A method for predicting power generation of a photovoltaic power plant

Abstract

The application relates to the technical field of photovoltaic power generation prediction, and discloses a power generation power prediction method for a photovoltaic power plant, which comprises the following steps: obtaining historical photovoltaic output data and corresponding historical meteorological data; performing multidimensional feature extraction on the historical meteorological data to generate primary meteorological features, verifying the primary meteorological features, calculating the contribution degree of each primary meteorological feature to photovoltaic power generation, and screening to generate secondary meteorological features; constructing a photovoltaic power generation power prediction model based on the secondary meteorological features and the historical photovoltaic output data; obtaining meteorological data of a target prediction period, and generating a power generation power prediction result of the photovoltaic power plant through the photovoltaic power generation power prediction model. By combining multidimensional analysis of meteorological data and deep learning technology, the accuracy and real-time performance of power generation prediction of the photovoltaic power plant are improved by dynamically adjusting model parameters.

Description

Technical Field

[0001] This invention relates to the field of photovoltaic power generation prediction technology, and in particular to a method for predicting the power generation of a photovoltaic power plant. Background Technology

[0002] Currently, photovoltaic (PV) power generation, as a renewable energy technology, has experienced rapid development and widespread application in recent years. With the increasing global demand for clean energy, PV power generation, due to its modularity, ease of deployment, and environmental friendliness, has become a significant force driving energy transformation.

[0003] However, the output power of photovoltaic (PV) power generation is affected by various factors, such as weather conditions, seasonal changes, and geographical location, making PV power generation forecasting a technical challenge. Accurate power forecasting is crucial for the optimized operation of power plants, the stable operation of the power grid, and the improvement of energy efficiency. Summary of the Invention

[0004] The purpose of this invention is to provide an accurate method for predicting photovoltaic power generation. By combining multi-dimensional analysis of meteorological data with deep learning technology and dynamically adjusting model parameters, the accuracy and real-time performance of photovoltaic power plant power generation prediction are improved.

[0005] To achieve the above objectives, the present invention provides a method for predicting the power generation of a photovoltaic power plant, comprising:

[0006] Obtain historical photovoltaic power output data and corresponding historical meteorological data;

[0007] Multi-dimensional feature extraction is performed on historical meteorological data to generate primary meteorological features. The primary meteorological features are then validated, and the contribution of each primary meteorological feature to photovoltaic power generation is calculated. Secondary meteorological features are then selected and generated.

[0008] Based on secondary meteorological characteristics and historical photovoltaic power output data, a photovoltaic power generation prediction model is constructed.

[0009] Meteorological data for the target forecast period is obtained, and the power generation forecast results of photovoltaic power plants are generated through the photovoltaic power generation forecast model.

[0010] In some embodiments of the present invention, multi-dimensional feature extraction is performed on historical meteorological data to generate primary meteorological features, including:

[0011] Classify historical meteorological data to generate classified historical meteorological data X', X' = [SR, TM, HU, W, C, P];

[0012] Where SR = [sr1, sr2, ..., sr] i …sr n ], TM = [tm1, tm2…tmi …tm n ],HU=[hu1,hu2…hu i …hu n ],W=[w1,w2…w i …w n ],C=[c1,c2…c i …c n ],P=[p1,p2…p i …p n ];

[0013] Wherein, SR represents historical solar radiation data, TM represents historical temperature data, HU represents historical humidity data, W represents historical wind speed data, C represents historical cloud cover data, and P represents historical precipitation data; sr i represents the i th historical solar radiation data, tm i represents the i th historical temperature data, hu i represents the i th historical humidity data, w i represents the i th historical wind speed radiation data, w i represents the i th historical cloud cover radiation data, p i represents the i th historical precipitation radiation data, and n represents the number of last historical meteorological data acquisition.

[0014] The covariance matrix Σ[Y ij ] is constructed for the classified historical meteorological data X'.

[0015] Wherein, Y ij =Cov(X' i ,X' j ), Y ij represents the i th row and j th column element in the covariance matrix.

[0016] The eigenvalues λ1, λ2, λ3, λ4, λ5 and λ6 of the covariance matrix are calculated.

[0017] The cumulative variance contribution rate is calculated

[0018] The value of i when k is greater than or equal to 0.9 is calculated, and the first-level meteorological feature is generated.

[0019] In some embodiments of the application, the relationship between the meteorological data and the photovoltaic output is analyzed based on the second-level meteorological feature and the historical photovoltaic output data, comprising:

[0020] The linear correlation degree between the photovoltaic power generation power and each meteorological factor is calculated.

[0021] The multiple perception machine model of meteorological factors on photovoltaic output is constructed, and the multiple perception machine model is trained to capture the nonlinear relationship between different types of historical meteorological data and photovoltaic power generation data.

[0022] Based on the historical meteorological data and real-time photovoltaic output data, the time bias b of meteorological factors on photovoltaic output is determined.

[0023] In some embodiments of the application, the linear correlation degree between the photovoltaic power generation power and each meteorological factor is calculated, including:

[0024] Based on the classified historical meteorological data X', the historical photovoltaic power generation data PW at the same time is obtained, PW=[pw1, pw2……pw n ] is obtained, and the correlation analysis between different types of historical meteorological data and photovoltaic power generation data is performed to obtain the correlation coefficient between historical meteorological data and photovoltaic power generation power:

[0025] The value of the correlation coefficient is between-1 and 1.

[0026] In some embodiments of the application, the multiple perception machine model of meteorological factors on photovoltaic output is constructed, including:

[0027] Input layer, hidden layer and output layer

[0028] The input layer includes:

[0029] 6 neurons, respectively corresponding to the input SR=[sr1, sr2……sr i ……sr n ], TM=[tm1, tm2……tm i ……tm n ], HU=[hu1, hu2……hu i ……hu n ], W=[w1, w2……w i ……w n ], C=[c1, c2……c i ……c n ], P=[p1, p2……p i ……p n ];

[0030] The hidden layer includes:

[0031] Two sub-hidden layers, each sub-hidden layer has a plurality of neurons, and each neuron is fully connected with all neurons of the next layer to form a fully connected network.

[0032] The output layer includes one neuron, and outputs the photovoltaic output data PW.

[0033] In some embodiments of the application, the training of the multiple perception machine model captures the nonlinear relationship between different types of historical meteorological data and photovoltaic power generation data, including:

[0034] Using historical data to train the multiple perception machine model of meteorological factors on photovoltaic output, comparing the predicted output of the network with the actual target output, using mean square error as the loss function, and calculating the mean square error MSE;

[0035]

[0036] Wherein, represents the i-th moment of the multiple perception machine model predicting the output data, PW i represents the actual photovoltaic power generation output data at the i-th moment.

[0037] According to the value of the mean square error, the weights and biases of the network are updated, and the nonlinear relationship between different types of historical meteorological data and photovoltaic power generation data is output.

[0038] In some embodiments of the application, the time bias b of the meteorological factor on the photovoltaic output is determined based on the historical meteorological data and the real-time photovoltaic output data, including:

[0039] Obtain historical meteorological data X' and photovoltaic output data PW s ,

[0040] Divide the historical meteorological data based on the real-time photovoltaic output data, wherein the time length of the collected real-time photovoltaic processing data is m times the time length of the divided historical meteorological data, m is a positive integer, and generate historical meteorological data of different time stages;

[0041] By comparing the historical meteorological data of different time stages with the photovoltaic output data PW s from s minutes ago to the current data acquisition time, generate the time bias b, and adjust the meteorological data observation period when predicting the photovoltaic output based on the time bias b.

[0042] In some embodiments of the application, the time bias b is generated by comparing the historical meteorological data of different time stages with the photovoltaic output data PW s from s minutes ago to the current data acquisition time, including:

[0043] Set the time window length to m, and calculate the covariance of the historical meteorological data and the real-time photovoltaic output data of different time windows respectively;

[0044] Calculate the standard deviation of the historical meteorological data and the standard deviation of the real-time photovoltaic output data of different time windows respectively;

[0045] The sliding correlation coefficient is calculated based on the covariance of the historical meteorological data and the real-time photovoltaic output data of different time windows, the standard deviation of the historical meteorological data, and the standard deviation of the real-time photovoltaic output data.

[0046] The sliding correlation coefficients of different time windows are compared.

[0047] The absolute value of the difference between the sliding correlation coefficient of all time windows and 1 is taken, the sliding correlation coefficient with the minimum absolute value is selected, and the time window corresponding to the selected sliding correlation coefficient is determined as the time bias b.

[0048] In some embodiments of the application, the photovoltaic power generation power prediction model is constructed, comprising:

[0049] A preliminary photovoltaic power generation power prediction model is established based on the linear and nonlinear correlation relationships between the obtained historical meteorological data and historical photovoltaic output data.

[0050] The parameters of the preliminary photovoltaic power generation power prediction model are corrected based on the time bias b.

[0051] The corrected results are output to obtain a corrected photovoltaic power generation power prediction model.

[0052] The corrected photovoltaic power generation power prediction model is trained to generate a photovoltaic power generation power prediction model.

[0053] Compared with the prior art, the photovoltaic power plant power prediction method of the embodiment of the application has the following beneficial effects:

[0054] The number k of eigenvalues when the cumulative variance contribution rate reaches or exceeds 90% is used to determine the first-level meteorological features, ensuring the balance between the interpretability and accuracy of the model, screening out meteorological features that have a significant contribution to photovoltaic power generation, and improving the pertinence and efficiency of the model prediction.

[0055] The selected meteorological features are verified to ensure their reliable contribution to photovoltaic power generation power prediction, and a 90% cumulative contribution rate threshold is set to ensure that the model only contains features that have a significant impact on photovoltaic power generation power, avoiding the interference of irrelevant variables.

[0056] By analyzing the Pearson correlation coefficients corresponding to different meteorological factors, meteorological factors that have a significant impact on photovoltaic power generation power can be identified, and the identified strongly correlated meteorological factors can be used as key inputs for model construction, which helps to construct a more accurate photovoltaic power generation power prediction model.

[0057] The multi-layer perception model constructed can effectively capture the complex relationship between meteorological factors and photovoltaic output through its nonlinear processing capability and deep learning architecture, improving the accuracy and stability of photovoltaic output prediction.

[0058] Through the determination of the time offset b, the prediction model can more accurately adjust the meteorological data use period, significantly improving the accuracy of photovoltaic power output prediction, and based on the dynamic adjustment of the sliding correlation coefficient, the model can flexibly adapt to the changes of the relationship between meteorological conditions and photovoltaic power output under different environmental conditions, enhancing the adaptability and reliability of the prediction.

[0059] By constructing a photovoltaic power generation power prediction model based on historical data training and combined with time offset optimization, and through regular verification and update, the continuous effectiveness and prediction accuracy of the model are ensured, providing solid technical support for the efficient operation of photovoltaic power stations. BRIEF DESCRIPTION OF DRAWINGS

[0060] Figure 1 is a power generation power prediction method flowchart of a photovoltaic power plant provided by an embodiment of the present application. DETAILED DESCRIPTION

[0061] The specific embodiments of the present application will be further described in detail below in conjunction with the drawings and examples. The following examples are used to illustrate the present application, but are not used to limit the scope of the present application.

[0062] In the description of the present application, it should be understood that the terms "center", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer" and the like indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present application and simplifying the description, and therefore cannot be understood as indicating or implying that the devices or elements referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as limiting the present application.

[0063] The terms "first", "second" are only for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the technical features indicated. Therefore, the features defined with "first", "second" can explicitly or implicitly include one or more of the features. In the description of the present application, unless otherwise specified, the meaning of "a plurality of" is two or more.

[0064] In the description of the present application, it should be noted that, unless otherwise explicitly specified and limited, the terms "mounting", "connecting", "connection" should be understood broadly, for example, it can be fixed connection, or detachable connection, or integrally connected; it can be mechanical connection, or electrical connection; it can be directly connected, or indirectly connected through an intermediate medium, or the communication inside two elements. For those skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.

[0065] Example 1:

[0066] As Figure 1 shown, embodiments of the present application disclose obtaining historical photovoltaic output data and corresponding historical meteorological data;

[0067] Multi-dimensional feature extraction is performed on the historical meteorological data to generate primary meteorological features, and the primary meteorological features are verified to calculate the contribution of each primary meteorological feature to photovoltaic power generation, and secondary meteorological features are screened and generated;

[0068] Based on the secondary meteorological features and the historical photovoltaic output data, a photovoltaic power generation power prediction model is constructed;

[0069] Obtain meteorological data of a target prediction period, and generate a photovoltaic power plant power prediction result through the photovoltaic power generation power prediction model.

[0070] Embodiment 2:

[0071] Multi-dimensional feature extraction is performed on the historical meteorological data to generate primary meteorological features, including:

[0072] The historical meteorological data is classified to generate classified historical meteorological data X', X' = [SR, TM, HU, W, C, P];

[0073] Wherein, SR = [sr1, sr2…sr i …sr n ], TM = [tm1, tm2…tm i …tm n ], HU = [hu1, hu2…hu i …hu n ], W = [w1, w2…w i …w n ], C = [c1, c2…c i …c n ], P = [p1, p2…p i …p n ];

[0074] Wherein, SR represents historical solar radiation data, TM represents historical temperature data, HU represents historical humidity data, W represents historical wind speed data, C represents historical cloud cover data, and P represents historical precipitation data; sr i represents the i-th historical solar radiation data, tm i represents the i-th historical temperature data, hu i represents the i-th historical humidity data, w i represents the i-th historical wind speed radiation data, w i represents the i-th historical cloud cover radiation data, and p iXi represents the i-th historical precipitation radiation data, and n represents the number of the last historical meteorological data obtained;

[0075] The covariance matrix Σ[Y ij ] is constructed for the classified historical meteorological data X';

[0076] Wherein, Y ij =Cov(X' i ,X' j ), Y ij represents the i-th row and j-th column element in the covariance matrix;

[0077] The eigenvalues λ1, λ2, λ3, λ4, λ5, λ6 of the covariance matrix are calculated;

[0078] The cumulative variance contribution rate is calculated

[0079] The value of i when k≥0.9 is calculated to generate the primary meteorological features.

[0080] In this embodiment

[0081] The main meteorological features are verified using the evidence optimization combination theory to generate the contribution of the main meteorological features to photovoltaic power generation. The secondary meteorological features are generated by screening the meteorological features with a cumulative contribution rate of more than 90%, including:

[0082] The basic probability assignment (BPA) function is constructed to assign a trust degree to each hypothesis (feature). The belief and likelihood calculation method in the Dempster-Shafer theory is used to generate the contribution of each feature to photovoltaic power generation.

[0083] In this embodiment, feature extraction and screening: by classifying historical meteorological data and constructing a covariance matrix, the primary meteorological features are extracted, and the contribution of each meteorological factor to photovoltaic power generation can be quantitatively evaluated. This method helps to identify the most influential meteorological parameters, thus establishing a more accurate relationship in the model.

[0084] In this embodiment, by classifying historical meteorological data into six categories: solar radiation (SR), temperature (TM), humidity (HU), wind speed (W), cloud cover (C), and precipitation (P), a feature matrix X' is formed, effectively organizing and preparing the data for subsequent analysis, providing a clear structure.

[0085] Covariance matrix and eigenvalue analysis

[0086] The calculated covariance matrix Σ reveals the statistical relationship between different meteorological factors, which is a key step in quantifying the correlation between features. The eigenvalues reveal the principal components of the covariance matrix, i.e., the main fluctuation direction of the data, which is crucial for understanding the data structure and information compression.

[0087] Cumulative variance contribution rate and feature screening

[0088] By determining the first-level meteorological features through the number of feature values k when the cumulative variance contribution rate reaches or exceeds 90%, the balance between the interpretability and accuracy of the model is ensured, the meteorological features that have a significant contribution to photovoltaic power generation are screened out, and the pertinence and efficiency of the model prediction are improved.

[0089] Meteorological feature verification and contribution quantification

[0090] Model verification: The selected meteorological features are verified to ensure their contribution to photovoltaic power generation power prediction is reliable, and by setting a cumulative contribution rate threshold of 90%, it is ensured that the model only contains features that have a significant impact on photovoltaic power generation power, avoiding the interference of irrelevant variables.

[0091] Example 3:

[0092] Based on the secondary meteorological features and historical photovoltaic output data, the relationship between meteorological data and photovoltaic output is analyzed, including:

[0093] Calculate the linear correlation between photovoltaic power generation power and each meteorological factor;

[0094] Construct a multiple perception machine model of meteorological factors on photovoltaic output, train the multiple perception machine model, and capture the nonlinear relationship between different types of historical meteorological data and photovoltaic power generation data;

[0095] Based on historical meteorological data and real-time photovoltaic output data, determine the time bias b of meteorological factors on photovoltaic output.

[0096] In this embodiment, through linear correlation analysis, nonlinear relationship construction, model training, and time bias determination, the technology can significantly improve the accuracy and stability of photovoltaic power generation power prediction. This not only provides strong technical support for the operation and management of photovoltaic power stations, but also optimizes power grid scheduling and improves the overall energy system operation efficiency and reliability.

[0097] Example 4:

[0098] The linear correlation between photovoltaic power generation power and each meteorological factor is calculated, including:

[0099] Based on the classified historical meteorological data X', the historical photovoltaic power generation data PW at the same time is obtained, PW = [pw1, pw2 …… pw n ], and the correlation analysis between different types of historical meteorological data and photovoltaic power generation data is performed, respectively obtaining the correlation coefficients between historical meteorological data and photovoltaic power generation power:

[0100] The value of the correlation coefficient is between -1 and 1.

[0101] Based on the classified historical meteorological data X', historical photovoltaic power generation data PW at the same time is obtained, PW = [pw1, pw2... pw n ], and correlation analysis is performed on different types of historical meteorological data and photovoltaic power generation data, respectively, to obtain the correlation coefficients between historical meteorological data and photovoltaic power generation power:

[0102] The value of the correlation coefficient is between -1 and 1, close to 1 or -1 indicating strong positive or strong negative correlation, close to 0 indicating no linear correlation.

[0103] In this embodiment, based on the classified historical meteorological data X', historical photovoltaic power generation data PW at the same time is obtained, PW = [pw1, pw2... pw n ], and correlation analysis is performed on different types of historical meteorological data and photovoltaic power generation data, respectively, to obtain the correlation coefficients between historical meteorological data and photovoltaic power generation power:

[0104] r SR represents the degree of correlation between historical solar radiation data and photovoltaic power generation power,

[0105]

[0106] r TM represents the degree of correlation between historical temperature data and photovoltaic power generation power,

[0107]

[0108] r HU represents the degree of correlation between historical humidity data and photovoltaic power generation power,

[0109]

[0110] r W represents the degree of correlation between historical wind speed data and photovoltaic power generation power,

[0111]

[0112] r C represents the degree of correlation between historical cloud cover data and photovoltaic power generation power,

[0113]

[0114] r P represents the degree of correlation between historical precipitation data and photovoltaic power generation power,

[0115]

[0116] In this embodiment, the linear correlation between historical meteorological data and photovoltaic power generation is quantified by calculating the Pearson correlation coefficient. The value of each coefficient is between -1 and 1, which provides an intuitive measure for understanding the relationship between various meteorological factors and photovoltaic power generation.

[0117] In this embodiment, by analyzing the Pearson correlation coefficients corresponding to different meteorological factors, meteorological factors that have a significant impact on photovoltaic power generation can be identified. For example, solar radiation (SR) is usually positively correlated with photovoltaic power generation, while temperature (TM) may also affect the efficiency of photovoltaic panels within a certain range, thereby affecting power generation. This helps to determine which factors should be prioritized in modeling and prediction.

[0118] In this embodiment, the strong correlation meteorological factors identified can be used as key inputs for model construction, which helps to build more accurate photovoltaic power generation prediction models. By excluding factors with weak correlation, the complexity of the model is reduced, while the prediction efficiency and accuracy of the model are improved.

[0119] Embodiment 5:

[0120] The constructed multiple perception machine model of meteorological factors on photovoltaic output includes:

[0121] Input layer, hidden layer and output layer

[0122] The input layer includes:

[0123] 6 neurons corresponding to input SR = [sr1, sr2…sr i …sr n ], TM = [tm1, tm2…tm i …tm n ], HU = [hu1, hu2…hu i …hu n ], W = [w1, w2…w i …w n ], C = [c1, c2…c i …c n ], P = [p1, p2…p i …p n ];

[0124] The hidden layer includes:

[0125] Two sub-hidden layers, each sub-hidden layer has multiple neurons, and each neuron is fully connected to all neurons in the next layer, forming a fully connected network.

[0126] The output layer includes 1 neuron, which outputs photovoltaic output data PW.

[0127] In this embodiment, the input layer directly receives six meteorological factors as inputs, including solar radiation (SR), temperature (TM), humidity (HU), wind speed (W), cloud cover (C), and precipitation (P). Through forward propagation of the neural network, these features are fused and interacted in the hidden layer to form more abstract feature representations, laying the foundation for more accurate prediction.

[0128] In this embodiment, the output layer only has one neuron, which is used to directly predict photovoltaic power output data (PW). This design makes the model's output directly correspond to the prediction target, simplifying the output structure and facilitating model training and optimization.

[0129] In this embodiment, the MLP model is trained through the backpropagation algorithm, which can automatically adjust the weights and biases of each layer of neurons to minimize prediction error. The use of ReLU activation function not only speeds up the training process, but also avoids the problem of gradient vanishing, improving the training efficiency and effectiveness of the model.

[0130] In this embodiment, the combination of complex network structure and nonlinear activation function makes the model have stronger generalization ability. Even when dealing with unseen meteorological data, the model can still make accurate predictions based on learned patterns, improving the practicality of the model.

[0131] Embodiment 6:

[0132] The multiple perception machine model is trained to capture the nonlinear relationship between different types of historical meteorological data and photovoltaic power generation data, including:

[0133] The multiple perception machine model of meteorological factors on photovoltaic power output is trained using historical data. The predicted output of the network is compared with the actual target output, and the mean square error is used as the loss function, and the mean square error MSE is calculated.

[0134]

[0135] wherein, represents the i-th moment of the multiple perception machine model predicting the output data, PW i represents the i-th moment of the actual photovoltaic power output data.

[0136] The weights and biases of the network are updated according to the value of the mean square error, and the nonlinear relationship between different types of historical meteorological data and photovoltaic power generation data is output.

[0137] The multiple perception machine model is trained to capture the nonlinear relationship between different types of historical meteorological data and photovoltaic power generation data, including:

[0138] The historical data is used to train a multiple perception machine model of meteorological factors on photovoltaic output, to minimize the mean square error between the predicted output and the actual output:

[0139] The predicted output of the network is compared with the actual target output, the mean square error is used as the loss function, and the mean square error MSE is calculated;

[0140]

[0141] wherein, represents the i-th moment of multiple perception machine model prediction output data, PW i represents the i-th moment of actual photovoltaic power generation output data;

[0142] According to the value of the mean square error, the gradient descent is performed on the network weight and bias, and the error gradient is calculated by back propagation along the network; according to the gradient calculated by back propagation, the weight and bias of the network are updated, and iterative optimization is performed, and the training period is set to 800-1000 times.

[0143] In this embodiment, the mean square error (MSE) is used as the loss function, which can effectively measure the difference between the model prediction value and the actual value, especially in dealing with continuous numerical prediction problems.

[0144] In this embodiment, the gradient of the loss function with respect to the network weight and bias is calculated, and the parameters are updated in the opposite direction of the gradient, and the gradient descent algorithm ensures that the model can gradually reduce the prediction error and achieve the optimization goal.

[0145] In this embodiment, the training period is set to 800-1000 times, which provides sufficient training opportunities for the model, avoids suboptimal solution caused by premature stopping of training, and also prevents overfitting, ensuring that the model has good generalization ability.

[0146] Embodiment 7:

[0147] The time bias b of meteorological factors on photovoltaic output is determined based on historical meteorological data and real-time photovoltaic output data, comprising:

[0148] The historical meteorological data X' and the photovoltaic output data PW s ,

[0149] The historical meteorological data is divided based on real-time photovoltaic output data, wherein the time length of the collected real-time photovoltaic processing data is m times the time length of the divided historical meteorological data, m is a positive integer, and the historical meteorological data of different time stages is generated;

[0150] By comparing historical meteorological data from different time periods with photovoltaic power output data (PW) from the previous s minutes to the current data acquisition time, the system was able to achieve the desired results. s Generate a time offset b, and adjust the meteorological data observation period for photovoltaic power output prediction based on the time offset b.

[0151] Acquire historical meteorological data X' and photovoltaic power output data PW for a period of time prior to the current sampling time. s ,

[0152] Among them, PW s This refers to the photovoltaic power output data from the previous s minutes to the current data acquisition time.

[0153] Based on real-time photovoltaic power output data, historical meteorological data are divided into different time periods to generate historical meteorological data for different time periods: X'1, X'2, X'3, X'4, X'5, X'6;

[0154] and PW s Statistical comparisons are performed, the time offset b is calculated, and the meteorological data observation period for photovoltaic power output prediction is adjusted based on the time offset b.

[0155] In this embodiment, determining the time offset b of photovoltaic output based on historical meteorological data and real-time photovoltaic output data includes:

[0156] Acquire historical meteorological data X' and real-time photovoltaic power output data PW within 30 minutes. n Based on real-time photovoltaic power output data, historical meteorological data are divided to generate historical meteorological data for different time periods: X'1, X'2, X'3, X'4, X'5, X'6;

[0157] Among them, X'1 corresponds to the historical meteorological data of the previous 10 minutes, X'2 corresponds to the historical meteorological data of the previous 20-10 minutes, X'3 corresponds to the historical meteorological data of the previous 30-20 minutes, X'4 corresponds to the historical meteorological data of the previous 40-30 minutes, X'5 corresponds to the historical meteorological data of the previous 50-40 minutes, and X'6 corresponds to the historical meteorological data of the previous 60-50 minutes.

[0158] The data is then statistically compared with real-time data to calculate the time offset b, and the meteorological data observation period for photovoltaic power output prediction is adjusted based on the time offset b.

[0159] Example 8:

[0160] The method involves comparing historical meteorological data from different time periods with photovoltaic power output data (PW) from the previous s minutes to the current data acquisition time. s Generate time offset b, including:

[0161] Setting the time window length as m, respectively calculating the covariance of the historical meteorological data and the real-time photovoltaic output data of different time windows;

[0162] Respectively calculating the standard deviation of the historical meteorological data and the standard deviation of the real-time photovoltaic output data of different time windows;

[0163] Based on the covariance of the historical meteorological data and the real-time photovoltaic output data of different time windows, the standard deviation of the historical meteorological data and the standard deviation of the real-time photovoltaic output data, the sliding correlation coefficient is calculated;

[0164] Comparing the sliding correlation coefficients of different time windows;

[0165] The absolute value of the difference between the sliding correlation coefficient of all time windows and 1 is taken, the sliding correlation coefficient with the smallest absolute value is selected, and the time window corresponding to the selected sliding correlation coefficient is determined as the time bias b.

[0166] The statistical comparison with the real-time data to calculate the time bias b includes:

[0167] Setting the time window as 3, respectively calculating the covariance of the historical meteorological data and the real-time photovoltaic output data of different time windows, generating the covariance of the historical meteorological data and the real-time photovoltaic output data of the previous 30 minutes cov(X'1+X'2+X'3, PW n );

[0168] The covariance of the historical meteorological data and the real-time photovoltaic output data of the previous 40-10 minutes cov(X'2+X'3+X'4, PW n );

[0169] The covariance of the historical meteorological data and the real-time photovoltaic output data of the previous 50-20 minutes cov(X'3+X'4+X'5, PW n );

[0170] The covariance of the historical meteorological data and the real-time photovoltaic output data of the previous 60-30 minutes cov(X'4+X'5+X'6, PW n );

[0171] Measuring the common trend of the historical meteorological data and the real-time photovoltaic output data;

[0172] Respectively calculating the standard deviation of the historical meteorological data of different time windows, including:

[0173] std(X'1+X'2+X'3) The standard deviation of the historical meteorological data of the previous 30 minutes;

[0174] std(X'2+X'3+X'4) The standard deviation of the historical meteorological data of the previous 40-10 minutes;

[0175] std(X'3+X'4+X'5) standard deviation of historical meteorological data in the previous 50-20 minutes;

[0176] std(X'4+X'5+X'6) standard deviation of historical meteorological data in the previous 60-30 minutes;

[0177] and the standard deviation std(PW n ) of real-time photovoltaic power output data, measuring the dispersion degree of historical meteorological data and real-time photovoltaic power output data;

[0178] Based on the covariance of historical meteorological data and real-time photovoltaic power output data in different time windows, the standard deviation of historical meteorological data and the standard deviation of real-time photovoltaic power output data, the sliding correlation coefficient is calculated:

[0179] the sliding correlation coefficient of historical meteorological data in the previous 30 minutes and real-time photovoltaic power output data,

[0180] the sliding correlation coefficient of historical meteorological data in the previous 40-10 minutes and real-time photovoltaic power output data,

[0181] the sliding correlation coefficient of historical meteorological data in the previous 50-20 minutes and real-time photovoltaic power output data,

[0182] the sliding correlation coefficient of historical meteorological data in the previous 50-20 minutes and real-time photovoltaic power output data,

[0183] By comparing the sliding correlation coefficients of different time windows, the sliding correlation coefficient closest to 1 is selected, and the time window corresponding to the current sliding correlation coefficient is determined as the time bias b.

[0184] In this embodiment, through the sliding window analysis, the time period in which the meteorological factor has the most significant influence on the photovoltaic power output can be accurately identified, providing key parameters for the prediction model.

[0185] The determination of the time bias b enables the prediction model to more accurately adjust the time period for using meteorological data, significantly improving the accuracy of photovoltaic power output prediction.

[0186] Based on the dynamic adjustment of the sliding correlation coefficient, the model can flexibly adapt to changes in the relationship between meteorological conditions and photovoltaic power output under different environmental conditions, enhancing the adaptability and reliability of the prediction.

[0187] Embodiment 9:

[0188] The photovoltaic power generation power prediction model is constructed, comprising:

[0189] A preliminary photovoltaic power generation power prediction model is established based on the linear correlation and nonlinear correlation between the obtained historical meteorological data and historical photovoltaic output data.

[0190] The preliminary photovoltaic power generation power prediction model parameters are corrected based on the time bias b.

[0191] A photovoltaic power generation power prediction correction model is output according to the correction result.

[0192] The photovoltaic power generation power prediction correction model is trained to generate a photovoltaic power generation power prediction model.

[0193] The photovoltaic power generation power prediction model is constructed, including:

[0194] A preliminary photovoltaic power generation power prediction model is established based on the linear correlation and nonlinear correlation between the obtained historical meteorological data and historical photovoltaic output data, and the model parameters are adjusted based on the time bias b. After training the preliminary photovoltaic power generation power prediction model, a photovoltaic power generation power prediction model is generated.

[0195] In this embodiment, the linear and nonlinear relationship analysis

[0196] The historical meteorological data and historical photovoltaic output data are used to determine the linear and nonlinear relationship through correlation analysis to provide a basis for model establishment. A preliminary photovoltaic power generation power prediction model is constructed, which comprehensively considers the complex relationship between meteorological factors and photovoltaic output. The time bias b is introduced to adjust the parameters of the preliminary model to ensure that the model accurately reflects the time delay of the influence of meteorological changes on photovoltaic output. The preliminary model is trained using historical data to optimize model parameters and improve prediction performance. After training and adjustment, the final photovoltaic power generation power prediction model is generated.

[0197] Through historical data training, the model can effectively capture the linear and nonlinear relationship between meteorological factors and photovoltaic output, improving prediction accuracy. The time bias b is introduced to optimize model parameters, making the model more accurately reflect the influence of meteorological changes on photovoltaic output.

[0198] The model is periodically verified and updated, including:

[0199] Real-time meteorological data is input into the photovoltaic power generation power prediction model to output power generation power prediction; actual photovoltaic output data is collected periodically and compared with prediction data to evaluate the prediction accuracy of the model; according to the verification result, if necessary, the model is retrained or the model parameters are adjusted to maintain the performance of the model

[0200] In this embodiment, real-time meteorological data is input into a trained photovoltaic power generation prediction model, and the predicted power generation value is output. Actual photovoltaic output data is collected periodically and compared with the predicted data to evaluate the model's prediction accuracy. Based on the verification results, if the model's prediction accuracy decreases or there is a large error, the model is retrained or the model parameters are adjusted.

[0201] Regular validation ensures stable model performance, and timely updates and optimization maintain high accuracy in model predictions. The model can be dynamically adjusted according to actual operating conditions, improving its adaptability to environmental changes and the flexibility of prediction.

[0202] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and substitutions can be made without departing from the technical principles of the present invention, and these improvements and substitutions should also be considered within the scope of protection of the present invention.

Claims

1. A method for predicting the power generation of a photovoltaic power plant, characterized in that, include: Obtain historical photovoltaic power output data and corresponding historical meteorological data; Multi-dimensional feature extraction is performed on historical meteorological data to generate primary meteorological features. The primary meteorological features are then validated, and the contribution of each primary meteorological feature to photovoltaic power generation is calculated. Secondary meteorological features are then selected and generated. Based on secondary meteorological characteristics and historical photovoltaic power output data, a photovoltaic power generation prediction model is constructed. Meteorological data for the target forecast period is obtained, and the power generation forecast results of photovoltaic power plants are generated through the photovoltaic power generation forecast model. The process of extracting multi-dimensional features from historical meteorological data to generate primary meteorological features includes: Classify historical meteorological data to generate classified historical meteorological data X',X'=[SR,TM,HU,W,C,P]; Where SR=[sr1,sr2…sr i …sr n ],TM=[tm1,tm2…tm i …tm n ],HU=[hu1,hu2…hu i …hu n ],W=[w1,w2…w i …w n ], C=[c1,c2…c i …c n ], P=[p1,p2…p i …p n ]; Where SR represents historical solar radiation data, TM represents historical temperature data, HU represents historical humidity data, W represents historical wind speed data, C represents historical cloud cover data, and P represents historical precipitation data; sr i tm represents the i-th historical solar radiation data. i Hu represents the i-th historical temperature data. i Let w represent the i-th historical humidity data. i Let w represent the i-th historical wind speed radiation data. i p represents the i-th historical cloud cover radiation data. i This represents the i-th historical precipitation radiation data, and n represents the number of historical meteorological data acquired in the last time. Construct the covariance matrix Σ[Y] of the classified historical meteorological data X'. ij ]; Among them, Y ij =Cov(X' i ,X' j ),Y ij This represents the element in the i-th row and j-th column of the covariance matrix; Calculate the eigenvalues ​​of the covariance matrix 1, 2, 3, 4, 5, 6; Calculate the cumulative variance contribution rate ; Calculate the value of i when k ≥ 0.9 to generate the first-level meteorological feature; The analysis of the relationship between meteorological data and photovoltaic output based on secondary meteorological characteristics and historical photovoltaic output data includes: Calculate the degree of linear correlation between photovoltaic power generation and various meteorological factors; A multiple sensing model of the effect of meteorological factors on photovoltaic power output was constructed. The multiple sensing model was trained, and different types of historical meteorological data were captured to establish nonlinear relationships with photovoltaic power generation data. Based on historical meteorological data and real-time photovoltaic power output data, the time offset b of meteorological factors on photovoltaic power output is determined.

2. The photovoltaic power generation prediction method for photovoltaic power plants as described in claim 1, characterized in that, The calculation of the linear correlation between photovoltaic power generation and various meteorological factors includes: Based on the classified historical meteorological data X', obtain the historical photovoltaic power generation data PW at the same time, where PW=[pw1,pw2……pw] n Correlation analysis was conducted between different types of historical meteorological data and photovoltaic power generation data to obtain the correlation coefficients between historical meteorological data and photovoltaic power generation. The correlation coefficient ranges between -1 and 1.

3. The photovoltaic power generation prediction method for photovoltaic power plants as described in claim 2, characterized in that, The construction of the multiple sensing machine model of meteorological factors on photovoltaic output includes: Input layer, hidden layer, and output layer; The input layer includes: Six neurons, corresponding to the input SR=[sr1,sr2…sr… i …sr n ],TM=[tm1,tm2…tm i …tm n ],HU=[hu1,hu2…hu i …hu n ],W=[w1,w2…w i …w n ], C=[c1,c2…c i …c n ], P=[p1,p2…p i …p n ]; The hidden layer includes: There are two sub-hidden layers, each with multiple neurons. Each neuron is fully connected to all neurons in the next layer, forming a fully connected network. The output layer includes one neuron that outputs photovoltaic power output data PW.

4. The photovoltaic power generation prediction method for photovoltaic power plants as described in claim 3, characterized in that, The training of the multiple perceptron model, capturing different types of historical meteorological data and establishing nonlinear relationships between them and photovoltaic power generation data, includes: Historical data was used to train a multiple perceptron model of the effect of meteorological factors on photovoltaic output. The network's predicted output was compared with the actual target output. The mean square error was used as the loss function, and the mean square error (MSE) was calculated. ； in, This represents the force output data predicted by the perceptron model multiple times at time i. This represents the actual photovoltaic power output data at time i. The weights and biases of the updated network are adjusted based on the mean square error value, and the nonlinear relationship between different types of historical meteorological data and photovoltaic power generation data is output.

5. The photovoltaic power generation prediction method for photovoltaic power plants as described in claim 4, characterized in that, The determination of the time offset b of photovoltaic output due to meteorological factors based on historical meteorological data and real-time photovoltaic output data includes: Acquire historical meteorological data X' and photovoltaic power output data PW from the previous s minutes to the current data acquisition time. s , Historical meteorological data is divided based on real-time photovoltaic power output data. The time length of the real-time photovoltaic processing data is m times the corresponding time length of the divided historical meteorological data, where m is a positive integer, to generate historical meteorological data for different time periods. By comparing historical meteorological data from different time periods with photovoltaic power output data (PW) from the previous s minutes to the current data acquisition time, the system was able to achieve the desired results. s Generate a time offset b, and adjust the meteorological data observation period for photovoltaic power output prediction based on the time offset b.

6. The photovoltaic power generation prediction method for photovoltaic power plants as described in claim 5, characterized in that, The method involves comparing historical meteorological data from different time periods with photovoltaic power output data (PW) from the previous s minutes to the current data acquisition time. s Generate time offset b, including: Set the time window length to m, and calculate the covariance between historical meteorological data and real-time photovoltaic power output data for different time windows; Calculate the standard deviation of historical meteorological data and the standard deviation of real-time photovoltaic power output data for different time windows; The sliding correlation coefficient is calculated based on the covariance of historical meteorological data and real-time photovoltaic power output data at different time windows, the standard deviation of historical meteorological data, and the standard deviation of real-time photovoltaic power output data. Compare the sliding correlation coefficients for different time windows; Subtract 1 from the sliding correlation coefficients of all time windows, take the absolute value of the result, select the sliding correlation coefficient with the smallest absolute value, and determine the time window corresponding to the selected sliding correlation coefficient as the time bias b.

7. The photovoltaic power generation prediction method for photovoltaic power plants as described in claim 6, characterized in that, The construction of the photovoltaic power generation prediction model includes: A preliminary model for predicting photovoltaic power generation is established based on the linear and nonlinear correlations between historical meteorological data and historical photovoltaic power output data. The parameters of the preliminary photovoltaic power generation prediction model based on time bias b are corrected. Based on the correction results, output the photovoltaic power generation prediction correction model; The photovoltaic power generation prediction correction model is trained to generate a photovoltaic power generation prediction model.

Images