Day-ahead scheduling optimization method and device based on photovoltaic power prediction

Abstract

The invention discloses a day-ahead scheduling optimization method and device based on photovoltaic power prediction, and the method comprises the steps: obtaining historical meteorological data, historical photovoltaic power data and day-ahead load before a target date, carrying out the preprocessing and screening of the historical meteorological data based on the historical photovoltaic power data, obtaining target historical meteorological data, and carrying out the prediction of the target historical meteorological data; and inputting the target historical meteorological data into the photovoltaic power prediction model to obtain photovoltaic predicted power of a target date, and constructing a day-ahead scheduling optimization model of the target date based on the photovoltaic predicted power and the day-ahead load to solve an optimal scheduling scheme of the target date. Since the meteorological data is combined for power prediction, the accuracy of the obtained photovoltaic predicted power of the target date is higher, the day-ahead scheduling optimization model is constructed through the obtained photovoltaic predicted power and the day-ahead load to solve the optimal scheduling scheme, scheduling optimization based on the day-ahead load and the predicted photovoltaic power is provided, and the scheduling efficiency is improved. Reasonable and accurate scheduling optimization can be carried out, and the operation cost of a power grid is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of power grid dispatching, and in particular to a method and device for optimizing day-ahead dispatching based on photovoltaic power prediction. Background Art

[0002] Photovoltaic power generation not only plays a key role in reducing greenhouse gas emissions, but also accelerates the transformation of the energy structure to a green and low-carbon one. As more and more distributed photovoltaic systems are connected to the power grid, the role of photovoltaic power prediction in the power system becomes increasingly important, which is of great significance to maintaining the normal operation and stability of the power system.

[0003] At present, the photovoltaic power of the previous day is mainly used as the photovoltaic predicted power of the next day. On the one hand, it is impossible to accurately determine the photovoltaic predicted power of the next day. On the other hand, how to reasonably optimize the day-ahead scheduling based on the photovoltaic predicted power of the next day in order to reduce the operating cost is a major problem and pain point at present. Summary of the invention

[0004] The present invention provides a method and device for optimizing day-ahead scheduling based on photovoltaic power prediction, so as to solve the problem that accurate photovoltaic prediction power cannot be obtained and reasonable optimization of day-ahead scheduling cannot be performed.

[0005] In a first aspect, the present invention provides a method for optimizing day-ahead scheduling based on photovoltaic power prediction, comprising:

[0006] Acquire historical meteorological data and historical photovoltaic power data before a target date, and acquire day-ahead load before the target date;

[0007] Preprocessing and screening the historical meteorological data based on the historical photovoltaic power data to obtain target historical meteorological data;

[0008] Inputting the target historical meteorological data into the photovoltaic power prediction model to obtain the photovoltaic prediction power on the target date;

[0009] A day-ahead scheduling optimization model for the target date is constructed based on the photovoltaic predicted power and the day-ahead load to solve the optimal scheduling plan for the target date.

[0010] In a second aspect, the present invention provides a day-ahead scheduling optimization device based on photovoltaic power prediction, comprising:

[0011] A historical data acquisition module, used to acquire historical meteorological data and historical photovoltaic power data before a target date, and to acquire day-ahead load before the target date;

[0012] A data preprocessing and screening module, used for preprocessing and screening the historical meteorological data based on the historical photovoltaic power data to obtain target historical meteorological data;

[0013] A photovoltaic power prediction module, used for inputting target historical meteorological data into a photovoltaic power prediction model to obtain photovoltaic prediction power on the target date;

[0014] The scheduling optimization module is used to construct a day-ahead scheduling optimization model for the target date based on the photovoltaic predicted power and the day-ahead load to solve the optimal scheduling plan for the target date.

[0015] In a third aspect, the present invention provides an electronic device, the electronic device comprising:

[0016] at least one processor; and

[0017] a memory communicatively connected to the at least one processor; wherein,

[0018] 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 day-ahead scheduling optimization method based on photovoltaic power prediction described in the first aspect of the present invention.

[0019] In a fourth aspect, the present invention provides a computer-readable storage medium storing computer instructions, wherein the computer instructions are used to enable a processor to implement the day-ahead scheduling optimization method based on photovoltaic power prediction described in the first aspect of the present invention when executed.

[0020] The embodiment of the present invention obtains historical meteorological data, historical photovoltaic power data and day-ahead load before a target date, pre-processes and screens historical meteorological data through historical photovoltaic power data to obtain target historical meteorological data, inputs historical meteorological data into a photovoltaic power prediction model to obtain photovoltaic predicted power on the target date, further constructs a day-ahead scheduling optimization model based on the day-ahead load and photovoltaic predicted power to solve the optimal scheduling scheme on the target date, predicts photovoltaic power on the target date through historical meteorological data, and compared with using the photovoltaic power of the previous day as the photovoltaic power of the next day, the photovoltaic predicted power on the target date has higher accuracy due to the combination of meteorological data for power prediction, and constructs a day-ahead scheduling optimization model through the obtained photovoltaic predicted power and day-ahead load to solve the optimal scheduling scheme, provides scheduling optimization based on day-ahead load and predicted photovoltaic power, can perform reasonable and accurate scheduling optimization, and reduce power grid operation costs.

[0021] 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

[0022] 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.

[0023] Figure 1 is a flow chart of a day-ahead scheduling optimization method based on photovoltaic power prediction provided by the first embodiment of the present invention;

[0024] Figure 2 is a flow chart of a day-ahead scheduling optimization method based on photovoltaic power prediction provided by Embodiment 2 of the present invention;

[0025] Figure 3 A schematic diagram of a network structure of a photovoltaic power prediction model according to an embodiment of the present invention;

[0026] Figure 4 It is a schematic diagram of the training process of the photovoltaic power prediction model;

[0027] Figure 5 This is a schematic diagram of the network structure of LSTM;

[0028] Figure 6 It is a flow chart of the optimization process of the day-ahead scheduling optimization model;

[0029] Figure 7 It is a structural schematic diagram of a day-ahead scheduling optimization device based on photovoltaic power prediction provided by Embodiment 3 of the present invention;

[0030] Figure 8 It is a schematic diagram of the structure of an electronic device provided in Embodiment 4 of the present invention. DETAILED DESCRIPTION

[0031] 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.

[0032] Embodiment 1

[0033] Figure 1 This is a flowchart of a method for optimizing day-ahead scheduling based on photovoltaic power prediction provided in the first embodiment of the present invention. This embodiment can be applied to the case of optimizing day-ahead scheduling by predicting photovoltaic power. This method can be executed by a device for optimizing day-ahead scheduling based on photovoltaic power prediction. The device for optimizing day-ahead scheduling based on photovoltaic power prediction can be implemented in the form of hardware and / or software and can be configured in an electronic device. Figure 1 As shown, the day-ahead scheduling optimization method based on photovoltaic power prediction includes:

[0034] S101. Acquire historical meteorological data and historical photovoltaic power data before a target date, and acquire day-ahead load before a target date.

[0035] This embodiment optimizes the day-ahead scheduling, so the target date is a date in days. For example, the target date is the date after the current date, that is, the day after the current day. The historical meteorological data and historical photovoltaic power data are the meteorological data and photovoltaic power data of the current day. The day-ahead load is the grid load of the current day. In short, the photovoltaic power of the next day is predicted based on the historical meteorological data and historical photovoltaic data of the current day, and the scheduling of the next day is optimized based on the predicted photovoltaic power and the load of the current day, that is, the day-ahead scheduling optimization.

[0036] Among them, historical meteorological data can be data collected according to a preset cycle during a day, which may include data such as light intensity, temperature, wind speed, humidity, etc. Historical photovoltaic power data can refer to the output power of a photovoltaic power station collected according to a preset cycle during a day. Similarly, the day-ahead load can refer to the load of the distribution network.

[0037] In one example, the output power and duration data of the photovoltaic power station of the day can be collected every 15 minutes to obtain historical meteorological data and historical photovoltaic data. For the day-ahead load, the load data of the day can be directly read from the power grid monitoring system.

[0038] S102: pre-process and filter the historical meteorological data based on the historical photovoltaic power data to obtain target historical meteorological data.

[0039] In an optional embodiment, outlier correction, missing value filling, and normalization processing can be performed on various types of meteorological data and historical photovoltaic power data in the historical meteorological data to obtain various types of first meteorological data and pre-processed historical photovoltaic power data, and then feature engineering processing is performed on each type of first meteorological data to obtain second meteorological data after feature engineering, and wavelet transform processing is performed on the pre-processed historical photovoltaic power data to obtain historical photovoltaic power data after wavelet transform processing, and the Pearson correlation coefficient between each type of second meteorological data and the historical photovoltaic power data after wavelet transform processing is calculated, and meteorological data with a Pearson correlation coefficient greater than a threshold is selected from the multiple types of second meteorological data to obtain target historical meteorological data.

[0040] Specifically, if a large amount of missing data occurs in a day, for example, 96 points should be collected in a day, but more than 30% of the data points are missing, the data collected on that day are unavailable. All data collected on that day are deleted, and the data from the previous day can be used to ensure data fluency.

[0041] If a small amount of data is missing during the day, considering that meteorological conditions are generally stable in a short period of time and the power generated by the photovoltaic power station is delayed, the missing data will be close to the adjacent data, and linear interpolation can be used to fill the data. For details, refer to the following formula to calculate the missing value:

[0042]

[0043] Among them, x1, x2, y1, and y2 are the values ​​of the points (x1, y1) and (x2, y2) respectively, and ψ(x) is the approximate value of the points (x1, y1) and (x2, y2), which can be used as missing values.

[0044] In addition to missing values, the data set may have outliers. If the outliers in the data collected for one day are accidental, it is considered that the data was interfered during collection. The average value of the data collected at the two previous and subsequent collection points can be used to replace them. However, if the outliers fluctuate greatly and last for a long time, the influence of extreme weather factors needs to be considered. The data cannot be directly replaced, but must be retained to ensure data integrity and improve the prediction accuracy of the model.

[0045] In addition to correcting outliers and supplementing missing values, the collected data can also be normalized to obtain data with values ​​between 0 and 1.

[0046] For power data, the collected power data signal can be processed to remove a certain amount of noise signal, because it includes periodic stable signals in the low-frequency band and random noise signals in the high-frequency band, so as to facilitate subsequent photovoltaic power prediction. Specifically, the historical photovoltaic power data can be processed with noise reduction by the three-layer wavelet improved threshold denoising method to obtain the historical photovoltaic power data after wavelet transform processing.

[0047] For each type of first historical meteorological data after preprocessing, feature engineering can be used to extract and reconstruct features. For example, feature engineering can use an interactive polynomial feature transformation method to construct features to obtain second meteorological data after feature engineering.

[0048] Furthermore, for different types of second meteorological data, the Pearson correlation coefficient between each type of second meteorological data and the historical photovoltaic power data after wavelet transform processing can be calculated. The specific calculation formula is as follows:

[0049]

[0050] m is the number of samples, x i ,y i is the observed value of point i corresponding to variables x and y, It is the average value of x and y samples. Exemplarily, the second meteorological data includes light intensity, temperature and humidity. The light intensity, temperature and humidity all include data from 96 collection points. The historical photovoltaic power data also includes the power of 96 points. For the Pearson correlation coefficient of light intensity and historical photovoltaic power data, the light intensity of 96 collection points and the power of 96 collection points can be used to calculate the Pearson correlation coefficient. Similarly, the Pearson correlation coefficients of temperature, humidity and historical photovoltaic power data can be calculated separately. The Pearson correlation coefficient can be negative, 0 or positive. A negative number indicates negative correlation, 0 indicates no correlation, and a positive number indicates positive correlation. The absolute value of the Pearson correlation coefficient can be taken, and the meteorological data whose absolute value of the Pearson correlation coefficient with the historical photovoltaic power data is greater than 0 will be retained.

[0051] The data preprocessing in this embodiment can ensure the quality of historical meteorological data, extract meteorological data with high correlation with photovoltaic power for power prediction, and improve the accuracy of photovoltaic power prediction.

[0052] S103: Input the target historical meteorological data into a photovoltaic power prediction model to obtain the photovoltaic prediction power on the target date.

[0053] The photovoltaic power prediction model of this embodiment can be a single neural network model or a hybrid neural network model. After training, the photovoltaic power prediction model can output photovoltaic predicted power when meteorological data is input. For example, the meteorological data of the day before (the current day) can be input into the photovoltaic power prediction model to obtain the photovoltaic predicted power for the next day.

[0054] S104. Construct a day-ahead dispatch optimization model for the target date based on the predicted photovoltaic power and the day-ahead load to solve the optimal dispatch plan for the target date.

[0055] In one embodiment, in addition to photovoltaic power stations, the distribution network also includes conventional units (thermal power, hydropower, etc.) and electrochemical energy storage power stations. Multiple scenarios can be generated through a scenario generation algorithm. The scenario data of each scenario includes the probability of the scenario occurring and the output data of conventional units, electrochemical energy storage power stations, and photovoltaic power stations. Then, a day-ahead scheduling optimization model is constructed through the scenario data. The day-ahead scheduling optimization model includes an objective function, which is the minimum value of the sum of the cost functions of conventional units, electrochemical energy storage power stations, photovoltaic power station units, and loads when multiple constraints are met. The optimal scenario can be obtained by solving the minimum value of the objective function. The output of conventional units, electrochemical energy storage power stations, and photovoltaic power stations in the optimal scenario is the optimal scheduling plan.

[0056] The embodiment of the present invention obtains historical meteorological data, historical photovoltaic power data and day-ahead load before a target date, pre-processes and screens historical meteorological data through historical photovoltaic power data to obtain target historical meteorological data, inputs historical meteorological data into a photovoltaic power prediction model to obtain photovoltaic predicted power on the target date, further constructs a day-ahead scheduling optimization model based on the day-ahead load and photovoltaic predicted power to solve the optimal scheduling scheme on the target date, predicts photovoltaic power on the target date through historical meteorological data, and compared with using the photovoltaic power of the previous day as the photovoltaic power of the next day, the photovoltaic predicted power on the target date has higher accuracy due to the combination of meteorological data for power prediction, and constructs a day-ahead scheduling optimization model through the obtained photovoltaic predicted power and day-ahead load to solve the optimal scheduling scheme, provides scheduling optimization based on day-ahead load and predicted photovoltaic power, can perform reasonable and accurate scheduling optimization, and reduce power grid operation costs.

[0057] Embodiment 2

[0058] Figure 2 Flow chart of a method for optimizing day-ahead scheduling based on photovoltaic power prediction provided by Embodiment 2 of the present invention. Embodiment 2 of the present invention is optimized on the basis of Embodiment 1 above. Figure 2 As shown, the day-ahead scheduling optimization method based on photovoltaic power prediction includes:

[0059] S201, obtaining historical meteorological data and historical photovoltaic power data before a target date, and obtaining day-ahead load before a target date;

[0060] S202, preprocessing and screening the historical meteorological data based on the historical photovoltaic power data to obtain target historical meteorological data;

[0061] S201-202 in this embodiment can refer to S101-102 in the first embodiment and will not be described in detail here.

[0062] S203, inputting the target historical meteorological data into the one-dimensional CNN network and the LSTM network respectively.

[0063] The photovoltaic power prediction model of this embodiment can be a model composed of a one-dimensional CNN network and an LSTM network, such as Figure 3 The figure shows a schematic diagram of the network structure of the photovoltaic power prediction model. The photovoltaic power prediction model includes a one-dimensional CNN network, an LSTM network, a feature fusion network and a power prediction network. The output layers of the one-dimensional CNN network and the LSTM network are connected to the input layer of the feature fusion network, and the output layer of the feature fusion network is connected to the input layer of the power prediction network.

[0064] The photovoltaic power prediction model of this embodiment can be trained by the following steps:

[0065] S1. Obtain a training data set, where the training data set includes preprocessed meteorological sample data and a marked first photovoltaic power;

[0066] S2. Construct a photovoltaic power prediction model including a CNN network and an LSTM network, and initialize the network parameters of the photovoltaic power prediction model;

[0067] S3, extracting meteorological sample data from the training data set and inputting it into the photovoltaic power prediction model to obtain the second photovoltaic power;

[0068] S4, calculating the loss rate based on the first photovoltaic power and the second photovoltaic power;

[0069] S5, determine whether the training stop condition is met, if so, execute S6, if not, execute S7;

[0070] S6. Determine that the photovoltaic power prediction model has completed training;

[0071] S7. Use the loss rate to adjust the model parameters of the photovoltaic power prediction model, and return to S3.

[0072] Specifically, Figure 4The figure shows a schematic diagram of the training process. Data samples can be generated by using historical data from multiple days after outlier processing, weather classification, feature changes, and normalization. The daily sample data includes meteorological data and real photovoltaic power (first photovoltaic power). The data samples are divided into training samples and test samples. The training samples are input into the neural network for training. During training, the predicted photovoltaic power (second photovoltaic power) and the real power are used to calculate the loss rate (such as calculating the mean square error loss) and determine whether the conditions for stopping training are met. For example, if the number of training times reaches the preset number of times, or the loss rate is less than the threshold, it is determined that the conditions for stopping training are met. If not, the network parameters are adjusted. After the training is completed, the test sample is input into the trained neural network model to obtain the result, and the result is verified with the real value. If the verification meets the requirements, the training is determined to be over. If the verification does not meet the requirements, the network parameters are adjusted to continue the training.

[0073] In a preferred embodiment, Dropout regularization and Adam algorithm (Adaptive Moment Estimation) can be selected to train the photovoltaic power prediction model to improve the performance, generalization ability and robustness of the photovoltaic power prediction model. Specifically, Dropout regularization is to randomly set the neuron activation value of a part (20%-50%) of the photovoltaic power prediction model to 0 during the training process, thereby reducing the sensitivity of the model to a single data point, so as to reduce the risk of overfitting and improve the robustness of the neural network. On the basis of using Dropout regularization, the Adam optimization algorithm with strong practicality and wide application range of deep learning is used. The Adam optimization algorithm is combined with the advantages of AdaGrad and RMSProp optimization algorithms to achieve adaptive learning rate adjustment, which can accelerate the training process and improve model performance.

[0074] After the photovoltaic power prediction model is trained, the target historical meteorological data can be input into the one-dimensional CNN network and the LSTM network respectively during photovoltaic power prediction.

[0075] S204, extracting spatial features of the target historical meteorological data in a one-dimensional CNN network, and extracting time series features of the target historical meteorological data in an LSTM network.

[0076] LSTM is a cyclic recursive neural network specially used for processing and predicting sequence data, aiming to solve the "long-term dependency problem" of traditional recurrent neural networks. Specifically, in the embodiment of the present invention, for photovoltaic power prediction, LSTM can well make up for the shortcomings of traditional recurrent neural networks and give full play to their advantages.

[0077] like Figure 5The figure shows the schematic diagram of the LSTM network structure. The main components of LSTM include cell state, input gate, forget gate, output gate, and hidden state. The cell state is the main channel that runs through the entire sequence and transmits important information. It has almost no attenuation and is maintained and updated through a gating mechanism. The input gate determines how much information in the current time step can be updated to the cell state. The forget gate determines which information in the cell state should be discarded based on the current input and the previous hidden state. The output gate is based on the current input and the previous hidden state to determine how much information to output in the next time step. The hidden state is used as an output to pass to the next layer or the next time step. The three gates are all processed by a fully connected layer with a sigmoid function to calculate the values ​​of the three. Therefore, the values ​​of the three are all in the range of (0,1).

[0078] The following mathematical expression explains the LSTM network, assuming that there are h hidden units, the batch size is n, and the number of inputs is d. So, X t ∈R n×d , the hidden state of the previous time step is H t-1 ∈R n×h , the gate at time step t can be defined as follows:

[0079] The input gate is I t ∈R n×h , the forget gate is F t ∈R n×h , the output gate is O t ∈R n×h , define candidate neurons Use the tanh function as the activation function, the function value range is (-1,1), and use the memory element C t To control input and forgetting, finally, define the hidden state H t , the specific mathematical expression is as follows:

[0080] I t =σ(X t W xi +H t-1 W hi +b i );

[0081] F t =σ(X t W xf +H t-1 W hf +b f );

[0082] O t =σ(X t W xo +H t-1 Who +b o );

[0083]

[0084] H t =O t ×tanh(C t );

[0085] Among them, W xi , W xf , W xo , W xc ∈R d×h and W hc ∈R h×h , is the weight coefficient; b i , b f , b o , b f ∈R 1×h , is the bias coefficient.

[0086] CNN is a deep learning algorithm. General CNN networks usually perform very well in image processing and visual recognition tasks. However, general CNN networks are two-dimensional convolutional networks and cannot process time series data. Therefore, a one-dimensional convolutional network (1D-CNN) suitable for processing time series data is used, and the ReLU function is selected as the activation function.

[0087] After the target historical meteorological data of this embodiment is input into the one-dimensional CNN network and the LSTM network, the time series feature X1 of the target historical meteorological data is extracted in the LSTM network, and the spatial feature X2 of the target historical meteorological data is extracted in the one-dimensional CNN network. The time series feature X1 and the spatial feature X2 are output to the feature fusion network.

[0088] S205. Concatenate and fuse the spatial features and the time series features in the feature fusion network to obtain fused meteorological features.

[0089] Specifically, the spatial feature X2 and the time series feature X1 are concatenated and fused in the feature fusion network to obtain the fused meteorological feature.

[0090] S206. Predicting the photovoltaic power forecast on the target date based on the meteorological characteristics in the power forecast network.

[0091] Specifically, in the power prediction network, the meteorological characteristics obtained by splicing and fusing the spatial feature X2 and the time series feature X1 are predicted to obtain the photovoltaic predicted power on the target date.

[0092] The photovoltaic power prediction model of this embodiment includes LSTM and one-dimensional CNN, giving full play to their respective advantages in extracting temporal features and spatial features, and adopting one-dimensional CNN, the entire model network structure is simple, which can not only ensure the accuracy of power prediction, but also reduce the amount of calculation and improve the efficiency of power prediction.

[0093] S207: Generate multiple scenes through a scene generation algorithm to obtain scene data of the multiple scenes.

[0094] The scenario in this embodiment may refer to the load and output conditions of conventional units, electrochemical energy storage power stations, photovoltaic power station units and power grid loads in the power grid at each moment. A scenario generation algorithm may be used to generate multiple scenarios in a day, and then the multiple scenarios may be reduced to obtain a target number of scenarios.

[0095] Exemplarily, multiple scenarios can be generated by a random programming method, and then the multiple scenarios can be clustered to reduce the scenarios to obtain a target number of scenarios. The scenario data of each scenario includes at least the probability of the scenario occurring and the number, output, status and other data of conventional units, electrochemical energy storage power stations, and photovoltaic power station units in each scenario.

[0096] Exemplarily, multiple scenarios can be generated by a Monte Carlo algorithm, and then a fast reduction algorithm based on probability distance is used to reduce the scenarios, reduce the scenarios to a target number, and give the probability of occurrence of each scenario. The specific scenario generation and reduction can refer to existing algorithms, and the embodiments of the present invention are not limited to this.

[0097] S208. Construct a day-ahead dispatch optimization model based on the generated scenario data of the multiple scenarios, the photovoltaic predicted power, and the day-ahead load.

[0098] Specifically, the expression of the constructed day-ahead scheduling optimization model is as follows:

[0099]

[0100] Among them, C t is the total operating cost of the power grid, C G,t , C EESS,t , C PV,t , C L,t are the cost functions of conventional units, electrochemical energy storage power stations, photovoltaic power station units and loads, respectively, N s 、N G 、N EESS 、N PV are the number of scenarios, the number of conventional units, the number of electrochemical energy storage power stations, and the number of photovoltaic power station units, respectively. Gi,t,s , P EESSi,t,s , P PVi,t,sare the outputs of conventional units, electrochemical energy storage power stations, and photovoltaic power station units at time t and in the sth scenario, respectively, and u Gi,t,s 、u EESSi,t,s 、u PVi,t,s are the start and stop states of conventional units, electrochemical energy storage power stations, and photovoltaic power station units at time t and in the sth scenario, respectively. s is the probability coefficient of the sth scenario, a i , b i 、c i is the power generation cost coefficient of conventional units, S i is the start-up and shutdown cost coefficient of conventional units, C(P EESSi,t,s )、M(P EESSi,t,s ) are the cost function and maintenance cost function of the electrochemical energy storage power station at time t and in the sth scenario, respectively. EESS is the unit time depreciation cost coefficient of the electrochemical energy storage power station, C(P PVi,t,s ) is the cost function of the photovoltaic power station unit at time t and in the sth scenario, k p,PV , k p,L are the penalty coefficient for abandoned light and the penalty cost coefficient for load power failure, P PVfor,t,s is the predicted value of the photovoltaic power station unit at time t and the sth scenario, k IDR is the coefficient of incentive demand response, Δ|P IDR,t,s | is the response amount of incentive demand in the sth scenario at time t, P L,t,s is the total load loss in the system at time t and in the sth scenario.

[0101] S209: Solve the day-ahead scheduling optimization model to obtain the optimal scheduling plan for the target date.

[0102] Find a solution that satisfies the following constraints so that C t The smallest scenario is taken as the target scenario, and the output of conventional units, electrochemical energy storage power stations, and photovoltaic power station units in the target scenario is determined to obtain the optimal scheduling plan for the target date, where the constraints include:

[0103] Power constraints:

[0104]

[0105] Among them, L fixed,t,s is the portion of the load that is not affected by electricity prices in the sth scenario at time t, ΔP PDR,t,s is the price-based demand response quantity in the sth scenario at time t, ΔP IDR,t,s is the amount of incentive demand response in the sth scenario at time t;

[0106] Output constraints of conventional units:

[0107] P Gi,min ≤P Gi,t,s ≤P Gi,max ;

[0108] |P Gi,t+1,s -P Gi,t,s |≤u Gi,t+1,s R i ;

[0109] Among them, P Gi,min , P Gi,max are the minimum and maximum output values ​​of each conventional unit, R i is the climbing power of each conventional unit;

[0110] Output constraints of photovoltaic power station units:

[0111] P PV,i,t,s ≤P PVfor,t,s ;

[0112] Output constraints of electrochemical energy storage power station:

[0113] P EESSfli,t,s ≤P EESSi,t,s ≤P EESSupi,t,s ;

[0114] S SOCmin ≤S SOC,t,s ≤S SOCmax ;

[0115] Among them, P EESSfli,t,s , P EESSupi,t,s are the lower and upper limits of the charge and discharge of the electrochemical energy storage power station at time t and in the sth scenario, respectively. SOC,t,s is the state of charge of the electrochemical energy storage power station at time t and the sth scenario, S SOCmin , S SOCmax They are the minimum and maximum charge states of the electrochemical energy storage power station respectively;

[0116] Scene constraints:

[0117] |P Gi,t,s -P Gi,t,bas |≤N Gi ;

[0118] |P EESSi,t,s -P EESSi,t,bas |≤N EESSi ;

[0119] Among them, P Gi,t,bas , P EESSi,t,bas are the basic output power of each conventional unit and electrochemical energy storage power station, N Gi 、N EESSiThey are the power fluctuation range constraint values ​​of each conventional unit and electrochemical energy storage power station respectively;

[0120] Network transmission power constraints:

[0121] |B ij (θ i,t,s -θ j,t,s )|≤P ijmax ;

[0122] Among them, B ij is the susceptance corresponding to line i, j, θ i,t,s is the phase angle of node i at time t and in the sth scenario, P ijmax Transmit maximum power to the line;

[0123] Constraints for load demand response:

[0124] P PDRmin ≤P PDR ≤P PDRmax ;

[0125]

[0126] Among them, P PDRmin , P PDRmax are the minimum and maximum values ​​of price-based demand response, is the load increase due to the incentive-based demand response in the sth scenario at time t, is the load reduction caused by the incentive-based demand response in the sth scenario at time t.

[0127] Figure 6 It is a flow chart of the optimization of the day-ahead dispatch optimization model. Specifically, for the day-ahead dispatch optimization, the day-ahead load and photovoltaic power forecast values ​​are input, and then the scenario generation and reduction technology is used to generate multiple typical scenarios, and an objective function with the lowest total grid operation cost is constructed. Then, it is determined whether each scenario meets the constraints. If so, the scenario with the smallest objective function value is used as the day-ahead dispatch plan. If not, the scenario is optimized again.

[0128] In this embodiment, the target historical meteorological data is input into a one-dimensional CNN network and an LSTM network respectively, spatial features are extracted from the target historical meteorological data in the one-dimensional CNN network, and time series features are extracted from the target historical meteorological data in the LSTM network. The spatial features and time series features are spliced ​​and fused in the feature fusion network to obtain fused meteorological features. The photovoltaic predicted power on the target date is predicted based on the meteorological features in the power prediction network. Multiple scenes are generated by a scene generation algorithm to obtain scene data of multiple scenes. A day-ahead scheduling optimization model is constructed based on the scene data of the generated multiple scenes, the photovoltaic predicted power, and the day-ahead load. The day-ahead scheduling optimization model is solved to obtain the optimal scheduling scheme on the target date. The photovoltaic power on the target date is predicted by the historical meteorological data. Compared with taking the photovoltaic power of the previous day as the photovoltaic power of the next day, the photovoltaic predicted power on the target date has higher accuracy due to the combination of meteorological data for power prediction. The day-ahead scheduling optimization model is constructed by the obtained photovoltaic predicted power and the day-ahead load to solve the optimal scheduling scheme, and scheduling optimization based on the day-ahead load and the predicted photovoltaic power is provided, which can perform reasonable and accurate scheduling optimization and reduce the operation cost of the power grid.

[0129] Embodiment 3

[0130] Figure 7 This is a schematic diagram of the structure of a day-ahead scheduling optimization device based on photovoltaic power prediction provided by the third embodiment of the present invention. Figure 7 As shown, the day-ahead dispatch optimization device based on photovoltaic power prediction includes:

[0131] The historical data acquisition module 701 is used to acquire the historical meteorological data and the historical photovoltaic power data before the target date, and acquire the day-ahead load before the target date;

[0132] A data preprocessing and screening module 702 is used to preprocess and screen the historical meteorological data based on the historical photovoltaic power data to obtain target historical meteorological data;

[0133] The photovoltaic power prediction module 703 is used to input the target historical meteorological data into the photovoltaic power prediction model to obtain the photovoltaic predicted power on the target date;

[0134] The scheduling optimization module 704 is used to construct a day-ahead scheduling optimization model for the target date based on the photovoltaic predicted power and the day-ahead load to solve the optimal scheduling plan for the target date.

[0135] Optionally, the data preprocessing and screening module 702 includes:

[0136] A preprocessing unit, used for performing outlier correction, missing value filling, and normalization processing on various types of meteorological data and historical photovoltaic power data in the historical meteorological data to obtain various types of first meteorological data and preprocessed historical photovoltaic power data;

[0137] A feature engineering unit, used for performing feature engineering processing on each type of first meteorological data to obtain second meteorological data after feature engineering;

[0138] A power data processing unit is used to perform wavelet transform processing on the pre-processed historical photovoltaic power data to obtain the historical photovoltaic power data after wavelet transform processing;

[0139] A correlation calculation unit, used for calculating the Pearson correlation coefficient between each type of second meteorological data and the historical photovoltaic power data after wavelet transformation;

[0140] The meteorological data screening unit is used to select meteorological data with a Pearson correlation coefficient greater than a threshold from multiple types of second meteorological data to obtain target historical meteorological data.

[0141] Optionally, the photovoltaic power prediction model includes a one-dimensional CNN network, an LSTM network, a feature fusion network and a power prediction network, the output layers of the one-dimensional CNN network and the LSTM network are connected to the input layer of the feature fusion network, and the output layer of the feature fusion network is connected to the input layer of the power prediction network;

[0142] The photovoltaic power prediction module 703 includes:

[0143] A meteorological data input unit, used to input the target historical meteorological data into the one-dimensional CNN network and the LSTM network respectively;

[0144] A spatial and temporal feature extraction unit, used for extracting spatial features from the target historical meteorological data in the one-dimensional CNN network, and extracting time series features from the target historical meteorological data in the LSTM network;

[0145] A feature fusion and splicing unit, used for splicing and fusing the spatial features and the time series features in the feature fusion network to obtain fused meteorological features;

[0146] A power prediction unit is used to predict the photovoltaic predicted power on the target date based on the meteorological characteristics in the power prediction network.

[0147] Optionally, the photovoltaic power prediction model is trained by the following modules:

[0148] A training data acquisition module, used to acquire a training data set, wherein the training data set includes preprocessed meteorological sample data and the marked first photovoltaic power;

[0149] A model building module is used to build a photovoltaic power prediction model including a CNN network and an LSTM network, and initialize the network parameters of the photovoltaic power prediction model;

[0150] A prediction module, used for extracting meteorological sample data from the training data set and inputting the data into a photovoltaic power prediction model to obtain a second photovoltaic power;

[0151] A loss rate calculation module, configured to calculate a loss rate based on the first photovoltaic power and the second photovoltaic power;

[0152] The training condition judgment module is used to judge whether the training stop condition is met; if so, the training completion determination module is executed; if not, the adjustment module is executed;

[0153] A training completion determination module, used to determine that the photovoltaic power prediction model has completed training;

[0154] The adjustment module is used to adjust the model parameters of the photovoltaic power prediction model by using the loss rate and return the result to the prediction module.

[0155] Optionally, the scheduling optimization module 704 includes:

[0156] A scenario generation unit, used to generate multiple scenarios through a scenario generation algorithm to obtain scenario data of the multiple scenarios, wherein the scenario data at least includes the probability of occurrence of the scenario and data of conventional units, electrochemical energy storage power stations, and photovoltaic power station units in each scenario;

[0157] A day-ahead dispatch optimization model building unit, configured to build a day-ahead dispatch optimization model based on the generated scenario data of the plurality of scenarios, the photovoltaic predicted power and the day-ahead load;

[0158] The optimization unit is used to solve the day-ahead scheduling optimization model to obtain the optimal scheduling plan for the target date.

[0159] Optionally, the expression of the constructed day-ahead scheduling optimization model is as follows:

[0160]

[0161] Among them, C t is the total operating cost of the power grid, C G,t , C EESS,t , C PV,t , C L,t are the cost functions of conventional units, electrochemical energy storage power stations, photovoltaic power station units and loads, respectively, N s 、NG 、N EESS 、N PV are the number of scenarios, the number of conventional units, the number of electrochemical energy storage power stations, and the number of photovoltaic power station units, respectively. Gi,t,s , P EESSi,t,s , P PVi,t,s are the outputs of conventional units, electrochemical energy storage power stations, and photovoltaic power station units at time t and in the sth scenario, respectively, and u Gi,t,s 、u EESSi,t,s 、u PVi,t,s are the start and stop states of conventional units, electrochemical energy storage power stations, and photovoltaic power station units at time t and in the sth scenario, respectively. s is the probability coefficient of the sth scenario, a i , b i 、c i is the power generation cost coefficient of conventional units, S i is the start-up and shutdown cost coefficient of conventional units, C(P EESSi,t,s )、M(P EESSi,t,s ) are the cost function and maintenance cost function of the electrochemical energy storage power station at time t and in the sth scenario, respectively. EESS is the unit time depreciation cost coefficient of the electrochemical energy storage power station, C(P PVi,t,s ) is the cost function of the photovoltaic power station unit at time t and in the sth scenario, k p,PV , k p,L are the penalty coefficient for abandoned light and the penalty cost coefficient for load power failure, P PVfor,t,s is the predicted value of the photovoltaic power station unit at time t and the sth scenario, k IDR is the coefficient of incentive demand response, Δ|P IDR,t,s | is the response amount of incentive demand in the sth scenario at time t, P L,t,s is the total load loss in the system at time t and in the sth scenario.

[0162] Optionally, the optimization unit includes:

[0163] The cost minimum solving subunit is used to solve the problem of C satisfying the following constraints. t The smallest scenario is taken as the target scenario, and the output of conventional units, electrochemical energy storage power stations, and photovoltaic power station units at the target scenario is determined to obtain the optimal scheduling plan for the target date, wherein the constraints include:

[0164] Power constraints:

[0165]

[0166] Among them, L fixed,t,s is the portion of the load that is not affected by electricity prices in the sth scenario at time t, ΔPPDR,t,s is the price-based demand response quantity in the sth scenario at time t, ΔP IDR,t,s is the amount of incentive demand response in the sth scenario at time t;

[0167] Output constraints of conventional units:

[0168] P Gi,min ≤P Gi,t,s ≤P Gi,max ;

[0169] |P Gi,t+1,s -P Gi,t,s |≤u Gi,t+1,s R i ;

[0170] Among them, P Gi,min , P Gi,max are the minimum and maximum output values ​​of each conventional unit, R i is the climbing power of each conventional unit;

[0171] Output constraints of photovoltaic power station units:

[0172] P PV,i,t,s ≤P PVfor,t,s ;

[0173] Output constraints of electrochemical energy storage power station:

[0174] P EESSfli,t,s ≤P EESSi,t,s ≤P EESSupi,t,s ;

[0175] S SOCmin ≤S SOC,t,s ≤S SOCmax ;

[0176] Among them, P EESSfli,t,s , P EESSupi,t,s are the lower and upper limits of the charge and discharge of the electrochemical energy storage power station at time t and in the sth scenario, respectively. SOC,t,s is the state of charge of the electrochemical energy storage power station at time t and the sth scenario, S SOCmin , S SOCmax They are the minimum and maximum charge states of the electrochemical energy storage power station respectively;

[0177] Scene constraints:

[0178] |P Gi,t,s -P Gi,t,bas |≤N Gi ;

[0179] |P EESSi,t,s -P EESSi,t,bas |≤N EESSi ;

[0180] Among them, P Gi,t,bas , P EESSi,t,bas are the basic output power of each conventional unit and electrochemical energy storage power station, N Gi 、N EESSi They are the power fluctuation range constraint values ​​of each conventional unit and electrochemical energy storage power station respectively;

[0181] Network transmission power constraints:

[0182] |B ij (θ i,t,s -θ j,t,s )|≤P ijmax ;

[0183] Among them, B ij is the susceptance corresponding to line i, j, θ i,t,s is the phase angle of node i at time t and in the sth scenario, P ijmax Transmit maximum power to the line;

[0184] Constraints for load demand response:

[0185] P PDRmin ≤P PDR ≤P PDRmax ;

[0186]

[0187]

[0188] Among them, P PDRmin , P PDRmax are the minimum and maximum values ​​of price-based demand response, is the load increase due to the incentive-based demand response in the sth scenario at time t, is the load reduction caused by the incentive-based demand response in the sth scenario at time t.

[0189] The day-ahead scheduling optimization device based on photovoltaic power prediction provided by the embodiment of the present invention can execute the day-ahead scheduling optimization method based on photovoltaic power prediction provided by any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0190] Embodiment 4

[0191] Figure 8A 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.

[0192] like Figure 8 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.

[0193] 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.

[0194] 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 special artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The processor 41 executes the various methods and processes described above, such as a day-ahead scheduling optimization method based on photovoltaic power prediction.

[0195] In some embodiments, the day-ahead scheduling optimization method based on photovoltaic power prediction can 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 can 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 day-ahead scheduling optimization method based on photovoltaic power prediction described above can be performed. Alternatively, in other embodiments, the processor 41 can be configured to execute the day-ahead scheduling optimization method based on photovoltaic power prediction in any other appropriate manner (for example, by means of firmware).

[0196] 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), complex 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.

[0197] 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.

[0198] 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.

[0199] 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).

[0200] 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.

[0201] 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.

[0202] 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.

[0203] 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 day-ahead scheduling optimization method based on photovoltaic power forecasting, characterized in that: include: Acquire historical meteorological data and historical photovoltaic power data before a target date, and acquire day-ahead load before the target date; Preprocessing and screening the historical meteorological data based on the historical photovoltaic power data to obtain target historical meteorological data; Inputting the target historical meteorological data into the photovoltaic power prediction model to obtain the photovoltaic prediction power on the target date; A day-ahead scheduling optimization model for the target date is constructed based on the photovoltaic predicted power and the day-ahead load to solve the optimal scheduling plan for the target date.

2. The method according to claim 1, characterized in that Preprocessing and screening the historical meteorological data based on the historical photovoltaic power data to obtain target historical meteorological data includes: Performing outlier correction, missing value filling, and normalization processing on various types of meteorological data and historical photovoltaic power data in the historical meteorological data to obtain various types of first meteorological data and pre-processed historical photovoltaic power data; Performing feature engineering processing on each type of first meteorological data to obtain second meteorological data after feature engineering; Performing wavelet transform processing on the pre-processed historical photovoltaic power data to obtain the wavelet transformed historical photovoltaic power data; Calculate the Pearson correlation coefficient between each type of second meteorological data and the historical photovoltaic power data processed by wavelet transform; Meteorological data with a Pearson correlation coefficient greater than a threshold value is selected from multiple types of second meteorological data to obtain target historical meteorological data.

3. The method according to claim 1, characterized in that The photovoltaic power prediction model includes a one-dimensional CNN network, an LSTM network, a feature fusion network and a power prediction network, the output layers of the one-dimensional CNN network and the LSTM network are connected to the input layer of the feature fusion network, and the output layer of the feature fusion network is connected to the input layer of the power prediction network; Inputting the target historical meteorological data into the photovoltaic power prediction model to obtain the photovoltaic power prediction on the target date includes: Inputting the target historical meteorological data into the one-dimensional CNN network and the LSTM network respectively; Extracting spatial features from the target historical meteorological data in the one-dimensional CNN network, and extracting time series features from the target historical meteorological data in the LSTM network; In the feature fusion network, the spatial features and the time series features are concatenated and fused to obtain fused meteorological features; The photovoltaic power forecast on the target date is forecasted in the power forecasting network based on the meteorological characteristics.

4. The method according to claim 1, characterized in that: The photovoltaic power prediction model is trained by the following steps: Acquire a training data set, wherein the training data set includes preprocessed meteorological sample data and annotated first photovoltaic power; Construct a photovoltaic power prediction model including a CNN network and an LSTM network, and initialize the network parameters of the photovoltaic power prediction model; Extracting meteorological sample data from the training data set and inputting it into a photovoltaic power prediction model to obtain a second photovoltaic power; calculating a loss rate based on the first photovoltaic power and the second photovoltaic power; Determine whether the training stop condition is met; If so, determining that the photovoltaic power prediction model has completed training; If not, the loss rate is used to adjust the model parameters of the photovoltaic power prediction model, and the process returns to the step of extracting meteorological sample data from the training data set and inputting the meteorological sample data into the photovoltaic power prediction model to obtain the second photovoltaic power.

5. The method according to any one of claims 1 to 4, characterized in that: Based on the photovoltaic predicted power and the day-ahead load, a day-ahead dispatch optimization model for the target date is constructed to solve the optimal dispatching plan for the target date, including: Generate multiple scenarios through a scenario generation algorithm to obtain scenario data for the multiple scenarios, where the scenario data at least includes the probability of the scenario occurring and data of conventional units, electrochemical energy storage power stations, and photovoltaic power station units in each scenario; Building a day-ahead dispatch optimization model based on the generated scenario data of the multiple scenarios, the photovoltaic predicted power and the day-ahead load; The day-ahead scheduling optimization model is solved to obtain the optimal scheduling plan for the target date.

6. The method according to claim 5, characterized in that The expression of the constructed day-ahead scheduling optimization model is as follows: Among them, C t is the total operating cost of the power grid, C G,t , C EESS,t , C PV,t , C L,t are the cost functions of conventional units, electrochemical energy storage power stations, photovoltaic power station units and loads, respectively, N s 、N G 、N EESS 、N PV are the number of scenarios, the number of conventional units, the number of electrochemical energy storage power stations, and the number of photovoltaic power station units, respectively. Gi,t,s , P EESSi,t,s , P PVi,t,s are the outputs of conventional units, electrochemical energy storage power stations, and photovoltaic power station units at time t and in the sth scenario, respectively, and u Gi,t,s 、u EESSi,t,s 、u PVi,t,s are the start and stop states of conventional units, electrochemical energy storage power stations, and photovoltaic power station units at time t and in the sth scenario, respectively. s is the probability coefficient of the sth scenario, a i 、b i 、c i is the power generation cost coefficient of conventional units, S i is the start-up and shutdown cost coefficient of conventional units, C(P EESSi,t,s )、M(P EESSi,t,s ) are the cost function and maintenance cost function of the electrochemical energy storage power station at time t and in the sth scenario, respectively. EESS is the unit time depreciation cost coefficient of the electrochemical energy storage power station, C(P PVi,t,s ) is the cost function of the photovoltaic power station unit at time t and in the sth scenario, k p,PV , k p,L are the penalty coefficient for abandoned light and the penalty cost coefficient for load power failure, P PVfor,t,s is the predicted value of the photovoltaic power station unit at time t and the sth scenario, k IDR is the coefficient of incentive demand response, Δ|P IDR,t,s | is the response amount of incentive demand in the sth scenario at time t, P L,t,s is the total load loss in the system at time t and in the sth scenario.

7. The method according to claim 5, characterized in that Solving the day-ahead scheduling optimization model to obtain the optimal scheduling solution for the target date includes: Find a solution that satisfies the following constraints so that C t The smallest scenario is taken as the target scenario, and the output of the conventional units, electrochemical energy storage power stations, and photovoltaic power station units of the target scenario is determined to obtain the optimal scheduling plan for the target date, wherein the constraints include: Power constraints: Among them, L fixed,t,s is the portion of the load that is not affected by electricity prices in the sth scenario at time t, ΔP PDR,t,s is the price-based demand response quantity in the sth scenario at time t, ΔP IDR,t,s is the amount of incentive demand response in the sth scenario at time t; Output constraints of conventional units: P Gi,min ≤P Gi,t,s ≤P Gi,max ; |P Gi,t+1,s -P Gi,t,s |≤u Gi,t+1,s R i ; Among them, P Gi,min , P Gi,max are the minimum and maximum output values ​​of each conventional unit, R i is the climbing power of each conventional unit; Output constraints of photovoltaic power station units: P PV,i,t,s ≤P PVfor,t,s ; Output constraints of electrochemical energy storage power station: P EESSfli,t,s ≤P EESSi,t,s ≤P EESSupi,t,s ; S SOCmin ≤S SOC,t,s ≤S SOCmax ; Among them, P EESSfli,t,s , P EESSupi,t,s are the lower and upper limits of the charge and discharge of the electrochemical energy storage power station at time t and in the sth scenario, respectively. SOC,t,s is the state of charge of the electrochemical energy storage power station at time t and the sth scenario, S SOCmin , S SOCmax They are the minimum and maximum charge states of the electrochemical energy storage power station respectively; Scene constraints: |P Gi,t,s -P Gi,t,bas |≤N Gi ; |P EESSi,t,s -P EESSi,t,bas |≤N EESSi ; Among them, P Gi,t,bas , P EESSi,t,bas are the basic output power of each conventional unit and electrochemical energy storage power station, N Gi 、N EESSi They are the power fluctuation range constraint values ​​of each conventional unit and electrochemical energy storage power station respectively; Network transmission power constraints: |B ij (i i,t,s -θ j,t,s )|≤P ijmax ; Among them, B ij is the susceptance corresponding to line i, j, θ i,t,s is the phase angle of node i at time t and in the sth scenario, P ijmax Transmit maximum power to the line; Constraints for load demand response: P PDRmin ≤P PDR ≤P PDRmax ; Among them, P PDRmin , P PDRmax are the minimum and maximum values ​​of price-based demand response, is the load increase due to the incentive-based demand response in the sth scenario at time t, is the load reduction caused by the incentive-based demand response in the sth scenario at time t.

8. A day-ahead dispatch optimization device based on photovoltaic power prediction, characterized in that: include: A historical data acquisition module, used to acquire historical meteorological data and historical photovoltaic power data before a target date, and to acquire day-ahead load before the target date; A data preprocessing and screening module, used for preprocessing and screening the historical meteorological data based on the historical photovoltaic power data to obtain target historical meteorological data; A photovoltaic power prediction module, used for inputting target historical meteorological data into a photovoltaic power prediction model to obtain photovoltaic prediction power on the target date; The scheduling optimization module is used to construct a scheduling optimization model for the target date based on the photovoltaic predicted power and the day-ahead load to solve the optimal scheduling plan for the target date.

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 day-ahead scheduling optimization method based on photovoltaic power prediction according to any one of claims 1 to 7.

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 day-ahead scheduling optimization method based on photovoltaic power prediction according to any one of claims 1 to 7 when executed.

Images