Photovoltaic output prediction method for supporting energy storage planning based on optimal scoring matrix

Abstract

The invention discloses a photovoltaic output prediction method and device for supporting energy storage planning based on an optimal scoring matrix, equipment and a medium. The method comprises the following steps: training a plurality of prediction models of different time scales according to combined features of a photovoltaic output training data set; performing applicable model testing on each time point according to the plurality of prediction models to obtain an optimal scoring matrix; and based on the optimal score matrix, according to photovoltaic output historical data and the plurality of prediction models, determining a photovoltaic output prediction value of a future time period. According to the embodiment of the invention, the prediction accuracy of medium-and-long-term photovoltaic output can be improved, so that more accurate energy storage planning can be carried out subsequently.

Description

Technical Field

[0001] The present invention relates to the technical field of power grid operation, and in particular to a photovoltaic output prediction method, device, equipment and medium for supporting energy storage planning based on an optimal scoring matrix. Background Art

[0002] Distributed photovoltaic power stations can be dispersed in various regions, can directly supply local electricity demand, and reduce energy consumption and losses during transmission. If the medium- and long-term photovoltaic output can be predicted in a timely and accurate manner, it will have important value for applications such as light resource evaluation, power grid operation, and planning of new photovoltaic power stations. Photovoltaic power stations are a typical intermittent power source; and they are affected by meteorological and environmental conditions, have strong diurnal and seasonal periodicity, and are characterized by many influencing factors, complex mechanisms, strong volatility and randomness, and fast frequency of change. These factors bring great difficulties to the accurate prediction of medium- and long-term distributed photovoltaic output. In some existing studies, the modeling methods based on mechanism models, the modeling methods based on probability statistics, the modeling methods based on artificial intelligence algorithms, and other direct or indirect prediction methods are mainly used to carry out research.

[0003] The mechanism modeling method mainly considers physical equations such as solar radiation transfer equation and photovoltaic component operation equation for prediction research. It does not require a large amount of historical data, but requires detailed geographic information, component parameters, and meteorological and solar radiation data of photovoltaic power stations. It is suitable for newly built photovoltaic power stations. The modeling process is complicated and it is difficult to simulate some extreme abnormal weather conditions and the slow changes of environmental and photovoltaic component parameters over time. The model has poor anti-interference ability and is not robust. The main prediction models include the Clear Sky Model (CSM), Solis model, ESRA model, Bird and Hulstrom model, Ineichen model, etc. Under clear sky conditions, photovoltaic power fluctuations are small, which can maximize the reflection of the power generation effect of irradiance. Therefore, the clear sky prediction model is often used as the basic calibration model. Cloud movement is the main reason for the fluctuation of ground irradiance and photovoltaic power. The prediction method based on cloud map has become one of the important technical directions for refined photovoltaic prediction. All cloud maps need to be processed by image processing to identify clouds, and block matching technology, cross-correlation algorithm and other methods are used to predict cloud movement, form cloud index map, and predict cloud occlusion. The analysis data of satellite cloud images is combined with ground telemetry data as the input of the prediction model for prediction, and the ground-based cloud image data is combined with satellite cloud image data and NWP data for prediction. The prediction method based on probability statistics uses the geographical information of the power station, numerical weather forecast, meteorological station observation information, and long-term power station operation information for a large amount of learning to establish a prediction model based on probability statistics analysis and multi-mode collection; it mainly includes gray theory, time series method and probability prediction method, etc., and uses parameter estimation and curve fitting to obtain the corresponding relationship between historical meteorological data and power generation, so as to realize the prediction of power generation of photovoltaic power stations.

[0004] In recent years, with the gradual maturity of artificial intelligence technology and big data analysis technology, traditional modeling methods have been gradually replaced by advanced technologies, especially artificial intelligence modeling methods represented by machine learning and deep learning technology, such as ensemble learning model (Ensemble Model), recurrent neural network model (RNN), long short-term memory network model (LSTM), Transformer network model, particle swarm algorithm, genetic algorithm, etc., all of which have achieved good results. At present, most studies consider the impact of meteorological data on power generation, ignoring the periodic variation characteristics of photovoltaic power generation data at different time scales. There are few comprehensive modeling and analysis studies on the output characteristics of photovoltaic power stations in different regions and the dynamic analysis of historical time series characteristics. At the same time, due to the limitations of algorithm structure and strategy, it is difficult to effectively integrate multiple models for accurate prediction on a long time scale. Summary of the invention

[0005] The present invention provides a photovoltaic output prediction method, device, equipment and medium for supporting energy storage planning based on an optimal scoring matrix, so as to improve the prediction accuracy of medium- and long-term photovoltaic output, so as to facilitate more accurate energy storage planning in the future.

[0006] According to one aspect of the present invention, a photovoltaic output prediction method for supporting energy storage planning based on an optimal scoring matrix is ​​provided, comprising:

[0007] Multiple prediction models at different time scales are trained based on the combined features of the PV output training dataset;

[0008] Performing applicable model tests at each time point according to the multiple prediction models to obtain an optimal scoring matrix;

[0009] Based on the optimal scoring matrix, the photovoltaic output forecast value for the future period is determined according to the photovoltaic output historical data and the multiple forecast models.

[0010] According to another aspect of the present invention, there is provided a photovoltaic output prediction device for supporting energy storage planning based on an optimal scoring matrix, comprising:

[0011] A model training module is used to train multiple prediction models of different time scales based on the combined features of the photovoltaic output training data set;

[0012] A model testing module, used to perform applicable model testing at each time point according to the multiple prediction models to obtain an optimal scoring matrix;

[0013] The model using module is used to determine the photovoltaic output forecast value for a future period based on the optimal scoring matrix, the photovoltaic output historical data and the multiple forecasting models.

[0014] According to another aspect of the present invention, a computer program product is provided, including a computer program, which, when executed by a processor, implements the photovoltaic output prediction method for supporting energy storage planning based on an optimal scoring matrix according to any embodiment of the present invention.

[0015] According to another aspect of the present invention, an electronic device is provided, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the photovoltaic output prediction method for supporting energy storage planning based on an optimal scoring matrix as described in any embodiment of the present invention.

[0016] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the photovoltaic output prediction method for supporting energy storage planning based on an optimal scoring matrix as described in any embodiment of the present invention when executed.

[0017] The embodiment of the present invention comprehensively models multiple features, thereby enhancing the expressive power and adaptability of the features of the input model; based on the multi-model fusion method of the optimal scoring matrix, in the process of rolling prediction, it is determined that the prediction model adopted at each time point is the optimal choice under the current circumstances, that is, by selecting the most appropriate prediction model in turn during the prediction process, the prediction accuracy of medium- and long-term photovoltaic output is improved, so that more accurate energy storage planning can be carried out subsequently.

[0018] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0020] Figure 1A is a flow chart of a photovoltaic output prediction method for supporting energy storage planning based on an optimal scoring matrix provided according to an embodiment of the present invention;

[0021] Figure 1B is a schematic diagram of a rolling prediction provided according to an embodiment of the present invention;

[0022] Figure 1C is a schematic diagram of a fixed period sequence provided according to an embodiment of the present invention;

[0023] Figure 1D is a schematic diagram of a dynamic periodic sequence provided according to an embodiment of the present invention;

[0024] Figure 2A is a flow chart of a photovoltaic output prediction method for supporting energy storage planning based on an optimal scoring matrix provided according to another embodiment of the present invention;

[0025] Figure 2B is a schematic diagram of a model training and use process provided according to another embodiment of the present invention;

[0026] Figure 2C is a schematic diagram of a medium- and long-term distributed photovoltaic output forecast capability trend provided according to another embodiment of the present invention;

[0027] Figure 2D is a schematic diagram of a medium- and long-term distributed photovoltaic output comparison curve provided according to another embodiment of the present invention;

[0028] Figure 3 is a structural schematic diagram of a photovoltaic output prediction method for supporting energy storage planning based on an optimal scoring matrix provided according to another embodiment of the present invention;

[0029] Figure 4 It is a schematic diagram of the structure of an electronic device implementing an embodiment of the present invention. DETAILED DESCRIPTION

[0030] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0031] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0032] Figure 1A A flowchart of a photovoltaic output prediction method based on an optimal scoring matrix to support energy storage planning is provided in one embodiment of the present invention. This embodiment can be applied to training prediction models of different time scales, and adopting the optimal scoring strategy to perform multi-model fusion prediction in the rolling prediction process. The method can be executed by a photovoltaic output prediction device based on an optimal scoring matrix to support energy storage planning. The device can be implemented in the form of hardware and / or software. The device can be configured in an electronic device with corresponding data processing capabilities, such as a server for predicting photovoltaic output and energy storage planning. Figure 1A As shown, the method includes:

[0033] S110, training multiple prediction models of different time scales according to the combined features of the photovoltaic output training data set.

[0034] S120: Perform applicable model tests at each time point according to the multiple prediction models to obtain an optimal scoring matrix.

[0035] S130: Based on the optimal scoring matrix, determine the photovoltaic output forecast value for the future period according to the photovoltaic output historical data and the multiple forecasting models.

[0036] Among them, the short-time scale prediction model can be used to directly predict the photovoltaic output forecast value in a shorter future period (such as after 1 month); the long-time scale prediction model can be used to directly predict the photovoltaic output forecast value in a longer future period (such as after 1 year).

[0037] Specifically, considering the volatility of photovoltaic output and the differences in data distribution in different regions, the features of photovoltaic output data at multiple latitudes are designed, such as fixed period features, dynamic features, local volatility features, and related meteorological auxiliary features, to increase the expressiveness and adaptability of the features of the input model. Accordingly, when it is necessary to extract features, the features of photovoltaic output data at different latitudes are extracted and fused into combined features as the input of the prediction model. It should be noted that the features input to the model during the subsequent backtesting and use of the prediction model are also the combined features of photovoltaic output data.

[0038] When using rolling forecasts for medium- and long-term forecast outputs, since the feature input in the subsequent forecast process may use the output of the previous forecast process, it is easy to introduce additional uncertainty and produce error accumulation. In order to avoid the influence of the error accumulation effect during the iteration process, the present invention adopts a multi-model fusion framework based on an integrated learning model, and uses combined features of different time scales to train multiple prediction models of different time scales. The time scales of these prediction models are different, so they are used in combination to reduce the number of iterations of rolling forecasts.

[0039] For example, Figure 1B As shown in the figure, if only prediction model m1 is used for rolling prediction, it needs to be rolled out 16 times to get the prediction result at the final time point. However, if prediction models m1, m3, and m12 of different time scales cooperate with each other, only three rolling predictions are needed to get the final prediction result. The number of rolling predictions is reduced by 13 times, and the error accumulation will be less. This shows that the cooperation of prediction models of multiple time scales can greatly reduce the number of iterations. However, as Figure 1BAs shown in Figure 1, there are multiple matching methods (forward order / reverse order) for prediction models of different time scales, and the final accumulated errors of each matching method are also different. How to determine the most appropriate matching method becomes a problem that needs to be considered.

[0040] To this end, the prediction models are tested one by one through the photovoltaic output backtest data set. According to the test results, the number of times each model has the best accuracy at each time point is counted as the model selection score dependency value at the corresponding time point. The model selection score dependency values ​​of all models in all time periods form the optimal score matrix. The optimal score matrix records the model selection score dependency values ​​of each prediction model at each time point.

[0041] When it is necessary to predict the photovoltaic output data for the future period, the future period is first divided into multiple time points, and then the applicable prediction model for each time point is determined according to the optimal scoring matrix. Based on the applicable prediction model at each time point, the photovoltaic output prediction value at each time point in the future period is determined in turn in a rolling prediction manner.

[0042] The embodiment of the present invention comprehensively models multiple features, thereby enhancing the expressive power and adaptability of the features of the input model; based on the multi-model fusion method of the optimal scoring matrix, in the process of rolling prediction, it is determined that the prediction model adopted at each time point is the optimal choice under the current circumstances, that is, by selecting the most appropriate prediction model in turn during the prediction process, the prediction accuracy of medium- and long-term photovoltaic output is improved, so that more accurate energy storage planning can be carried out subsequently.

[0043] On the basis of the above embodiment, optionally, the combined features include fixed period features, dynamic period features based on similar pattern matching, volatility features based on local changes and other auxiliary features.

[0044] Specifically, 1) fixed period feature means that multiple historical values ​​are selected from the historical sequence with a certain step length as the period as input features. According to the analysis of photovoltaic output characteristics and autocorrelation sequence, the correlation coefficients of the same time in adjacent days are all above 0.8. Therefore, the day is selected as the step length (96 time points), and m steps are selected forward (in this experiment, m = 12) to construct fixed period features, such as Figure 1C shown.

[0045] In the figure, P(d,h) represents the predicted photovoltaic output value at the hth time point on the dth day, and v(dn,h) represents the output value at the same time point with a delay of n days, where n is a positive integer, and the periodic feature sequence C = {v(dn,h)n = 1,2,…,m; h∈[1,96]} is obtained. The fixed periodic feature expresses the data change trend at a specific time scale.

[0046] 2) Affected by meteorology and geographical location, photovoltaic output has certain volatility and randomness. Dynamic periodic features are constructed to cope with the impact of changes in photovoltaic output. The dynamic periodic feature is based on the current pattern string pattern_t (defined as a sequence of continuous data points with a certain time span in the historical sequence close to the current time t). Through similarity calculation, the historical output sequence (such as top K) with high trend and numerical change consistency with pattern_t is matched from the historical pattern library, and the value of the last position is selected to form the input feature sequence D_t = {d_i, i = 1, 2, ..., K}, where K is the feature length. Dynamic feature selection is based on an assumption that at the same time scale, the sequence matched based on similarity has consistent data point change trends and numerical laws in most time periods, so the changes of single data points in local locations should also be consistent. The process of constructing dynamic features is as follows:

[0047] In the training phase, the sliding window size is set (such as 96), the current pattern string is determined (a certain time scale is set, such as an interval of 1 time point or 96 time points, etc.), the used historical sequence is split into multiple pattern sequences and traversed in sequence, and the sequence similarity values ​​except the last position data point are calculated according to a similarity algorithm, and the largest K (indicating the dynamic feature length) sequences are selected. The data points at the last position are taken to form a dynamic feature sequence. The label is designed to correspond to the time point feature of the sequence (such as month-day-time point), which expresses the consistency of photovoltaic output data at different times, and finally combined into a pattern library Pattern_Tag with different labels for the corresponding site.

[0048] In the prediction stage, according to each prediction time point i, the pattern string pattern_i to be matched is combined according to the same time scale, and the most relevant feature sequence D_i is selected from Pattern_Tag based on similarity calculation, that is, D_i = Similarity(Pattern_Tag, pattern_i). Similarity is a certain type of similarity calculation method. In actual scenarios, Euclidean distance, cosine distance, and dynamic time warping (DTW) algorithm are generally selected. In this invention, we mainly use DTW algorithm to calculate pattern similarity.

[0049] The basic idea of ​​the DTW algorithm is to measure the similarity of two time series by calculating the minimum distance between them. According to the principle of shortest distance, the correspondence between the elements of two sequences (A / B) of different lengths is constructed; the construction process is mainly divided into two parts: constructing the cumulative distance matrix and finding the shortest path. When constructing the correspondence between the elements of two sequences, the sequences need to be extended or compressed, as follows Figure 1D shown.

[0050] 3) The volatility characteristics based on local changes refer to the statistical characteristics of the peak and valley values ​​of photovoltaic output, fluctuation variance, fluctuation amplitude, climbing times, etc. that describe the extreme value and stability of output changes in a period of time before the forecast target. The definition and calculation of the main volatility characteristics are as follows:

[0051] γ i =v i+1 -v i (1)

[0052]

[0053] Among them, v i+1 , v i It represents the photovoltaic output value at adjacent moments, and n is the size of the statistical window. The fluctuation amplitude FA and the fluctuation rate FR reflect the fluctuation degree of photovoltaic output at adjacent moments, which are usually calculated by the average value within a time window. The smaller the value, the smaller the fluctuation and the better the stability.

[0054]

[0055] Among them, t j , t i represents the j, i time point, v j , v i represents the photovoltaic output value at time point j, i, ρ m It represents the climbing rate of a certain climbing; N represents the number of climbing times in a period of time, and the climbing rate RR represents the average climbing rate in a period of time. The climbing rate fully describes the ups and downs of the photovoltaic output time series.

[0056]

[0057] A=P max -P min (8)

[0058] in, Represents the average value of photovoltaic output during the statistical period, P max , P min Respectively represent the maximum and minimum values ​​of photovoltaic output; the fluctuation variance σ reflects the degree to which photovoltaic output deviates from its average value at any time within a period of time. The larger the fluctuation variance, the more violent the fluctuation; the maximum fluctuation amplitude A reflects the maximum range of photovoltaic output within a period of time. The larger A is, the greater the fluctuation. In addition, the peak-to-valley average range PP, peak-to-valley mean PM, etc. are also used as volatility feature sequences. The photovoltaic output volatility feature sequence W_t = {FA, FR, RP, σ, A, PP, PM} at time t is obtained.

[0059] 4) Other auxiliary features

[0060] Meteorological factors have the most direct impact on photovoltaic output, especially meteorological elements such as shortwave and solar irradiance. Through correlation coefficient analysis, meteorological characteristics such as downward solar shortwave radiation (Surface Short-wave Radiation Downwards, SSRD) reaching the surface, total sky direct solar radiation (Total Sky Direct Solar Radiation at Surface, FDIR) received by the ground, and direct solar radiation (Direct Solar Radiation, DSRP) are selected. See Table 1 for details.

[0061] Table 1 Main meteorological factors affecting photovoltaic output

[0062]

[0063]

[0064] In addition, One-Hot coding feature conversion is performed on season, month, and time point, where the coding length of season is 4, the coding length of month is 12, and the coding length of time point is 96.

[0065] Figure 2A A flowchart of a photovoltaic output prediction method for supporting energy storage planning based on an optimal scoring matrix is ​​provided in another embodiment of the present invention. This embodiment is optimized and improved on the basis of the above embodiment. Figure 2A As shown, the method includes:

[0066] S210, obtaining the combined features of different time scales of the photovoltaic output training data set; for the combined features of each time scale, using the extreme gradient boosting tree algorithm of the boosting integrated architecture as a medium- and long-term distributed photovoltaic output prediction learning framework to train the prediction model of the time scale.

[0067] Specifically, the present invention adopts the extreme gradient boosting tree (eXtreme GradientBoosting, XGBoost) algorithm based on the boosting integrated architecture as the medium- and long-term distributed photovoltaic output prediction learning framework. The basic idea of ​​Boosting integrated technology is to perform serial training by combining multiple base (weak) learners, and the latter base learner corrects the error of the previous base learner. Finally, all base learners are integrated into a strong learner, which is a trained prediction model. The model function based on the Boost integrated learning architecture is generally expressed as:

[0068]

[0069] Where K is the number of base learners (usually decision trees are used as base learners). The basic idea is to use the distributed forward additive model to iteratively improve and generate base learners by optimizing the structured loss function. The principle of the XGBoost model is as follows:

[0070] First, construct the objective function. Introduce the regularization term into the loss function and get the basic objective function:

[0071]

[0072] In the formula, the first term represents the loss function of the model, which is the loss between the true value and the predicted value of each sample; the second term represents the complexity of the tree, such as the number of tree layers, the number of leaf nodes, etc. Since the algorithm uses superposition training, that is, the predicted value of k models is equal to the predicted value of the previous k models plus the predicted value of the kth model currently being trained, we have:

[0073]

[0074] Introducing the objective function, we get:

[0075]

[0076] For the kth tree being trained, the first k-1 trees are all constants, so the objective function becomes:

[0077]

[0078] in, is the predicted value of the first k-1 tree models, also a constant. The objective function is approximated by the Taylor series. Definition:

[0079]

[0080] Because g i and h i They are all related to the first k-1 learners, so they are all constants. is also a constant and has nothing to do with optimizing the kth tree, so omitting the constant term yields:

[0081]

[0082] Then the tree structure is introduced into the objective function to optimize the unknown quantity f k (x i ) and Ω(f k ), it needs to be parameterized to obtain the parameterized objective function:

[0083]

[0084] Among them, T represents the number of leaf nodes, represents the sum of the values ​​of each leaf node, α and λ are hyperparameters that control the degree of penalty. After introducing the parametric expression of the tree structure, a greedy algorithm is used to construct K optimal trees (to minimize the objective function), and finally all the trees are weighted to form our final prediction model:

[0085]

[0086] For each time scale combination feature of the photovoltaic output training data, the prediction model of the corresponding time scale can be obtained through this framework and training process.

[0087] S220: construct an initial matrix, wherein the rows and columns of the initial matrix respectively represent different prediction models and different time points.

[0088] S230. For each time point, determine the prediction performance of each prediction model at that time point through the photovoltaic output backtest data set, and increase the number of selections of the prediction model with the best prediction performance at that time point in the initial matrix by one; normalize the initial matrix to obtain the optimal scoring matrix.

[0089] Specifically, the scoring matrix M is initialized, and a standard time length (e.g., 1 month) is set, with one matrix for each standard time length. The rows of the matrix represent the prediction models, and the columns represent the time points of the standard time length, all of which are initially set to 0.

[0090] For each trained prediction model, the training set RD is backtested respectively, and the residual between the predicted value and the true value of all prediction models at all time points is calculated as the prediction performance of the prediction model.

[0091] According to the size of the residual value, select the model at the time point corresponding to the minimum value, and update the number of times of the corresponding position in M, that is, M[mid_i, timepoint_i] + = 1;

[0092] After all models are backtested and updated at all time points, the number of selections in the normalized matrix M is used as the score to obtain the optimal score matrix.

[0093] For example, there are two prediction models for photovoltaic output. The first one is a prediction model M1 obtained by extracting features at a scale of 1 point (interval of 15 minutes); the second one is a prediction model M96 obtained by extracting features at a scale of 96 points (interval of a whole day); if the current time is 12:00 on day d, the output value at 13:00 on day d+1 needs to be predicted; if only the first model is used, the prediction path is: M1[d(12:15)], M1[d(12:30)], …, M1[d+1(12: 45)], M1[d+1(13:00)]; rolling forecast 99 times, each prediction needs to be based on the previous result, which increases the risk of error accumulation; if there are two prediction models, then the prediction path can be: M96[d+1(12:00)], M1[d+1(12:15)], M1[d+1(12:30)], M1[d+1(12:45)], M1[d+1(13:00)]; rolling forecast 5 times, try to avoid less rolling to reduce the risk of error accumulation. As for the performance at different time points, it is mainly analyzed from the perspective of probability. In the backtest data, which models perform well and which perform poorly at each data point, and finally convert them into probability choices according to the performance statistics, which can be understood as the preference weight of any model at any point.

[0094] S240, dividing the future time period to be predicted into multiple future time points; for each future time point, determining the prediction model with the highest score for that future time point in the optimal scoring matrix as the applicable prediction model for that future time point; based on the applicable prediction model for each future time point, determining the photovoltaic output prediction value for each future time point in the future time period in turn in a rolling prediction manner.

[0095] Specifically, Figure 2B As shown, the future time period to be predicted is divided into multiple (N) future time points, and prediction is performed point by point.

[0096] For any future time point, match the position of the current predicted time point i in the optimal score [*,i], that is, the column where the predicted time point i is located.

[0097] Get the probability of each prediction model being selected in the column where the prediction time point is located;

[0098] According to the probability maximization principle, the position with the largest probability value is selected, the prediction model corresponding to the row of the corresponding position is used as the applicable prediction model at the current prediction time point, the combined features are input into the applicable prediction model, and the photovoltaic output prediction value at the current prediction time point is output.

[0099] S250. Determine, based on the photovoltaic output forecast value, a first energy storage power configuration requirement required for energy storage to smooth power fluctuations and achieve grid-connected power quality, and a second energy storage power configuration requirement required to track a scheduling plan curve.

[0100] S260: Determine the larger of the first energy storage power configuration requirement and the second energy storage power configuration requirement as the actual maximum planned energy storage configuration requirement.

[0101] Specifically, based on the predicted distributed photovoltaic output data, the power range required for energy storage to smooth power fluctuations and reach the grid-connected power quality is analyzed. The active power of renewable energy power generation has a large volatility, and its power fluctuation can be smoothed using the sliding average method. The sliding average method is a method that selects a fixed-length window, takes the arithmetic average of all values ​​in the window, and uses the average value as the value of the window center point. This method can be used to obtain the smoothed renewable energy output curve, and compared with the curve before smoothing, the power demand that the energy storage system needs to meet can be obtained.

[0102]

[0103] In the above formula, N is the window length; P t is the predicted photovoltaic output value at the tth moment in the predicted future period; P ft is the power after smoothing; P ES0 Configure the required energy storage power.

[0104] From another dimension, combined with the scheduling plan power generation curve, the power range required to track the scheduling plan is calculated.

[0105] P ES1 =max[P a (i)-P g (i)] (22)

[0106] In the above formula, P a (i) is the active power required by the dispatch plan at the i-th moment; P g (i) is the predicted photovoltaic output value at the tth moment in the predicted future period; P ES1 Configure the required energy storage power required to track the dispatch plan curve.

[0107] Finally, combined with P ES0 and P ES1 , take the larger value as the actual maximum planning configuration requirement for energy storage.

[0108] For example, the performance of different features and their combinations are compared. For the convenience of display, the fixed period feature is represented as T1, the dynamic period feature is represented as T2, the local volatility feature is represented as T3, the meteorological feature is represented as T4, and the time coding feature is represented as T5. A variety of feature combination schemes are designed, such as T1+T4, which represents the combination of fixed period features and meteorological features, and the others are similar. SX region and SC region represent Shanxi region and Sichuan region respectively. The results of multiple photovoltaic power stations in the same region in the four typical months of the experiment (March, June, September, and December) are averaged. The results of this section are carried out according to the basic rolling prediction method. The specific results are shown in the following table:

[0109] Table 2 Comparison of different feature combinations in the same model

[0110]

[0111] By comparing the results of multiple features and their combinations, in the case of single features, T1 and T2 have comparable effects, and the mean absolute percentage error (MAPE) values ​​in the SX and SC regions are both 0.19, which is better than the result of the single T3 feature; in terms of accuracy index comparison, the two regions also show consistent results. In the case of combined features, the five features are combined, and the index performance is the best under the set experimental conditions, among which the mean absolute percentage error in the SX region is 0.16, R2 is 0.94, and ACC is 0.85; the mean absolute percentage error in the SC region is 0.19, R2 is 0.92, and ACC is 0.84, which are better than other combinations; in particular, the combination of T2 and T4 is more ideal. From the average results of distributed photovoltaic power stations in different regions, the overall performance of the SX region is slightly better than that of the SC region. It shows that the designed combined features are effective in photovoltaic output prediction. By comparing the average conditions of multiple groups of different feature combinations and photovoltaic power stations in different regions, it is shown that the stability of the combined features is also better than any single feature.

[0112] The model framework results proposed in the present invention are compared with some popular models in current research, and models such as LSTM, ET, SVM, Transformer, RF, and LightGBM are selected to further illustrate the advantages of the model framework and strategy proposed in the present invention. It should be noted that in the experimental verification of different models, the combined features (T1+T2+T3+T4+T5) are fixedly used as the combined feature input of all models, and the optimal scoring strategy is used for iteration. The comparison of the results of various models is shown in the following table:

[0113] Table 3 Comparison of different models for the same set of features

[0114]

[0115] Among them, M2E-DPV* is the abbreviation of the main model framework and strategy proposed in this invention. Judging from the indicators of the two regions, several comparative indicators of the M2E-DPV* method have significantly improved performance compared with other comparative models, and the indicator results in different regions are more stable, which further illustrates the advantages of the model framework strategy proposed in this invention.

[0116] Finally, according to the above experimental results, the forecasting capabilities of the two regions are statistically analyzed and the forecast trends are displayed, such as Figure 2C and Figure 2D .in, Figure 2C The middle left and right parts are the forecast capability statistics of the three indicators for SX and SC regions for 30 consecutive days. Figure 2D In the middle, a comparison chart of the predicted value and actual value is extracted with a certain time point as the starting time and predicted for the next 30 days (30*96 data points).

[0117] The embodiment of the present invention determines the planning configuration requirements of the corresponding energy storage through the predicted photovoltaic output prediction data, thereby improving the utilization value of the photovoltaic output prediction data.

[0118] Figure 3 A schematic diagram of the structure of a photovoltaic output prediction device supporting energy storage planning based on an optimal scoring matrix is ​​provided as another embodiment of the present invention. Figure 3 As shown, the device comprises:

[0119] A model training module 310, for training multiple prediction models of different time scales according to the combined features of the photovoltaic output training data set;

[0120] A model testing module 320 is used to perform applicable model testing at each time point according to the multiple prediction models to obtain an optimal scoring matrix;

[0121] The model using module 330 is used to determine the photovoltaic output prediction value in the future period based on the optimal scoring matrix, the photovoltaic output historical data and the multiple prediction models.

[0122] The photovoltaic output prediction device for supporting energy storage planning based on the optimal scoring matrix provided in an embodiment of the present invention can execute the photovoltaic output prediction method for supporting energy storage planning based on the optimal scoring matrix provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0123] Optionally, the model training module 310 includes:

[0124] A feature extraction unit, used to obtain the combined features of different time scales of the photovoltaic output training data set;

[0125] The model training unit is used to train the prediction model of each time scale by using the extreme gradient boosting tree algorithm of the boosting integrated architecture as a medium- and long-term distributed photovoltaic output prediction learning framework for the combined features of each time scale.

[0126] Optionally, the model testing module 320 includes:

[0127] A matrix construction unit, used to construct an initial matrix, wherein the rows and columns of the initial matrix represent different prediction models and different time points respectively;

[0128] A matrix updating unit, for each time point, determining the prediction performance of each prediction model at the time point through the photovoltaic output backtest data set, and increasing the number of selections of the prediction model with the best prediction performance at the time point in the initial matrix by one;

[0129] The matrix normalization unit is used to normalize the initial matrix to obtain an optimal scoring matrix.

[0130] Optionally, the model using module 330 includes:

[0131] A time period division unit, used for dividing the future time period to be predicted into multiple future time points;

[0132] A model determination unit, for determining, at each time point, the prediction model with the highest score at that time point in the optimal scoring matrix as the applicable prediction model at that time point;

[0133] The model using unit is used to determine the photovoltaic output forecast value at each time point in the future period in a rolling forecast manner based on the applicable forecast model at each time point.

[0134] Optionally, the combined features include fixed period features, dynamic period features based on similar pattern matching, volatility features based on local changes, and other auxiliary features.

[0135] Optionally, the device further comprises:

[0136] A demand calculation unit, used to determine a first energy storage power configuration requirement required for energy storage to smooth power fluctuations and achieve grid-connected power quality, and a second energy storage power configuration requirement required to track a scheduling plan curve according to the photovoltaic output forecast value;

[0137] The demand selection unit is used to determine the larger of the first energy storage power configuration demand and the second energy storage power configuration demand as the actual maximum planned configuration demand of energy storage.

[0138] The photovoltaic output prediction device for supporting energy storage planning based on the optimal scoring matrix further described can also execute the photovoltaic output prediction method for supporting energy storage planning based on the optimal scoring matrix provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0139] Figure 4 A schematic diagram of the structure of an electronic device 40 that can be used to implement an embodiment of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0140] like Figure 4 As shown, the electronic device 40 includes at least one processor 41, and a memory connected to the at least one processor 41, such as a read-only memory (ROM) 42, a random access memory (RAM) 43, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 41 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 42 or the computer program loaded from the storage unit 48 to the random access memory (RAM) 43. In the RAM 43, various programs and data required for the operation of the electronic device 40 can also be stored. The processor 41, the ROM 42, and the RAM 43 are connected to each other through a bus 44. An input / output (I / O) interface 45 is also connected to the bus 44.

[0141] A number of components in the electronic device 40 are connected to the I / O interface 45, including: an input unit 46, such as a keyboard, a mouse, etc.; an output unit 47, such as various types of displays, speakers, etc.; a storage unit 48, such as a disk, an optical disk, etc.; and a communication unit 49, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 49 allows the electronic device 40 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0142] The processor 41 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 41 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 41 executes the various methods and processes described above, such as a photovoltaic output prediction method supporting energy storage planning based on an optimal scoring matrix.

[0143] In some embodiments, the photovoltaic output prediction method for supporting energy storage planning based on the optimal scoring matrix may be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 48. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 40 via the ROM 42 and / or the communication unit 49. When the computer program is loaded into the RAM 43 and executed by the processor 41, one or more steps of the photovoltaic output prediction method for supporting energy storage planning based on the optimal scoring matrix described above may be performed. Alternatively, in other embodiments, the processor 41 may be configured to execute the photovoltaic output prediction method for supporting energy storage planning based on the optimal scoring matrix in any other appropriate manner (e.g., by means of firmware).

[0144] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0145] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0146] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in combination with an instruction execution system, device or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0147] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0148] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0149] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.

[0150] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0151] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A photovoltaic output prediction method for supporting energy storage planning based on an optimal scoring matrix, characterized in that: The method comprises: Multiple prediction models at different time scales are trained based on the combined features of the PV output training dataset; Performing applicable model tests at each time point according to the multiple prediction models to obtain an optimal scoring matrix; Based on the optimal scoring matrix, the photovoltaic output forecast value for the future period is determined according to the photovoltaic output historical data and the multiple forecast models.

2. The method according to claim 1, characterized in that The method of training multiple prediction models of different time scales according to the combined features of the photovoltaic output training data set includes: Obtain the combined features of different time scales of the photovoltaic output training dataset; For each time scale combination feature, the extreme gradient boosting tree algorithm with enhanced integrated architecture is used as the learning framework for medium- and long-term distributed photovoltaic output prediction to train the prediction model for this time scale.

3. The method according to claim 1, characterized in that The performing applicable model tests at each time point according to the multiple prediction models to obtain the optimal scoring matrix comprises: Constructing an initial matrix, wherein the rows and columns of the initial matrix respectively represent different prediction models and different time points; For each time point, the prediction performance of each prediction model at that time point is determined by the photovoltaic output backtest data set, and the number of selections of the prediction model with the best prediction performance at that time point is increased by one in the initial matrix; The initial matrix is ​​normalized to obtain an optimal scoring matrix.

4. The method according to claim 2, characterized in that: Based on the optimal scoring matrix, determining the photovoltaic output forecast value for the future period according to the photovoltaic output historical data and the multiple forecasting models includes: Divide the future period to be predicted into multiple future time points; For each time point, the prediction model with the highest score at that time point in the optimal scoring matrix is ​​determined as the applicable prediction model at that time point; Based on the applicable prediction model at each time point, the photovoltaic output forecast value at each time point in the future period is determined in turn by rolling prediction.

5. The method according to any one of claims 1 to 4, characterized in that: The combined features include fixed period features, dynamic period features based on similar pattern matching, volatility features based on local changes, and other auxiliary features.

6. The method according to claim 1, characterized in that After determining the photovoltaic output forecast value for the future period based on the optimal scoring matrix and according to the photovoltaic output historical data and the multiple forecasting models, the method further includes: Determine, according to the photovoltaic output forecast value, a first energy storage power configuration requirement required for energy storage to smooth power fluctuations and achieve grid-connected power quality, and a second energy storage power configuration requirement required to track the scheduling plan curve; The larger of the first energy storage power configuration requirement and the second energy storage power configuration requirement is determined as the actual maximum planned configuration requirement for energy storage.

7. A photovoltaic output prediction device supporting energy storage planning based on an optimal scoring matrix, characterized in that: The device comprises: A model training module is used to train multiple prediction models of different time scales based on the combined features of the photovoltaic output training data set; A model testing module, used to perform applicable model testing at each time point according to the multiple prediction models to obtain an optimal scoring matrix; The model using module is used to determine the photovoltaic output forecast value for a future period based on the optimal scoring matrix, the photovoltaic output historical data and the multiple forecasting models.

8. The device according to claim 7, characterized in that The model training module includes: A feature extraction unit, used to obtain the combined features of different time scales of the photovoltaic output training data set; The model training unit is used to train the prediction model of each time scale by using the extreme gradient boosting tree algorithm of the boosting integrated architecture as a medium- and long-term distributed photovoltaic output prediction learning framework for the combined features of each time scale.

9. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the photovoltaic output prediction method for supporting energy storage planning based on the optimal scoring matrix as described in any one of claims 1-6.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the photovoltaic output prediction method for supporting energy storage planning based on an optimal scoring matrix according to any one of claims 1 to 6 when executed.

Images