of photovoltaic modules, providing a reliable data foundation for subsequent accurate quantification of degradation. This eliminates the need for additional hardware, reducing deployment difficulty. By calculating the target degradation coefficient of the photovoltaic power plant based on operational and environmental data, the method can quantify the synergistic power loss caused by module aging, surface dust accumulation, and their interaction, thus characterizing the total degradation ratio in actual operation. Correcting the initial power prediction value determined by meteorological data based on the target degradation coefficient allows for dynamic adjustment of the prediction results, effectively compensating for power deviations caused by the interaction of aging and dust accumulation. While ensuring prediction accuracy, the method maintains the lightweight model structure and low computational requirements, making the prediction results more closely match the actual output of older power plants, thus improving the accuracy and engineering applicability of power prediction. Attached Figure Description

[0025] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0026] Figure 1 This is a schematic flowchart of a photovoltaic power plant power prediction method according to some embodiments of the present invention.

[0027] Figure 2 This is a schematic flowchart of another photovoltaic power plant power prediction method according to some embodiments of the present invention.

[0028] Figure 3 This is a structural block diagram of a photovoltaic power plant power prediction device according to an embodiment of the present invention.

[0029] Figure 4This is a schematic diagram of the hardware structure of a computer device according to an embodiment of the present invention. Detailed Implementation

[0030] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0031] According to embodiments of the present invention, a method, apparatus, computer equipment, and storage medium for predicting the power of a photovoltaic power plant are provided. It should be noted that the steps shown in the flowcharts in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowcharts, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0032] This embodiment provides a method for predicting the power output of a photovoltaic power plant. Figure 1 This is a flowchart of a photovoltaic power plant power prediction method according to an embodiment of the present invention, such as... Figure 1 As shown, the process includes the following steps:

[0033] Step S101: Obtain the operation data and environmental data of the photovoltaic power station.

[0034] In this embodiment, two types of core input data are collected from the local monitoring system of the photovoltaic power plant, historical operation records, and public or commercial meteorological databases. The first type is operational data, which consists of parameters characterizing the photovoltaic power plant's own state and history. These mainly include: historical power data, i.e., the historical power generation time series recorded by the photovoltaic power plant's inverters or meters; the power plant's operating years, i.e., the number of years since the power plant was officially put into operation, directly reflecting the aging time of the modules; and the photovoltaic module's first-year rated power parameters, i.e., the nominal power under standard test conditions at the initial stage of the power plant's operation, serving as a benchmark for evaluating performance degradation. The second type is environmental data, which consists of external meteorological and climate parameters affecting the power plant's output. These mainly include: historical meteorological data, such as temperature, solar irradiance, and rainfall during the same period as historical power data; and real-time or forecast meteorological data used for future prediction, such as irradiance and temperature for the predicted day / time period. To ensure data quality, the collected raw historical power data needs to be preprocessed. For example, statistical methods such as the 3σ criterion (Laida criterion) can be used to identify and remove abnormal power points caused by equipment failure (such as inverter shutdown) or extreme weather events (such as rainstorms, hail, and typhoons), retaining only valid data samples under normal power plant operation conditions, thereby constructing a reliable dataset for subsequent modeling and analysis.

[0035] It should be noted that data acquisition can be expanded based on the actual situation of the data source. For example, in addition to reading directly from the monitoring system, the operating years can also be obtained and verified from the power plant's engineering archives or operation and maintenance logs; if the rated power parameters for the first year cannot be directly obtained, theoretical values ​​from the power plant's design documents can be used, or estimated by back-calculating based on actual power generation data from typical sunny days in the initial stage of operation. Meteorological data sources can be data from small weather stations installed at the power plant site, data from nearby meteorological bureaus, or gridded data from satellite inversion or numerical weather prediction models. When multiple data sources are available, cross-validation and fusion can be performed to improve the spatiotemporal accuracy and reliability of the data. In addition to the 3σ criterion, anomaly removal can also be performed according to business rules. For example, non-zero power data during periods when the theoretical power generation is zero at night can be directly marked as invalid; or points that significantly deviate from the normal irradiance-power range can be filtered in conjunction with irradiance data. Depending on the forecasting requirements, operational data and environmental data can be collected and aligned at hourly, daily, or finer time granularities. For this method, calculating the daily average rainfall usually requires daily granular data.

[0036] Step S102: Calculate the aging degradation coefficient of the photovoltaic power station based on the operating data.

[0037] The implementation method is as follows: extract the operating years of the power plant from the operating data. (Directly reflects aging duration) and annual average decay coefficient (Reflecting the impact of local climate on aging rate), and simultaneously acquiring the photovoltaic panel's property parameters, including material characteristic parameters. (Characteristics of the inherent anti-aging ability of component materials) and shape characteristics (Characterizing the nonlinearity of the aging process); then, these parameters are substituted into a nonlinear function model constructed based on aging theory (such as the Arrhenius model) and measured data to calculate the aging degradation coefficient, which directly quantifies the proportion of component output loss caused by the number of years of operation.

[0038] Step S103: Calculate the dust accumulation attenuation coefficient of the photovoltaic power station based on environmental data. This is achieved by: extracting the daily average rainfall d within a first time period (e.g., the last 7 days) from the environmental data (as a key inverse indicator characterizing the degree of dust accumulation); determining the range of the dust accumulation attenuation coefficient and its lower limit (e.g., 0.85) based on engineering experience, where the lower limit corresponds to the maximum dust accumulation attenuation rate (e.g., 15%); and finally, based on the linear correlation characteristic of "more rainfall, less dust accumulation, and lighter attenuation," constructing a calculation model to calculate the dust accumulation attenuation coefficient using the daily average rainfall, the lower limit of the coefficient, and the maximum attenuation rate. This coefficient quantifies the impact of recent rainfall on dust shading on the module surface.

[0039] Step S104: Calculate the interaction effect coefficient between aging characteristics and dust accumulation characteristics in the photovoltaic power station based on operational and environmental data. This is achieved by: obtaining the range of increases in the dust accumulation rate of photovoltaic modules (e.g., 30%-50% increase in dust accumulation rate for older modules compared to newer modules), and determining the upper and lower limits of this ratio for calculation; then, constructing a mathematical model coupling the power station's operating years t (aging characteristics) and average daily rainfall d (dust accumulation characteristics). This model ensures that the longer the operating years and the lower the rainfall, the larger the calculated interaction effect coefficient, thereby quantifying the additional power loss multiplication effect caused by the synergistic effect of aging leading to easier dust accumulation and dust accumulation accelerating aging. Step S105: Calculate the interaction effect coefficient between aging characteristics and dust accumulation characteristics in the photovoltaic power station based on the operational and environmental data. The calculated aging attenuation coefficient, ash accumulation attenuation coefficient, and interaction effect coefficient are combined (typically by multiplication) to obtain a comprehensive target attenuation coefficient, which ultimately characterizes the total output attenuation ratio under the combined effects of aging, ash accumulation, and their interaction effects.

[0040] It should be noted that the annual average decay coefficient The determination of the value can be based on empirical values ​​selected according to the regional climate type (arid / mild), or it can be derived by using historical power data of the power plant. This can be achieved by comparing typical output performance at the beginning of operation with recent performance, and using curve fitting or regression analysis to better reflect the actual aging rate of the power plant. Material characteristic parameters With shape characteristic parameters If it is not possible to obtain the data directly from the component manufacturer, it can be calibrated by collecting degradation test data of the same model of components under different operating years and using parameter estimation methods (such as maximum likelihood estimation).

[0041] The time period is not limited to 7 days and can be adjusted to 5 or 10 days based on the local dust settling and rain-induced cleaning patterns. The calculation model can incorporate a rainfall threshold constraint. For example, when the average daily rainfall exceeds a certain threshold (such as 20 mm), the accumulated dust is considered to be completely cleaned, and the dust attenuation coefficient is directly set to 1.

[0042] The range of the increased proportion can be obtained by analyzing measured data of dust deposition samples from power plants of different operating years in the same region. The coefficients in the calculation formula of the interaction effect coefficient (such as the coefficients after converting the upper and lower limits of the proportion) can be fine-tuned based on the historical data and performance loss analysis of specific power plants to more accurately reflect the coupling strength between aging and dust accumulation in that power plant.

[0043] As an example, suppose a photovoltaic power station has been in operation for a certain number of years. Annual average attenuation coefficient The value was determined to be 0.018 after fitting, which is the material property parameter. Shape characteristic parameters First, calculate the aging degradation coefficient and substitute it into the formula:

[0044]

[0045] in, This is the aging degradation coefficient; These are material property parameters; The annual average attenuation coefficient; The operating life of the power station; These are shape characteristic parameters.

[0046] Calculated (That is, aging results in 85% output retention). Secondly, the average daily rainfall over the past 7 days, d=2 mm, was obtained from environmental data. The lower limit of the ash accumulation attenuation coefficient was set at 0.85 (corresponding to a maximum attenuation rate of 15%), and the ash accumulation attenuation coefficient was calculated using a linear model. Next, given that the rate of dust accumulation on old components increases by 30%-50%, a proportional parameter is set based on this. Combining t=10 years and d=2 mm, the interaction effect coefficient formula is substituted into the formula (e.g., ...). ), calculated Finally, by combining the three factors, the target attenuation coefficient is obtained. This value is approximately 1, but the exact value needs to be calculated according to a specific formula. It comprehensively represents the total attenuation effect of the power station under its current aging and ash accumulation conditions.

[0047] Step S106: Correct the initial power prediction value of the photovoltaic power station based on the target attenuation coefficient to obtain the target power prediction value. The initial power prediction value is determined based on the meteorological data of the photovoltaic power station.

[0048] In this embodiment, firstly, meteorological data of the photovoltaic power station within a second time period (i.e., the future period to be predicted, such as the forecast irradiance and temperature for the next 24 hours) is acquired, and the initial power prediction value of the photovoltaic power station is calculated based on the meteorological data. This initial value is a raw prediction derived from a basic "meteorological-power" correlation model (such as a physical formula or a simple regression model) without considering specific degradation states. Subsequently, a dynamic weighting factor is determined based on the operating years of the power station in the operating data. This factor is a value between 0 and 1, used to balance the contribution ratio between the initial prediction and degradation correction. Its value is usually set as follows: when the operating years are shorter, the basic meteorological correlation is given a higher weight; when the operating years are longer, the weight of degradation correction is increased to reflect the dominance of the impact of aging and ash accumulation. Finally, the initial power prediction is corrected using a dynamic weighting factor and a target attenuation coefficient. Specifically, the target attenuation coefficient is applied to the initial prediction to obtain an attenuation correction term. Then, the initial prediction and this attenuation correction term are weighted and fused using the dynamic weighting factor. The calculation formula is: Target power prediction = Dynamic weighting factor × Initial power prediction + (1 - Dynamic weighting factor) × (Target attenuation coefficient × Initial power prediction). Through this mechanism, the model can adaptively adjust the prediction benchmark according to the aging of the power plant: for newer power plants, it relies more on the original meteorological correlation; for older power plants, it focuses more on output reduction based on the attenuation coefficient. This allows the target power prediction to more accurately reflect the actual power generation capacity of the power plant within a lightweight structure.

[0049] It should be noted that, in addition to being directly linearly or piecewise related to the operating years, the dynamic weighting factor can also incorporate the target attenuation coefficient itself as an auxiliary variable. For example, when the target attenuation coefficient is below a certain threshold (indicating severe overall attenuation), the dynamic weighting factor can be appropriately reduced (i.e., the weight of attenuation correction can be increased), even if the operating years have not reached the preset threshold. The model used for calculation based on meteorological data can be a variety of lightweight models, such as: formula-based methods based on the physical characteristics of photovoltaic modules (substituting irradiance and temperature into standard formulas), simple linear regression models based on historical data, or time series models (such as ARIMA). For older power plants, simple models calibrated or commonly used during their initial commissioning can be directly adopted, without the need to retrain complex models. To ensure numerical stability, the results after weighted calculation can also be validated for reasonableness, for example, ensuring that the predicted target power value is not greater than the theoretical limit power (module rated power × irradiance utilization coefficient) and not lower than zero.

[0050] As an example, assuming a photovoltaic power station has an operating life of t=12 years, the calculated target degradation coefficient is 0.92. The predicted daily total irradiance for tomorrow (the second time period) is 5 kWh / m², and the daily average temperature is 25°C. Using a simple physical formula model, combined with the power station's installed capacity, the initial power prediction value is calculated to be 8500 kWh. According to the rule, for t=12 years (≤14 years), the dynamic weighting factor is set to... Calculate the target power prediction value: Substituting into the correction formula, the target power prediction value = 0.6×8500+(1−0.6)×(0.92×8500)=5100+0.4×7820=5100+3128=8228 kWh. In this example, the final prediction value of 8228 kWh is lower than the initial prediction value of 8500 kWh, reflecting the impact of the attenuation correction; at the same time, since the weighting factor is 0.6, the correction process still retains a considerable proportion (60%) of the basic meteorological forecast results, which is consistent with the degree of attenuation impact of the power plant at this operating age.

[0051] In this embodiment of the application, the target degradation coefficient of the photovoltaic power station is calculated based on operating data and environmental data, including:

[0052] Step A1: Calculate the aging degradation coefficient of the photovoltaic power station based on the operating data.

[0053] In this embodiment of the application, the aging degradation coefficient of the photovoltaic power station is calculated based on operating data, including:

[0054] Step A101: Extract the operating years of the power station and the annual average attenuation coefficient from the operating data.

[0055] Specifically, the raw operating data is analyzed and extracted as follows: First, the operating years of the power station are extracted, a scalar value characterizing the aging time of photovoltaic modules. This is usually obtained by subtracting the commissioning year from the current year by consulting the power station's commissioning records or grid connection documents, and it directly serves as the time variable in the aging model. Second, the annual average degradation coefficient is extracted, a key parameter characterizing the annual performance degradation rate of photovoltaic modules under specific operating conditions. Its value reflects the accelerating effect of local climate conditions (such as temperature, humidity, and UV intensity) on material aging. Extraction methods can be based on prior knowledge, such as consulting experience comparison tables to obtain empirical values ​​according to the climate type of the power station's location (e.g., arid and hot regions or mild and humid regions); or it can be obtained by fitting the rated power of the power station in its first year with the rated power in the current year. That is, by using historical performance records in the operating data, comparing the rated output capacity in the initial commissioning period (first year) with the recent rated output capacity (current year), and combining this with the operating years, the average annual degradation rate value can be derived. This coefficient is the basis for subsequently constructing a nonlinear aging degradation model and quantifying the performance loss accumulated over time.

[0056] Furthermore, the service life can be calculated annually or precisely to the month or day to more precisely reflect the aging process. For power plants that have undergone component replacement or large-scale maintenance, the concept of equivalent service life can be introduced, using a weighted average based on the commissioning time of different batches of components. The annual average degradation coefficient k is a parameter that is finely adjusted with the operational stage. For example, in the early stages of power plant operation (such as the first 5 years), the standard annual degradation rate provided by the component manufacturer is used; in the middle and later stages of operation, the actual power monitoring data of the power plant in recent years is used, and a sliding window approach is employed to perform linear or nonlinear regression to refit the data. The value is adjusted to better reflect the actual aging state of the component.

[0057] Step A102: Obtain the property parameters of the photovoltaic panels in the photovoltaic power station, including material property parameters and shape property parameters.

[0058] Specifically, two core attribute parameters need to be identified and obtained. The first is the material characteristic parameter (usually denoted as λ), a dimensionless constant used to characterize the inherent ability of photovoltaic panel encapsulation materials (such as EVA film and backsheet) and the cells themselves to resist performance degradation caused by environmental stresses (such as heat, humidity, and ultraviolet radiation). A larger value indicates stronger anti-aging performance and less aging degradation over the same operating years. The second is the shape characteristic parameter (usually denoted as μ), also a dimensionless constant, used to characterize the nonlinear curve shape of the module's output degradation over operating time. It reflects that the aging process is not a simple linear relationship but may include initial stabilization periods, acceleration periods, etc. A larger value indicates more significant nonlinear characteristics of the degradation process. Typical methods for obtaining these parameters include: directly searching from the product manual or test report corresponding to the photovoltaic module; and analyzing historical operating data by statistically analyzing a large number (e.g., more than 50) of measured degradation data from similar old photovoltaic power plants with different operating years, using parameter estimation methods (such as nonlinear least squares) to fit the aging degradation model, thereby deriving typical parameters applicable to this type of module. and Values; refer to empirical values ​​from databases of similar material components or expert knowledge bases.

[0059] It should be noted that if precise parameters for a specific module model are unavailable, statistical averages from publicly available long-term aging studies of that module category (e.g., monocrystalline PERC, polycrystalline) and encapsulation material type (e.g., double-glass, standard backsheet) can be used as a substitute. Regarding shape characteristic parameters... In addition to fitting statistical data across power plants, historical performance data from a longer time series (such as from commissioning to the present) of the power plant can also be used to estimate the aging trend by fitting its aging trend line, making the parameters more consistent with the actual aging process of the power plant. In a more refined model, the material property parameter λ can be considered as a variable that changes slightly with time or cumulative irradiation dose, but for the sake of simplifying the calculation, it can be treated as a constant value.

[0060] Step A103: Calculate the aging degradation coefficient of the photovoltaic power station based on the power station's operating years, annual average degradation coefficient, material characteristic parameters, and shape characteristic parameters.

[0061] Specifically, the four key parameters (power plant operating years) will be used to... Annual average attenuation coefficient Material property parameters and shape characteristic parameters Substituting the values ​​into a pre-defined nonlinear mathematical model, the proportion of photoelectric conversion efficiency loss due to material aging in the component is calculated, i.e., the aging degradation coefficient (usually denoted as ). The implementation involves substituting the four parameters mentioned above as input variables into a decay function constructed based on Arrhenius aging theory and a large amount of historical measured data. The typical form of this function is:

[0062]

[0063] in, The aging degradation coefficient after t years of operation, with a value range of 0-0.4, is dimensionless and represents the proportion of output degradation caused by aging of the component. These are material property parameters, fundamental parameters characterizing the anti-aging ability of materials. The larger the value, the stronger the anti-aging performance of the material. This item integrates the environmental acceleration effect and the time accumulation effect; The annual average attenuation coefficient is 0.02 in arid and high-temperature regions due to accelerated material aging, and 0.015 in temperate and humid regions. The operational lifespan of the facility is obtained directly from the facility's commissioning records. The shape characteristic parameters (e.g., convexity) of the decay curve are controlled by the exponent.

[0064] Among them, the annual average attenuation coefficient The equivalent rated power of a photovoltaic power plant in its first year can be obtained by fitting the rated power of the current year to the rated power of the photovoltaic power plant, as shown in the following formula:

[0065]

[0066] The annual average attenuation coefficient, The rated power of the power plant under standard test conditions during its first year of operation; This is the equivalent rated power assessed under the same conditions in the current year. It is determined by fitting measured aging degradation data from more than 50 aging power stations with different operating years, and is used to characterize the nonlinearity of aging degradation. The larger the value, the more significant the nonlinear characteristics of aging degradation.

[0067] Calculated using this formula It is a dimensionless value between 0 and a certain upper limit (such as 0.4), and its physical meaning is directly expressed as: the proportion of rated output capacity retained by the photovoltaic module due to the aging of its own materials after t years of operation (for example, a result of 0.88 means that 88% of the original capacity is retained, and 12% has been degraded). This coefficient accurately describes the nonlinear change of aging degradation with the years.

[0068] It should be noted that when data or computing resources are limited, a simplified version or empirical formula variation of this nonlinear model can be used. For example, when the shape parameter μ approaches 1, the model can be simplified to an approximate exponential or linear decay model for calculation, sacrificing some nonlinear accuracy in exchange for computational speed. If the power plant contains components of different models or batches, and their attribute parameters ( , Given that the aging degradation coefficients of various components can be calculated separately, and then a weighted average can be performed based on their installed capacity ratio to obtain the comprehensive aging degradation coefficient of the entire power plant. When the material characteristic parameter λ or shape parameter... When there is a certain range of uncertainty, sensitivity analysis can be performed. For example, optimistic and pessimistic values ​​can be used to calculate the aging degradation coefficient to obtain a possible degradation range, which can provide a reference for risk assessment.

[0069] By extracting the power plant's operating years and annual average degradation coefficient from operational data, key time and rate parameters reflecting aging duration and the accelerated effects of local climate can be directly obtained, providing accurate input for aging modeling. By obtaining the material and shape characteristics of photovoltaic panels, the inherent anti-aging capabilities and nonlinear characteristics of the degradation process of the modules themselves can be introduced, making the model more physically meaningful and targeted. By calculating the aging degradation coefficient based on the above-mentioned years, degradation coefficient, and attribute parameters, the nonlinear model accurately depicts the nonlinear degradation law of output as the years of operation increase, realizing a leap from simple linear estimation to accurate nonlinear fitting of the aging process.

[0070] Step A2: Calculate the dust accumulation attenuation coefficient of the photovoltaic power station based on environmental data.

[0071] In this embodiment of the application, the calculation of the dust accumulation attenuation coefficient of the photovoltaic power station based on environmental data includes:

[0072] Step A201: Extract the daily average rainfall within the first time period from the environmental data.

[0073] Specifically, from environmental data (usually meteorological time-series data including rainfall), for a specific time period reflecting recent dust accumulation and cleanliness (e.g., the last 7 days in engineering practice), daily rainfall data within this period is extracted, and the arithmetic mean of these data is calculated to obtain the daily average rainfall (usually denoted as d). This value directly characterizes the frequency and intensity of natural rainfall cleaning dust from the photovoltaic module surface in the period preceding the assessment time: the higher the daily average rainfall, the stronger the recent rainwater cleaning effect and the less dust accumulates on the module surface; conversely, it means that dust accumulation may be more severe. Therefore, this parameter is a key driving variable for subsequently constructing a "rainfall-dust accumulation attenuation" correlation model and ultimately quantifying the proportion of photovoltaic output loss caused by dust accumulation. The extraction process typically involves data querying, time window alignment, and simple arithmetic mean calculation.

[0074] It should be noted that the length of the first time period can be adjusted according to the local dust deposition rate and climate characteristics. In windy, sandy, and arid areas, dust accumulates quickly, so the period can be shortened (e.g., the last 5 days) to more sensitively reflect changes in dust accumulation. In humid and rainy areas, the period can be extended (e.g., the last 10 days) to smooth the impact of a single rainfall event and obtain a more stable trend. When extracting daily rainfall, potential data gaps or anomalies need to be addressed. For example, if data for a particular day is completely missing, values ​​from the preceding and following days can be used for interpolation, or that day can be directly removed and the base for calculating the average value adjusted. Significantly unreasonable extreme values ​​(e.g., excessively large values ​​due to instrument errors) should be identified and corrected. Rainfall data in environmental data can come from a single data source (e.g., on-site rain gauges at power plants) or can be fused from multiple data sources (e.g., data from nearby meteorological stations, satellite remote sensing precipitation products) to improve data reliability and spatial representativeness. When using fused data, daily rainfall can be the average of multiple sources or a quality-weighted value.

[0075] Step A202: Determine the lower limit of the coefficient based on the range of values ​​corresponding to the ash attenuation coefficient, and obtain the maximum ash attenuation rate corresponding to the lower limit of the coefficient.

[0076] Specifically, based on the physical laws and engineering experience regarding the output degradation of photovoltaic modules due to dust accumulation, the theoretical range of the dust accumulation degradation coefficient is defined, and the lower limit of this range is determined accordingly. Then, the maximum dust accumulation degradation rate corresponding to this lower limit is calculated. First, based on statistical analysis of a large amount of photovoltaic power plant operation data and consensus on engineering practice, a typical range for the dust accumulation degradation coefficient (representing the output retention ratio) is preset, for example, from 0.85 to 1.0. The lower limit of the coefficient (e.g., 0.85) is a key constant, representing the minimum proportion of output that the photovoltaic module can retain when there is no effective rainfall for cleaning over a long period and dust accumulation reaches an observable stable or saturated state. Second, the maximum dust accumulation degradation rate is another key constant directly derived from this lower limit of the coefficient. Its calculation formula is: Maximum dust accumulation degradation rate = 1 - Lower limit of the coefficient. For example, when the lower limit of the coefficient is 0.85, the corresponding maximum dust accumulation degradation rate is 1 - 0.85 = 0.15 (or 15%). This 15% represents the theoretical upper limit of power loss under the worst-case ash accumulation conditions, as accepted by the model. These two parameters together define the computational boundary of the ash accumulation attenuation model, ensuring that subsequent linear interpolation calculations based on rainfall are constrained within a reasonable physical range.

[0077] It should be noted that the value range and lower limit of the coefficient are not absolutely fixed and can be fine-tuned according to the specific environment of the photovoltaic power station location (such as wind and sand intensity, dust type, and module tilt angle). For example, in heavy industrial areas or on the edge of deserts, where dust accumulates faster and thicker, the lower limit of the coefficient may be set to 0.80, corresponding to a maximum attenuation rate of 20%; while in humid areas with clean air, it can be set to 0.90, corresponding to a maximum attenuation rate of 10%.

[0078] For a specific power station, its long-term historical data can be analyzed. Several typical long-term rainless periods (such as more than 30 consecutive days without rainfall) are selected, and the relative output of photovoltaic modules (relative to the clean state) at the end of these periods is statistically analyzed. The statistical lower limit of these relative outputs (e.g., the 5th percentile) is taken as an empirical estimate of the actual lower limit of the coefficients for the power station, and the corresponding maximum degradation rate is calculated accordingly, so that the model parameters are more in line with the actual situation of the power station.

[0079] Step A203: Calculate the dust accumulation attenuation coefficient of the photovoltaic power station based on the daily average rainfall, the lower limit of the coefficient, and the maximum dust accumulation attenuation rate.

[0080] Specifically, a pre-defined calculation model quantifies the proportion of photovoltaic power output degradation caused by dust accumulation on the component surface, i.e., the dust accumulation degradation coefficient (denoted as ). The implementation method is to use the daily average rainfall d (reflecting the recent intensity of natural cleansing) and the lower limit of the coefficient ( (e.g., 0.85, corresponding to the output retention ratio when there is no rainfall) and the maximum attenuation rate of ash accumulation ( The average daily rainfall (e.g., 0.15, or 15%) is used as input and substituted into a mathematical model that reflects the linear relationship between rainfall, dust accumulation, and attenuation rate. The core logic of this model is: the larger the average daily rainfall (d), the stronger the recent rainwater washing effect and the less dust accumulates; therefore, the calculated dust attenuation coefficient should be closer to 1 (indicating no attenuation). Conversely, when d is smaller or even 0, the calculated coefficient should approach the lower limit of the coefficient. At this point, the output attenuation reaches the maximum value set by the model. The calculation formula is as follows:

[0081]

[0082] in, The ash accumulation attenuation coefficient ranges from 0.85 to 1.0, representing the proportion of power output retained by the module due to ash accumulation. The lower limit of 0.85 corresponds to a maximum attenuation rate of 15% due to ash accumulation, which is a common limit value for long-term ash accumulation attenuation in engineering practice without rainfall. d represents the average daily rainfall over the past 7 days. Through this calculation, a value between the lower limit of the coefficient (e.g., 0.85) and 1.0 is obtained, directly representing the proportion of photovoltaic module power output retained due to ash accumulation under the current rainfall conditions.

[0083] It should be noted that a rainfall saturation threshold (e.g., 10 mm / day) can be defined. When d is greater than or equal to this threshold, linear calculations are no longer performed; instead, it is directly assumed that the accumulated dust has been effectively removed, and the dust attenuation coefficient is set to 1.0, simplifying calculations under extreme rainfall conditions. Besides linear models, nonlinear functions (such as exponential or logarithmic relationships) can also be used to describe the relationship between rainfall and cleaning effectiveness, to more accurately fit the phenomenon observed in some areas where "initial rainfall has high cleaning efficiency, followed by decreasing efficiency." (Coefficient lower limit value) and maximum attenuation rate It doesn't have to be a globally fixed value; instead, it can be fine-tuned based on the season or the local prevailing wind direction (which affects the source of dust). For example, during dry, windy seasons, a lower value can be used. Values ​​(such as 0.82) are used to reflect more severe potential dust accumulation effects.

[0084] By extracting the daily average rainfall within the first time period from environmental data, key dynamic environmental indicators that directly reflect the recent natural cleaning intensity and inversely characterize the degree of dust accumulation can be obtained. By determining the lower limit of the dust accumulation attenuation coefficient and the corresponding maximum attenuation rate based on the range of the dust accumulation attenuation coefficient, reasonable attenuation boundaries can be set according to engineering practice to ensure the physical rationality of the model calculation. By calculating the dust accumulation attenuation coefficient based on the daily average rainfall, the lower limit of the coefficient, and the maximum attenuation rate, a linear correlation model between rainfall and dust accumulation attenuation is constructed, realizing a rapid, simple, and effective dynamic quantification of the power output loss caused by dust accumulation.

[0085] Step A3: Calculate the interaction effect coefficient between aging characteristics and dust accumulation characteristics in the photovoltaic power station based on the operating data and environmental data.

[0086] In this embodiment of the application, the interaction effect coefficient between aging characteristics and dust accumulation characteristics in a photovoltaic power station is calculated based on operational data and environmental data, including:

[0087] Step A301: Obtain the range of the increase in the dust adhesion rate of photovoltaic modules in the photovoltaic power station.

[0088] Specifically, through long-term observation and comparative experiments on similar aging photovoltaic power plants, the range of the increase in the dust accumulation rate of photovoltaic modules was obtained. This range specifically refers to the fact that, compared to brand-new or well-maintained photovoltaic modules, aging modules, due to factors such as yellowing of their encapsulation materials leading to increased surface hydrophilicity, make dust adhesion easier and more robust, resulting in a significant increase in the dust accumulation rate per unit time (i.e., the speed at which dust accumulates on the module surface). This range of increases is a parameter quantifying the degree of this increase, usually expressed as a percentage range, such as "30%-50%". This numerical range indicates that the dust accumulation rate of older modules is 30% to 50% faster than that of newer modules, which is one of the core bases for quantifying the directional impact of "aging exacerbates dust accumulation" in the interaction effect model. Obtaining this parameter provides a benchmark for determining the specific upper and lower limits of the proportion.

[0089] It should be noted that for specific power plants or regions, measured data on dust deposition or power output degradation of modules with different operating years under similar environmental conditions (such as continuous rainless periods) can be collected. By comparing the dust accumulation of new modules (short operating years) with that of older modules (long operating years), the distribution range of the adhesion rate improvement ratio can be statistically calculated. The improvement ratio range can be refined according to the photovoltaic module encapsulation type (such as glass-backsheet, double-glass) and the climate zone of the power plant location (such as arid and dusty areas, humid and low-dust areas). For example, in arid and dusty areas, changes in surface hydrophobicity may have a more significant impact on dust accumulation, and the improvement ratio range may be set at 40%-60%; while in humid areas, this range may be adjusted to 20%-40%. Publicly available experimental data or consensus values ​​from the photovoltaic industry regarding the impact of module aging on dust accumulation characteristics can be directly cited as this ratio range.

[0090] Step A302: Determine the upper limit and lower limit of the ratio based on the range of the increase ratio.

[0091] Specifically, the obtained range of increases in the dust accumulation rate of photovoltaic modules (a percentage range, e.g., 30%-50%) is analyzed, and the two boundary values ​​of this range are directly determined as the upper and lower limits of the required proportion for calculation. Typically, the maximum value of this range (e.g., 50%) is determined as the upper limit, representing the maximum intensity of the interaction effect; the minimum value of this range (e.g., 30%) is determined as the lower limit, representing the baseline or minimum intensity of the interaction effect. These two values ​​will be used as key parameters in constructing the mathematical model of the interaction effect coefficient, enabling the model to dynamically calculate the specific interaction effect coefficient within this intensity range using variables such as the number of years of operation and rainfall, thereby quantifying the degree of synergistic effect between aging and dust accumulation. For example, in one implementation, the upper limit might be used to calculate the coefficient term of the interaction effect increasing with the number of years, and the lower limit might be used to calculate the coefficient term related to rainfall.

[0092] It should be noted that if the improvement ratio range is based on a large amount of statistical data and the data distribution shows a central tendency, a certain statistic (such as the median or mean) within that range can be taken as a single ratio value. The upper and lower limits of the ratio can then be redefined by fluctuating around this center by a certain percentage (such as ±10%) to better represent typical cases. The degree of aging (e.g., segmented by years of operation or yellowing level) can be correlated with different sub-ranges of improvement ratios. For example, mild aging (e.g., t≤8 years) corresponds to a ratio range of 20%-35%, with upper and lower limits of 35% and 20% respectively; moderate to severe aging (t>8 years) corresponds to a ratio range of 35%-50%, with corresponding upper and lower limits of 50% and 35% respectively. This allows for a more precise link between the intensity of the interaction effect and the degree of aging. In areas with strong winds and abundant dust sources, the base rate of dust accumulation is already high, and the improvement ratio due to aging may be more significant. Therefore, a larger ratio range (e.g., 40%-60%) can be used, and new upper and lower limits can be determined accordingly.

[0093] Step A303: Calculate the interaction effect coefficient based on the upper limit of the ratio, the lower limit of the ratio, the operating years of the power plant in the operating data, and the daily average rainfall in the environmental data.

[0094] Specifically, the upper limit of the proportion ( ) and the lower limit of the proportion ( The interaction coefficient (usually denoted as ), along with the power plant's operating years (t) and average daily rainfall (d), is substituted into a pre-defined interaction coefficient model to calculate the interaction coefficient. The model can be pre-set with fixed initial coefficients (e.g., constants a, b, and c in the model formula). In an optional embodiment, the model aims to mathematically characterize two physical processes: the deeper the aging (characterized by the number of years t), the stronger the hydrophilicity of the component surface, and the greater the increase in the dust accumulation rate (reflected by the upper and lower limits of the ratio), resulting in a stronger interaction effect; the less recent rainfall (characterized by the daily average rainfall d), the heavier the dust accumulation, and the greater the synergistic loss caused by the superposition with aging. This is achieved by converting the upper and lower limits of the ratio into coefficients in the model; for example, setting the coefficient for the number of years t as a quantity related to the upper limit of the ratio (e.g., ...). (Approximation), setting the coefficients of rainfall-related terms to quantities related to the lower limit of the proportion (e.g.) (approximation), and then construct the formula in linear summation form:

[0095]

[0096] Where a, b, and c are constants determined based on the proportional range and physical meaning. Through this calculation, a value greater than or equal to 1 is finally obtained. This value is used as the multiplier of the product of the aging attenuation coefficient and the ash accumulation attenuation coefficient, and is used to quantify the proportion of additional power loss caused by the combined effect of the two.

[0097] It should be noted that the interaction effect coefficient model is not necessarily limited to a linear form. For example, a model can be constructed that includes an interaction term (e.g., t × (1 / d)) involving the operating years t and rainfall d to more precisely characterize the synergistic amplification effect of the two under extreme conditions, but attention should be paid to the increased model complexity. The constants a, b, and c in the model (e.g., 0.03, 0.05, and 10 in the example above) do not need to be fixed values, but rather calibrated using regression analysis based on historical data of the climate characteristics (e.g., average rainfall, dust type) and component aging characteristics of the power plant area, making the model more closely reflect local realities. Different calculation formulas or coefficients are used depending on the different intervals of the operating years t or the average daily rainfall d. For example, when t is less than a certain threshold, the interaction effect is considered weak, and a smaller value of a is used; when t exceeds the threshold, a larger value of a is used to reflect the abrupt change in the interaction effect after accelerated aging.

[0098] As an example, suppose we obtain the upper limit value of the ratio. The lower limit of the ratio The operating data reveals the power plant's service life. Annual average daily rainfall extracted from environmental data Millimeters. Formula used: , where 0.03 and 0.05 are constants determined empirically based on the proportional range. Therefore, the calculated photovoltaic panel aging-dust accumulation interaction effect coefficient is 1.34. This value is greater than 1, indicating the existence of a synergistic enhancement effect, which will increase the total attenuation coefficient, i.e., an additional synergistic loss of about 34% (relative to independent action).

[0099] By obtaining the range of the increase in the dust adhesion rate of photovoltaic modules, we can quantitatively understand the physical phenomenon that changes in the surface characteristics of aging modules lead to easier dust adhesion, providing a basis for quantifying the interaction effect. By determining the upper and lower limits of the increase range, we can transform the qualitative influence range into specific parameter boundaries usable by the model, which is convenient for mathematical modeling. By calculating the interaction effect coefficient based on the upper and lower limits of the increase range, the number of years of operation, and the average daily rainfall, we constructed an enhancement factor model that couples the degree of aging (number of years) and the degree of dust accumulation (rainfall), which scientifically characterizes and quantifies the synergistic loss effect of mutual aggravation between aging and dust accumulation.

[0100] Step A4: The aging attenuation coefficient, dust accumulation attenuation coefficient, and interaction effect coefficient are fused to obtain the target attenuation coefficient of the photovoltaic power station.

[0101] Specifically, three independent coefficients were calculated: aging degradation coefficient ( ), dust accumulation attenuation coefficient ( ) and interaction effect coefficient ( Through mathematical fusion rules, it is integrated into a comprehensive target attenuation coefficient (denoted as ). This coefficient is used to comprehensively characterize the total output degradation ratio of a photovoltaic power station under its current aging level, dust accumulation status, and the combined effect of both. Its core implementation method is to use a multiplicative fusion model, that is, directly multiplying the three coefficients together:

[0102]

[0103] in, The target attenuation coefficient; This is the aging degradation coefficient; This is the dust accumulation attenuation coefficient; This represents the interaction effect coefficient.

[0104] In this model, and As coefficients less than or equal to 1, the output retention ratios due to aging and dust accumulation were independently quantified; and A multiplicative amplification factor greater than or equal to 1 is used to quantify the additional power loss multiplication effect resulting from the interaction between the two. Through this multiplicative coupling method, the final result is... It is a value between 0.6 and 1.0, and its physical meaning represents the current operating life. and rainfall conditions This coefficient represents the proportion of the actual power output of a photovoltaic module, under the combined effects of aging, dust accumulation, and their interactions, relative to its original (clean, unaged) theoretical output. This coefficient will be directly used for power correction calculations in subsequent dynamic weighted prediction models.

[0105] It should be noted that, based on multiplicative fusion, weighting factors can be introduced to adjust the relative importance of different attenuation factors in the comprehensive evaluation. For example, an exponential weighting method can be used: ,in , , The weights are determined based on experience or data fitting, and typically satisfy... This ensures consistency in dimensions. This allows the model to flexibly adjust the contribution of each attenuation factor based on different power plant types or climate zone characteristics.

[0106] If the power plant has other known significant attenuation factors (such as slight shading, localized losses due to hot spot effects, etc.), their corresponding attenuation coefficients can also be included in the multiplicative fusion model to more comprehensively reflect the total attenuation. If the three coefficients have large differences in their range or dimensions, appropriate normalization or standardization can be performed before fusion to make them comparable in magnitude, and then multiplication or other fusion operations can be performed to ensure the stability and rationality of the fusion result.

[0107] By calculating the aging attenuation coefficient based on operational data, the long-term performance degradation of photovoltaic modules caused by natural material aging can be quantified, providing a basic attenuation benchmark for prediction. By calculating the dust accumulation attenuation coefficient based on environmental data, the cleaning effect of recent rainfall on the dust on the module surface can be dynamically reflected, quantifying the output fluctuations caused by short-term environmental factors. By calculating the interaction effect coefficient between aging characteristics and dust accumulation characteristics based on operational and environmental data, the additional power loss caused by the synergistic effect of aging aggravating dust accumulation and dust accumulation accelerating aging is quantified for the first time, breaking through the limitations of single-factor correction. By fusing the three coefficients of aging, dust accumulation, and interaction effect, a comprehensive target attenuation coefficient is obtained, which comprehensively and coupledly characterizes the multiple attenuation mechanisms affecting the output of aging power plants, providing accurate attenuation input for subsequent high-precision correction.

[0108] In this embodiment of the application, the initial power prediction value of the photovoltaic power station is corrected based on the target attenuation coefficient to obtain the target power prediction value, including:

[0109] Step B1: Obtain meteorological data for the photovoltaic power station during the second time period.

[0110] Specifically, when power forecasting is required, meteorological data for the location of the photovoltaic power plant within a second time period is obtained from external weather forecasting services or databases. The second time period specifically refers to the future time period to be predicted, such as the next 24 hours of the current forecast day, or the morning to afternoon of the forecast day. The acquired meteorological data consists of meteorological elements directly related to photovoltaic power generation output, primarily including: daily total irradiance or hourly irradiance, which is the most direct factor affecting power generation; daily average temperature or hourly temperature, which affects the conversion efficiency of photovoltaic modules; and other auxiliary factors, such as wind speed (affecting module heat dissipation and dust deposition), relative humidity, etc. This forecast meteorological data is the necessary input for calculations using basic "meteorology-power" correlation models (such as physical formulas, simple statistical models, or historical fitting models) to derive initial power forecast values ​​that do not consider the specific degradation state of the power plant. Data acquisition methods typically include calling meteorological APIs, accessing data interfaces of public weather forecasting websites, or using the output of numerical weather prediction models.

[0111] It should be noted that, depending on the forecasting requirements, the second time period can be several hours into the future (ultra-short-term forecast), 1-3 days into the future (short-term forecast), or even longer. Correspondingly, the temporal resolution of the meteorological data must also be matched. For example, ultra-short-term forecasts may require 15-minute data for the next 1-6 hours, while daily forecasts typically require daily totals or daily averages. For power plants with special geographical locations or complex terrain, single gridded forecast data may contain biases. Data fusion techniques can be used, such as combining global forecast models, regional mesoscale models, and station observation data for calibration, or using statistical downscaling methods to refine coarse-resolution forecast data to the power plant location, to improve the spatial accuracy of the forecast data. The acquired forecast data should undergo a reasonableness check (e.g., non-negative irradiance, temperature within a reasonable range). Furthermore, data from multiple forecast sources or ensemble forecast data can be acquired simultaneously to assess the range of forecast uncertainty and provide possible error intervals for subsequent forecasts.

[0112] Step B2: Calculate the initial power prediction value of the photovoltaic power station based on meteorological data.

[0113] Specifically, key meteorological inputs (such as daily total irradiance and daily average temperature) are substituted into a mathematical model describing the standard correlation between "meteorology and power." The basic model is typically a physical formula model or a lightweight statistical model (such as linear regression) trained on historical operating data of the power plant (especially data from its early operating stages). For example, a physical formula model might directly calculate theoretical output based on the rated power, conversion efficiency, and temperature coefficient of photovoltaic modules, combined with forecasted irradiance and temperature; a statistical model might establish a regression relationship between meteorological factors and power generation. This calculation yields a baseline value for future power prediction of the photovoltaic power plant under ideal or standard module conditions (without considering specific aging and dust accumulation), i.e., the initial power prediction value. This value serves as a benchmark for subsequent dynamic weight adjustments, providing a power prediction reference purely based on meteorological conditions.

[0114] It should be noted that the choice of base model can be diverse. Besides simple linear models, model chains from open-source libraries such as PVlib, which are based on physical principles, can be used for simulation. These chains consider factors such as solar position, module tilt angle, conversion efficiency, and temperature loss, thus obtaining initial predictions closer to theoretical values. For power plants with a long data history, a time series model (such as ARIMA) or a machine learning model (such as support vector machines) can be trained using data from the first few years after commissioning. However, it is essential to ensure that the model complexity is low to meet lightweight requirements.

[0115] If the meteorological data is hourly, but the base model requires daily totals or averages, then corresponding aggregation calculations are necessary (e.g., integrating irradiance to obtain the total daily irradiance, or averaging temperature). Simultaneously, the input meteorological data should be validated for reasonableness, such as removing negative irradiance values ​​or abnormal temperature values.

[0116] The parameters required for the physical formula model (such as component rated power and efficiency temperature coefficient) are obtained from the power plant design documents or component specifications. The parameters for the statistical model are determined by fitting historical data.

[0117] Step B3: Determine the dynamic weighting factor based on the operating years of the power plant in the operating data.

[0118] Specifically, extract the power plant's operating years from the operational data. Then, based on the preset piecewise or continuous function rules, the running years are... The mapping is a numerical value between 0 and 1, i.e., a dynamic weighting factor (denoted as ). The principle behind this rule is that a shorter operating life indicates less component aging and less impact from degradation; therefore, the initial power forecast value based on meteorological data should be given a higher weight, i.e., a dynamic weighting factor. A large value (e.g., close to 1) indicates severe component aging and significant degradation effects when the component has been in operation for a long time. In this case, the weight of the initial prediction should be reduced, and the weight of the correction term based on the target degradation coefficient should be increased, i.e., the dynamic weighting factor. The value is relatively small. The piecewise function rule could be: when the running years... When the period is less than or equal to 14 years, the dynamic weighting factor It decreases linearly with age, for example When the operating period t is greater than or equal to 14 years, the dynamic weighting factor... The weighting factor is stabilized at a low value (e.g., 0.3) to avoid the complete failure of the basic meteorological correlation due to excessively low weighting, thus ensuring the stability of the forecast. In this way, the dynamic weighting factor can adapt to the aging state of the power plant and achieve intelligent adjustment of the forecast benchmark while keeping the model lightweight.

[0119] Step B4: Correct the initial power prediction value using the dynamic weighting factor and the target attenuation coefficient to obtain the target power prediction value.

[0120] Specifically, in the dynamic weighting factor ( ) and target attenuation coefficient ( ) applied to the initial power prediction value ( The target power prediction value is obtained by correcting the result through a preset weighted fusion formula, which is more in line with the current attenuation state of the power plant and has higher accuracy. Its core implementation method is to use a linear weighted model: In this model, the first term This represents the original forecast contribution based on fundamental meteorological associations, and its weight is determined by a dynamic weighting factor. Decision; Second item This represents the predicted contribution after reduction by a comprehensive attenuation coefficient, and its weight is... This formula allows for the determination of the aging level of the power plant (based on...). This reflects a dynamic balance between the proportions of "uncorrected weather forecasts" and "forecasts after full attenuation correction" in the final result. When the power plant is relatively new ( When the weather is relatively large, the final forecast is closer to the original weather forecast; when the power station is old ( When the value is relatively small, the final prediction will approach the attenuation-corrected value more closely. This approach significantly improves the adaptability of power prediction for aging power plants to actual attenuation conditions while ensuring the model is lightweight and computationally simple.

[0121] By acquiring meteorological data from photovoltaic power plants during the second time period, key meteorological conditions for future periods can be input for power prediction, laying the physical foundation for the prediction. By calculating the initial power prediction value of photovoltaic power plants based on meteorological data, a baseline prediction that does not consider specific degradation states can be generated using a basic meteorological-power correlation model, serving as the starting point for correction. By determining dynamic weighting factors based on the operating years of the power plants in the operational data, the prediction strategy can be adaptively adjusted according to the degree of power plant aging, achieving an intelligent balance between the baseline prediction and degradation correction. By using dynamic weighting factors and target degradation coefficients to correct the initial power prediction value, the final prediction result can retain the rationality of meteorological correlation while accurately reflecting the actual degradation state of the power plant within a lightweight computational framework, significantly improving prediction accuracy.

[0122] In the embodiments of this application, such as Figure 2 As shown, the method also includes:

[0123] Step S201: Obtain the measured power value of the photovoltaic power station during the third time period.

[0124] In this embodiment, the actual power generation measurement sequence, i.e., the measured power values, is extracted from the local monitoring system or Supervisory Control and Data Acquisition (SCADA) system of the photovoltaic power plant within a specific historical period, namely the third time period (e.g., the most recent month). This third time period is a past time window with complete measured data used for model validation and parameter optimization. The acquired measured power values ​​are typically high-temporal-resolution (e.g., continuous data every 15 minutes or hour), reflecting the actual power output of the power plant at each moment within this period. These data will serve as a standard for comparison with subsequent model predictions for the same period, thereby calculating the prediction error and providing a basis for iterative optimization of key model parameters (such as the annual average degradation coefficient). The acquisition process typically involves accessing the power plant's historical database, querying and exporting the corresponding power data records by time range.

[0125] Step S202: Obtain the power prediction value of the photovoltaic power station after correction based on the target attenuation coefficient within the third time period, and compare the measured power value with the power prediction value to obtain the comparison result.

[0126] In this embodiment, firstly, the power prediction value of the photovoltaic power station after correction based on the target attenuation coefficient within the third time period is obtained. That is, the third time period (the past verification period) is used as the simulation prediction object. Historical meteorological data corresponding to each day or time period within this period (used as input for the second time period) and the current state data (operating years calculated according to historical dates, environmental data using historical values) are re-inputted into the model. The target attenuation coefficient and dynamic weighting factor are calculated sequentially, and the initial power prediction value is corrected, thereby generating a power prediction value time series that completely corresponds to the third time period and considers attenuation. Secondly, the measured power value and the power prediction value are compared to obtain the comparison result. This is achieved by comparing the time series of measured power values ​​within the same third time period with the newly calculated power prediction value time series, under the premise of strict timestamp alignment, point-by-point or statistical comparison. The comparison result is usually presented in the form of quantitative error indicators, such as calculating the root mean square error (RMSE) and mean absolute error of both, where RMSE is a commonly used indicator for comprehensively evaluating prediction deviation. This comparison result directly reflects the degree of deviation between the prediction accuracy and the measured value under the current model parameters, providing a clear quantitative basis for subsequent model parameter optimization.

[0127] It should be noted that, in addition to RMSE, the comparison results can include various error statistics, such as mean absolute percentage error (MAPE), correlation coefficient (R²), and error distribution statistics (e.g., histograms, quantiles), thus comprehensively evaluating the model's predictive performance from different dimensions. The comparison results can also be analyzed, for example, calculating the prediction errors for sunny, cloudy, and rainy days separately, or comparing the errors during daytime power generation with those at night, to identify the model's predictive weaknesses under different weather conditions or operating states. For longer third time periods, batch processing programs can be developed to automatically and cyclically complete daily attenuation coefficient calculations and power predictions, and vectorized or parallel computing techniques can be used to improve processing efficiency.

[0128] Step S203: Based on the comparison results, iteratively optimize the annual average attenuation coefficient in the running data to obtain the optimized annual average attenuation coefficient.

[0129] In this embodiment, the adjustable parameters in the model, such as the annual average decay coefficient k in the operating data, are automatically fine-tuned using the comparison results (quantitative indicators of prediction error, such as root mean square error RMSE) to reduce prediction error and obtain an optimized annual average decay coefficient that better reflects the actual aging rate of the power plant. This can be achieved using an iterative optimization algorithm (such as gradient descent). Specifically, starting with the currently used k value, the prediction error of the model within the third time period (i.e., the comparison result) under the current k value is calculated. If the error (e.g., RMSE) exceeds a preset tolerance threshold (e.g., 10%), the k value is adjusted in the direction of error reduction according to the trend of error relative to k (gradient direction) with a fixed small step size (e.g., 0.001). (For example, if increasing k reduces the error, the k value is increased.) Then, the target decay coefficient is recalculated using the new k value to obtain a new comparison result. This process is repeated until the calculated error indicator (e.g., RMSE) drops below the threshold (e.g., ≤10%) or the maximum number of iterations is reached. The k value obtained at this point is the optimized annual average decay coefficient. Through iteration, the model can adaptively calibrate its representation of the accelerated aging effects of the local climate, making the aging degradation model more accurate.

[0130] It should be noted that, in addition to the basic gradient descent method, more efficient optimization algorithms, such as the conjugate gradient method or the Adam optimizer, can be used to accelerate convergence. Dynamic step size adjustments can also be set, using larger step sizes for rapid convergence when the error is large, and smaller step sizes for fine-tuning when the error approaches a threshold. The optimization object is not limited to the annual average decay coefficient. If the model allows and the data supports it, other parameters in the aging degradation model (such as shape characteristic parameters) can also be adjusted simultaneously. The coefficients in the model can be jointly optimized in a small manner to seek a better global solution. The termination condition can not only depend on whether the RMSE is less than a threshold, but can also be combined with other indicators, such as the error decrease being less than a certain minimum value for several consecutive iterations, or reaching a preset upper limit for the number of iterations, to avoid infinite loops.

[0131] As an example, assume the comparison result is RMSE = 12% (greater than the threshold of 10%). The currently used annual average decay coefficient k = 0.018. Perform iterative optimization: use gradient descent, setting the adjustment step size Δ = 0.001. First, it is necessary to estimate the gradient direction of the error (expressed as RMSE) with respect to k. This can be done using a perturbation method: calculate the RMSE (12%) when k = 0.018, and then calculate the RMSE obtained by rerunning the model when k = 0.019 (k + Δ) (assumed to be 11.5%). Since the RMSE decreases from 12% to 11.5%, it indicates that the error decreases in this direction (increasing k). Therefore, update the k value along the direction of error reduction: the new k = 0.018 + 0.001 = 0.019. Using k = 0.019, recalculate the target decay coefficient and power prediction value in the third time period, and perform the comparison operation again to obtain a new comparison result, assuming the RMSE has decreased to 10.5%. Since 10.5% is still greater than 10%, continue iterating. After re-estimating the gradient, it is assumed that the gradient continues to increase. Reducing the value to 0.020 lowers the RMSE to 9.8%. At this point, RMSE = 9.8% ≤ 10%, satisfying the termination condition. The iteration stops. Finally, the optimized annual average depreciation coefficient is determined to be k = 0.020. This value will replace the original 0.018 and be used for subsequent depreciation calculations, thus enabling the model to more accurately reflect the actual annual aging rate of the power plant.

[0132] Step S204: Update the target attenuation coefficient using the optimized annual average attenuation coefficient to obtain the updated target attenuation coefficient.

[0133] In this embodiment of the application, the optimized annual average attenuation coefficient (denoted as ) obtained through iterative optimization, which better reflects the actual aging rate of the power plant, is used. This is integrated into the attenuation calculation model to generate a more accurate updated target attenuation coefficient. The implementation method is: using... Replace the previously used annual average attenuation coefficient k, and then recalculate the aging attenuation coefficient using the same operating data (power plant operating years t) and property parameters (material characteristic parameter λ, shape characteristic parameter μ) through the aging attenuation coefficient calculation formula. Subsequently, this updated version... Compared with the calculated (unchanged under current environmental data) ash attenuation coefficient and the calculated interaction coefficient Together, they are fused again to obtain the updated target attenuation coefficient. This ensures that the core parameters of the degradation model can be calibrated based on the latest measured data, thereby enabling the entire power prediction model to dynamically track and adapt to changes in the aging rate of photovoltaic modules and maintain high-precision prediction capabilities over the long term.

[0134] It should be noted that while updating the annual average decay coefficient k, if the optimization process involves fine-tuning other parameters (such as shape characteristic parameters μ or coefficients in the interaction effect model), the latest values ​​of these parameters can also be substituted in to recalculate the aging decay coefficient and interaction effect coefficient, and then fused to achieve a comprehensive update of the model parameters. The triggering of model updates is not limited to a fixed monthly occurrence. More sensitive triggering conditions can be set; for example, when the error of multiple consecutive short-term (e.g., one week) validations exceeds a threshold, the optimization and update process can be triggered immediately; or, updates can be proactively performed during seasonal transitions or significant changes in climate characteristics. In addition to updating the annual average decay coefficient k, the parameters (including λ and μ) of the entire aging decay model can be refitted using the most recent longer period of running data (e.g., the last two years), thereby recalibrating the decay curve over a longer historical period, and then using the new aging decay coefficient to participate in the update of the target decay coefficient.

[0135] By obtaining the measured power values ​​of photovoltaic power plants during the third time period, reliable performance benchmark data can be provided for model verification and optimization. By obtaining the predicted power values ​​based on the target attenuation coefficient and comparing them, the prediction accuracy and error level of the current model can be objectively and quantitatively evaluated. By iteratively optimizing the annual average attenuation coefficient in the operating data based on the comparison results, the key parameters of the model can be adaptively calibrated using optimization algorithms to better match the actual aging rate of the power plant. By updating the target attenuation coefficient using the optimized annual average attenuation coefficient, dynamic closed-loop optimization and self-learning of model parameters can be achieved, ensuring that the prediction model can track changes in component performance over a long period of time and maintain the durability and robustness of prediction accuracy.

[0136] This embodiment also provides a photovoltaic power plant power prediction device, which is used to implement the above embodiments and preferred embodiments, and will not be repeated as already described. As used below, the term "module" can be a combination of software and / or hardware that implements a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0137] This embodiment provides a photovoltaic power plant power prediction device, such as... Figure 3 As shown, it includes:

[0138] The acquisition module 31 is used to acquire the operating data and environmental data of the photovoltaic power station;

[0139] The first coefficient calculation module 32 is used to calculate the aging degradation coefficient of the photovoltaic power station based on the operating data.

[0140] The second coefficient calculation module 33 is used to calculate the dust accumulation attenuation coefficient of the photovoltaic power station based on the environmental data.

[0141] The third coefficient calculation module 34 is used to calculate the interaction effect coefficient of aging characteristics and dust accumulation characteristics in the photovoltaic power station based on the operating data and the environmental data.

[0142] The fourth coefficient calculation module 35 is used to fuse the aging attenuation coefficient, the dust accumulation attenuation coefficient and the interaction effect coefficient to obtain the target attenuation coefficient of the photovoltaic power station.

[0143] The correction module 36 is used to correct the initial power prediction value of the photovoltaic power station based on the target attenuation coefficient to obtain the target power prediction value. The initial power prediction value is determined based on the meteorological data of the photovoltaic power station.

[0144] In this embodiment of the application, the correction module 33 is specifically used to acquire meteorological data of the photovoltaic power station in the second time period; calculate the initial power prediction value of the photovoltaic power station based on the meteorological data; determine the dynamic weighting factor based on the operating years of the power station in the operating data; and correct the initial power prediction value using the dynamic weighting factor and the target attenuation coefficient to obtain the target power prediction value.

[0145] In this embodiment of the application, the device further includes: an update module, configured to acquire the measured power value of the photovoltaic power station in the third time period; acquire the predicted power value of the photovoltaic power station after correction based on the target attenuation coefficient in the third time period, and compare the measured power value with the predicted power value to obtain a comparison result; iteratively optimize the annual average attenuation coefficient in the operating data based on the comparison result to obtain the optimized annual average attenuation coefficient; and update the target attenuation coefficient using the optimized annual average attenuation coefficient to obtain the updated target attenuation coefficient.

[0146] Please see Figure 4 , Figure 4 This is a schematic diagram of the structure of a computer device provided in an optional embodiment of the present invention, such as... Figure 4As shown, the computer device includes one or more processors 10, memory 20, and interfaces for connecting the components, including high-speed interfaces and low-speed interfaces. The components communicate with each other via different buses and can be mounted on a common motherboard or otherwise installed as needed. The processors can process instructions executed within the computer device, including instructions stored in or on memory to display graphical information of a GUI on external input / output devices (such as display devices coupled to the interfaces). In some alternative implementations, multiple processors and / or multiple buses can be used with multiple memories and multiple memory modules, if desired. Similarly, multiple computer devices can be connected, each providing some of the necessary operations (e.g., as a server array, a group of blade servers, or a multiprocessor system).

[0147] Processor 10 may be a central processing unit, a network processor, or a combination thereof. Processor 10 may further include a hardware chip. The hardware chip may be an application-specific integrated circuit (ASIC), a programmable logic device (PLD), or a combination thereof. The programmable logic device may be a complex programmable logic device (CAMP), a field-programmable gate array (FPGA), a general-purpose array logic (GDA), or any combination thereof.

[0148] The memory 20 stores instructions executable by at least one processor 10 to cause at least one processor 10 to perform the method shown in the above embodiments.

[0149] The memory 20 may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created based on the use of the computer device as shown by a landing page for an app. Furthermore, the memory 20 may include high-speed random access memory and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some alternative embodiments, the memory 20 may optionally include memory remotely located relative to the processor 10, which can be connected to the computer device via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0150] The memory 20 may include volatile memory, such as random access memory; the memory may also include non-volatile memory, such as flash memory, hard disk or solid-state drive; the memory 20 may also include a combination of the above types of memory.

[0151] The computer device also includes a communication interface 30 for communicating with other devices or communication networks.

[0152] This invention also provides a computer-readable storage medium. The methods described above according to embodiments of the invention can be implemented in hardware or firmware, or implemented as computer code that can be recorded on a storage medium, or implemented as computer code downloaded via a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and then stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code, which, when accessed and executed by the computer, processor, or hardware, implements the methods shown in the above embodiments.

[0153] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A method for predicting the power output of a photovoltaic power plant, characterized in that, The method includes: Obtain operational and environmental data from photovoltaic power plants; The aging degradation coefficient of the photovoltaic power station is calculated based on the operating data, wherein the aging degradation is calculated based on the operating years of the power station, the annual average degradation coefficient, and the attribute parameters of the photovoltaic panels in the photovoltaic power station. The dust accumulation attenuation coefficient of the photovoltaic power station is calculated based on the environmental data. The dust accumulation attenuation coefficient is calculated based on the average daily rainfall in the first time period of the environmental data, the lower limit of the coefficient corresponding to the range of the dust accumulation attenuation coefficient, and the maximum dust accumulation attenuation rate corresponding to the lower limit of the coefficient. The interaction effect coefficient between aging characteristics and dust accumulation characteristics in the photovoltaic power station is calculated based on the operating data and the environmental data. The aging attenuation coefficient, the dust accumulation attenuation coefficient, and the interaction effect coefficient are fused to obtain the target attenuation coefficient of the photovoltaic power station. The initial power prediction value of the photovoltaic power station is corrected based on the target attenuation coefficient to obtain the target power prediction value, which is determined based on the meteorological data of the photovoltaic power station.

2. The method according to claim 1, characterized in that, The calculation of the aging degradation coefficient of the photovoltaic power station based on the operating data includes: Extract the power plant's operating years and average annual degradation coefficient from the operational data; Obtain the property parameters of the photovoltaic panels in the photovoltaic power station, wherein the property parameters include material property parameters and shape property parameters; The aging degradation coefficient of the photovoltaic power station is calculated based on the operating years of the power station, the annual average degradation coefficient, the material characteristic parameters, and the shape characteristic parameters.

3. The method according to claim 1, characterized in that, The calculation of the dust accumulation attenuation coefficient of the photovoltaic power station based on the environmental data includes: Extract the daily average rainfall within the first time period from the environmental data; The lower limit of the coefficient is determined based on the range of values ​​corresponding to the ash attenuation coefficient, and the maximum ash attenuation rate corresponding to the lower limit of the coefficient is obtained. The ash accumulation attenuation coefficient of the photovoltaic power station is calculated based on the daily average rainfall, the lower limit of the coefficient, and the maximum ash accumulation attenuation rate.

4. The method according to claim 1, characterized in that, The calculation of the interaction effect coefficient between aging characteristics and dust accumulation characteristics in the photovoltaic power station based on the operating data and the environmental data includes: Obtain the range of the increase in the dust adhesion rate of photovoltaic modules in the photovoltaic power station; The upper and lower limits of the ratio are determined based on the aforementioned range of increases. The interaction effect coefficient is calculated based on the upper limit of the ratio, the lower limit of the ratio, the operating years of the power plant in the operating data, and the daily average rainfall in the environmental data.

5. The method as described in claim 4, characterized in that, The calculation of the interaction effect coefficient based on the upper limit of the ratio, the lower limit of the ratio, the operating years of the power plant in the operating data, and the daily average rainfall in the environmental data includes: The interaction effect coefficient is obtained by pre-setting the upper limit of the ratio, the lower limit of the ratio, the operating years of the power plant in the operating data, and the daily average rainfall in the environmental data into an interaction effect coefficient model. The preset interaction effect coefficient model has initial coefficients, which are calibrated using regression analysis based on the upper limit of the proportion, the lower limit of the proportion, the average rainfall, the dust type, and the number of years of operation.

6. The method according to claim 1, characterized in that, The step of correcting the initial power prediction value of the photovoltaic power station based on the target attenuation coefficient to obtain the target power prediction value includes: Obtain meteorological data of the photovoltaic power station during the second time period; The initial power forecast value of the photovoltaic power station is calculated based on the meteorological data. A dynamic weighting factor is determined based on the power plant's operating years in the aforementioned operational data; The initial power prediction value is corrected using the dynamic weighting factor and the target attenuation coefficient to obtain the target power prediction value.

7. The method according to claim 1, characterized in that, The method further includes: Obtain the measured power value of the photovoltaic power station during the third time period; The power prediction value of the photovoltaic power station after correction based on the target attenuation coefficient is obtained in the third time period, and the measured power value is compared with the power prediction value to obtain the comparison result; Based on the comparison results, the annual average attenuation coefficient in the running data is iteratively optimized to obtain the optimized annual average attenuation coefficient. The target attenuation coefficient is updated using the optimized annual average attenuation coefficient to obtain the updated target attenuation coefficient.

8. A photovoltaic power plant power prediction device, characterized in that, The device includes: The acquisition module is used to acquire operational data and environmental data of the photovoltaic power station; The first coefficient calculation module is used to calculate the aging degradation coefficient of the photovoltaic power station based on the operating data, wherein the aging degradation is calculated based on the operating years of the power station, the annual average degradation coefficient, and the attribute parameters of the photovoltaic panels in the photovoltaic power station. The second coefficient calculation module is used to calculate the dust accumulation attenuation coefficient of the photovoltaic power station based on the environmental data. The dust accumulation attenuation coefficient is calculated based on the daily average rainfall in the first time period in the environmental data, the lower limit of the coefficient corresponding to the range of the dust accumulation attenuation coefficient, and the maximum dust accumulation attenuation rate corresponding to the lower limit of the coefficient. The third coefficient calculation module is used to calculate the interaction effect coefficient between aging characteristics and dust accumulation characteristics in the photovoltaic power station based on the operating data and the environmental data. The fourth coefficient calculation module is used to fuse the aging attenuation coefficient, the dust accumulation attenuation coefficient, and the interaction effect coefficient to obtain the target attenuation coefficient of the photovoltaic power station. The correction module is used to correct the initial power prediction value of the photovoltaic power station based on the target attenuation coefficient to obtain the target power prediction value, wherein the initial power prediction value is determined based on the meteorological data of the photovoltaic power station.

9. A computer device, characterized in that, include: A memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, the processor executing the computer instructions to perform the method of any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing the computer to perform the method of any one of claims 1 to 7.

Images