New energy base delivery quantitative analysis method and system based on price difference contract mechanism

Abstract

The invention provides a new energy base delivery quantitative analysis method and system based on a price difference contract mechanism. The method comprises the following steps: selecting a new energy base receiving end; calculating the outgoing electricity price of the new energy base; analyzing the external power transmission quantity composition of the new energy base; predicting the new energy marketization transaction electricity price; performing new energy base delivery quantitative analysis based on a price difference contract mechanism; and making a new energy base delivery strategy. According to the method, the influence of various factors such as energy supply and demand, adjustment and climate conditions on the new energy transaction price is fully considered by adopting a hybrid deep learning and reinforcement learning method, and the relationship between the mechanism electric quantity and the mechanism electricity price is accurately quantified by constructing the quantitative analysis model, so that the new energy transaction price is accurately analyzed. A scientific basis is provided for a new energy base operator to formulate a competitive delivery strategy and assess the income risk, the competitive power of the new energy base in the receiving end power market is enhanced, and the economical efficiency and sustainability of a project are ensured.

Description

Technical Field

[0001] The present invention relates to the field of electricity market, and in particular to a method and system for quantitative analysis of power transmission from a new energy base based on a contract for difference mechanism. Background Art

[0002] At present, many scholars have carried out research on large-scale new energy bases, but most researchers mainly focus on the capacity optimization configuration and coordinated operation strategies of different power types such as photovoltaic, wind power, and thermal power. They ignore the key points of market trading links in the context of full entry of new energy into the market, and it is difficult to effectively adapt to market dynamics.

[0003] While some scholars have explored price forecasting for new energy market transactions, they often employ time series forecasting models. These models primarily rely on historical price data to infer trends, failing to fully account for the numerous complex and intertwined factors influencing new energy market transactions, such as energy supply and demand and climate conditions. This results in significant deviations between forecasts and actual prices, making it difficult to meet the demands of accurate decision-making.

[0004] Furthermore, there is a lack of quantitative analysis of energy transmission from new energy bases based on the CFD settlement mechanism. This means that new energy base operators lack a scientific and quantitative basis when formulating transmission strategies and assessing returns and risks, making it difficult to achieve optimal resource allocation and maximize benefits in a complex market environment.

[0005] Therefore, existing research still has much room for improvement and innovation in improving the prediction of new energy market transaction prices and quantitative analysis of new energy base delivery under the CFD mechanism.

[0006] At present, how to accurately assess the uncertainties brought about by the market-oriented transaction prices, mechanism electricity quantities and mechanism electricity prices of new energy has become a key issue that needs to be urgently solved in the development of new energy bases. Summary of the Invention

[0007] In order to overcome the above-mentioned deficiencies of the prior art, the present invention discloses a method and system for quantitative analysis of new energy base transmission based on a contract for difference mechanism.

[0008] The technical solution provided by the present invention is: a quantitative analysis method for new energy base transmission based on the difference contract mechanism, comprising the following steps:

[0009] New energy base receiving terminal selection: Select the new energy base power receiving terminal target, obtain the receiving terminal's power and lighting requirements, and the power receiving terminal coal benchmark price;

[0010] Calculation of electricity price for new energy bases: Based on the benchmark price of coal at the receiving end and the channel costs between the transmission and reception, the upper limit of the comprehensive on-grid electricity price of the new energy base is calculated;

[0011] Analysis of the composition of electricity output from new energy bases: Based on the power and electricity demand of the receiving end, calculate the electricity output of different types of power sources such as wind, solar, thermal, and nuclear at the new energy bases;

[0012] Predicting the electricity price of renewable energy market transactions: Build a new energy market transaction price prediction model based on hybrid deep learning and reinforcement learning to predict the on-grid electricity price of wind and solar renewable energy in new energy bases;

[0013] Quantitative analysis of new energy base transmission based on the CFD mechanism: Build a quantitative analysis model for new energy base transmission based on the CFD mechanism, and quantitatively analyze the relationship between mechanism electricity quantity and mechanism electricity price under the full participation of new energy in market-based electricity trading prices;

[0014] Formulate a transmission strategy for new energy bases: Based on the quantitatively analyzed mechanism electricity volume and mechanism electricity price, formulate a transmission strategy that is suitable for the new energy base itself and competitive to the receiving end.

[0015] Preferably, the identification of unit types specifically includes: screening units that meet the compensation conditions based on historical cost recovery data, classifying and formulating subsidy standards based on the differences in cost-effectiveness of unit scale and type, classifying according to capacity into 1000MW, 600MW, and 300MW levels, and distinguishing between coal-fired and gas-fired units in terms of type.

[0016] Further preferably, the power receiving object is R, and the power demand of the receiving object is represented by vector P R Indicates that the power demand in a certain period is expressed as Q R Indicates that the coal base price at the power receiving end is C R .

[0017] More preferably, the upper limit of the comprehensive grid-connected electricity price of the energy base can be expressed by the following formula:

[0018] C up =C R -C C (1)

[0019] Among them, C up is the upper limit of the comprehensive on-grid electricity price of the new energy base; C R C is the coal benchmark price at the power receiving end; C It is the cost of the power grid channel between the sending and receiving ends.

[0020] Further preferably, the relationship between the power amounts of different power types is shown as follows:

[0021] Q R =Q w +Q S +Q T +Q N (2)

[0022] Among them, Q R Q is the power demand of the receiving power grid in a certain period of time, W is the wind power generation of the new energy base in a certain period of time, Q S is the photovoltaic power generation of the new energy base in a certain period of time, Q T is the thermal power generation of the new energy base in a certain period of time, Q N It refers to the amount of nuclear power generated by the new energy base during a certain period of time.

[0023] Further preferably, the construction of the new energy market-based transaction electricity price prediction model includes the following steps:

[0024] S1: Collect historical electricity price data, meteorological data, and market supply and demand data for new energy market transactions, clean and standardize the data, remove outliers and missing values, and standardize the data using the Z-score standardization method:

[0025]

[0026] Where X is the original data, μ is the mean, σ is the standard deviation, and the data is divided into training set and test set with a ratio of 80% and 20%;

[0027] S2: Extract key features, including seasonality, trend, and correlation between meteorological data and electricity prices;

[0028] Among them, seasonal features: one-hot encoding is used to convert season and week into binary feature vectors;

[0029] Trend characteristics: Use a rolling window to calculate the moving average and moving standard deviation of historical electricity prices. The rolling average over the past 7 days is:

[0030]

[0031] Among them, P t-i is the electricity price on day ti;

[0032] Meteorological characteristics: Process meteorological data, including wind speed, light intensity, and temperature, use linear interpolation to fill missing values, and perform standardization;

[0033] Construct feature vector X t , contains all the above features;

[0034] S3: Build a deep neural network (DNN) model to capture the nonlinear characteristics and complex patterns in electricity price data. The structure of the DNN model is as follows:

[0035] Input layer: feature vector Xt;

[0036] Hidden layer: ReLU activation function is used, and the number of hidden layer neurons is 128 and 64 respectively;

[0037] Output layer: predicted electricity price

[0038] The stochastic gradient descent (SGD) algorithm is used to train the DNN model, and the loss function uses the mean square error (MSE):

[0039]

[0040] Among them, P i is the real electricity price, To predict electricity prices, n is the number of samples;

[0041] Use Bayesian optimization to adjust the hyperparameters of the DNN model, including the learning rate η and the batch size B. The goal of Bayesian optimization is to minimize the MSE on the validation set.

[0042] S4: Construct a Lasso estimated autoregressive (LEAR) model to handle the autocorrelation and linear trend of electricity price data. The formula of the LEAR model is:

[0043]

[0044] in:

[0045] is the electricity price predicted by the LEAR model; β0 is the intercept term; β i is the autoregressive coefficient; γ j is the moving average coefficient;∈ t is a white noise sequence; p is the autoregressive order; q is the moving average order;

[0046] The model parameters β0, β i , γ j , and use the Akaike Information Criterion (AIC) to determine the optimal p and q;

[0047] S5: Perform weighted fusion on the outputs of the DNN model and the LEAR model to obtain a comprehensive prediction result:

[0048]

[0049] Among them, the weight α is obtained through historical data training, and the calculation method of the weight α is:

[0050]

[0051] Use the gradient descent algorithm to optimize α to minimize the MSE on the validation set;

[0052] S6: Introduce the reinforcement learning module, use the Q-learning algorithm to dynamically adjust the weight α, and define the state St as:

[0053] S t =[P t ,ε t ,ΔP t ] (9)

[0054] Among them, ε t is the prediction error:

[0055]

[0056] ΔP t The rate of change of electricity price:

[0057] ΔP t =P t -P t-1 (11)

[0058] The action At is the change Δα of the adjustment weight α;

[0059] The reward Rt is the negative of the prediction error, that is:

[0060] R t =-|∈ t | (12)

[0061] The goal of Q-learning is to maximize the cumulative reward, and the update formula of Q value is:

[0062] Q(S t ,A t )=Q(S t ,A t )+α(R t +γmax a Q(S t+1 ,α)-Q(S t ,A t )) (13)

[0063] Among them, α is the learning rate and γ is the discount factor;

[0064] By simulating market operations, dynamically adjusting weight α, and optimizing forecasting strategies;

[0065] S7: Use mean square error (MSE), mean absolute error (MAE), and mean absolute percentage error (MAPE) to evaluate model performance. The calculation formula is as follows:

[0066]

[0067] Further preferably, the quantitative analysis model for new energy base transmission based on the CFD mechanism is as follows:

[0068] The income from wind and solar power generation is as follows:

[0069]

[0070] Among them, I n , I ni , I no They are renewable energy power generation income, on-site income, and off-site income;

[0071] Q m，t 、 They are the amount of renewable energy transmitted and the market transaction price within a certain period of time;

[0072] Q j C j They are the mechanism electricity quantity and mechanism electricity price included in the renewable energy transmission electricity quantity;

[0073] C ma The average market transaction price is determined by the weighted average price of similar projects in the monthly real-time market on the power generation side in areas where the power spot market operates continuously. The average market transaction price is determined by the weighted average price of similar projects in the medium- and long-term transactions on the power generation side during the active trading period in areas where the power spot market does not operate continuously.

[0074] The quantitative analysis model for new energy base transmission based on the CFD mechanism is as follows:

[0075]

[0076] Among them, I t , I N They are the electricity transmission income of flexible power sources thermal power and nuclear power in the new energy base, C T 、C N The on-grid electricity price of thermal power and nuclear power in the location of the new energy base;

[0077] By combining formula (16) and formula (17), the relationship between the mechanism electricity quantity included in the renewable energy transmission electricity quantity and the mechanism electricity price is quantified.

[0078] According to another aspect of the present invention, a new energy base delivery quantitative analysis system based on a contract for difference mechanism is provided, comprising:

[0079] The new energy base receiving terminal selection unit is used to obtain the power and lighting requirements of the receiving terminal and the coal benchmark price of the power receiving terminal;

[0080] The new energy base outbound electricity price calculation unit is used to calculate the upper limit of the comprehensive on-grid electricity price of the new energy base;

[0081] The new energy base's outbound power supply composition analysis unit is used to calculate the power of different types of power sources such as wind, solar, thermal, and nuclear at the new energy base;

[0082] Prediction of new energy market-based transaction electricity price unit, used to predict wind and solar power grid prices in new energy bases;

[0083] A quantitative analysis unit for outbound transmission from new energy bases based on the CFD mechanism is used to quantitatively analyze the relationship between mechanism electricity quantity and mechanism electricity price under the electricity price of full participation of new energy in market-based transactions;

[0084] The new energy base delivery strategy formulation unit is used to formulate a delivery strategy that is suitable for the new energy base itself and competitive to the receiving end.

[0085] The beneficial effects of the present invention are:

[0086] This invention proposes a quantitative analysis method and system for new energy base outbound delivery based on the CFD mechanism, which significantly improves the accuracy of new energy market transaction price prediction;

[0087] Among them, the new energy market transaction electricity price forecasting method based on hybrid deep learning and reinforcement learning: innovatively integrates deep neural networks (DNN) and Lasso estimated autoregressive (LEAR) models. Through steps such as data collection and preprocessing, feature engineering, model training and optimization, and reinforcement learning modules, it accurately predicts the market transaction electricity price of new energy, overcoming the limitations of traditional time series forecasting models and significantly improving forecast accuracy.

[0088] A quantitative analysis model for new energy base transmission based on the CFD mechanism: Under the CFD settlement mechanism, a quantitative analysis model is constructed to accurately quantify the relationship between the mechanism-based electricity volume and the mechanism-based electricity price in the transmission volume of new energy bases. This fills the methodological gap in this field and provides a scientific basis for new energy base operators to formulate transmission strategies and assess revenue risks.

[0089] In summary, the present invention fully considers the impact of multiple factors such as energy supply and demand, climate conditions, etc. on the transaction price of new energy by adopting a hybrid deep learning and reinforcement learning method, overcoming the limitations of traditional time series prediction models. In addition, this method also fills the gap in the quantitative analysis method of new energy base transmission under the difference contract settlement mechanism. By constructing a quantitative analysis model, it accurately quantifies the relationship between mechanism electricity volume and mechanism electricity price, providing a scientific basis for the operators of new energy bases to formulate competitive transmission strategies and evaluate profit risks, which helps to enhance the competitiveness of new energy bases in the receiving power market and ensure the economy and sustainability of the project. BRIEF DESCRIPTION OF THE DRAWINGS

[0090] Figure 1 This is a flow chart of the quantitative analysis method for new energy base transmission based on the difference contract mechanism provided by the present invention. DETAILED DESCRIPTION

[0091] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0092] The present invention provides a quantitative analysis method for the outbound delivery of new energy bases based on the CFD mechanism. Figure 1 Shown, including:

[0093] New energy base receiving terminal selection: Select the new energy base power receiving terminal target, obtain the receiving terminal's power and lighting requirements, and the power receiving terminal coal benchmark price;

[0094] When building a complete and accurate quantitative analysis model for new energy base transmission based on the CFD mechanism, the first and core task is to accurately select the power source receiving target. The power source receiving object is set as R, and the power demand of the receiving object is represented by the vector P. R Indicates that the power demand in a certain period is expressed as Q R Indicates that the power demand P R and power demand Q R Together, they characterize the scale and characteristics of the electric energy demand of the receiving object R;

[0095] At the same time, the coal benchmark price at the receiving end of the power source is also an important factor in the decision-making and quantitative analysis of the new energy base's transmission. When the bundled electricity price of each power source type of the new energy base is higher than the coal benchmark price at the receiving end, the market competitiveness is at a disadvantage. The coal benchmark price at the receiving end of the power source is C R ;

[0096] Calculation of electricity price for new energy bases: Based on the benchmark price of coal at the receiving end and the channel costs between the transmission and reception, the upper limit of the comprehensive on-grid electricity price of the new energy base is calculated;

[0097] When operators of large-scale renewable energy bases evaluate the feasibility of transmitting electricity from their power source to the receiving end, the price of electricity transmitted from the renewable energy base is a key indicator. To ensure that the price of electricity transmitted from the renewable energy base is competitive, they need to comprehensively consider the coal base price at the receiving end and the channel costs between the transmitting and receiving ends. The upper limit of the comprehensive on-grid electricity price of the renewable energy base can be expressed as follows:

[0098] C up =C R -C C (1)

[0099] Among them, C up is the upper limit of the comprehensive on-grid electricity price of the new energy base; C R C is the coal benchmark price at the power receiving end; C The cost of the power grid channel between the transmitter and receiver. Whether building a new transmission line or utilizing an existing transmission channel, the power industry has a mature and scientific transmission cost accounting system to determine the specific value.

[0100] Analysis of the composition of electricity output from new energy bases: Based on the power and electricity demand of the receiving end, calculate the electricity output of different types of power sources such as wind, solar, thermal, and nuclear at the new energy bases;

[0101] Taking into account the power and electricity demand of the power receiving end, as well as the installed capacity and output characteristics of different power sources such as wind, solar, thermal, and nuclear at the new energy base, the power composition of different power types within a certain period is estimated. The relationship between the power of different power types is shown in the following formula:

[0102] Q R =Q W +Q S +Q T +Q N (2)

[0103] Among them, Q R Q is the power demand of the receiving power grid in a certain period of time, W is the wind power generation of the new energy base in a certain period of time, Q S is the photovoltaic power generation of the new energy base in a certain period of time, Q T is the thermal power generation of the new energy base in a certain period of time, Q N The amount of nuclear power generated by the new energy base during a certain period of time;

[0104] Predicting the electricity price of renewable energy market transactions: Build a new energy market transaction price prediction model based on hybrid deep learning and reinforcement learning to predict the on-grid electricity price of wind and solar renewable energy in new energy bases;

[0105] The construction of the new energy market-based transaction electricity price prediction model includes the following steps:

[0106] S1: Collect historical electricity price data, meteorological data, and market supply and demand data for new energy market transactions, clean and standardize the data, remove outliers and missing values, and standardize the data using the Z-score standardization method:

[0107]

[0108] Where X is the original data, μ is the mean, σ is the standard deviation, and the data is divided into training set and test set with a ratio of 80% and 20%;

[0109] S2: Extract key features, including seasonality, trend, and correlation between meteorological data and electricity prices;

[0110] Among them, seasonal features: one-hot encoding is used to convert season and week into binary feature vectors;

[0111] Trend characteristics: Use a rolling window to calculate the moving average and moving standard deviation of historical electricity prices. The rolling average over the past 7 days is:

[0112]

[0113] Among them, P t-i is the electricity price on day ti;

[0114] Meteorological characteristics: Process meteorological data, including wind speed, light intensity, and temperature, use linear interpolation to fill missing values, and perform standardization;

[0115] Construct feature vector X t , contains all the above features;

[0116] S3: Build a deep neural network (DNN) model to capture the nonlinear characteristics and complex patterns in electricity price data. The structure of the DNN model is as follows:

[0117] Input layer: feature vector Xt;

[0118] Hidden layer: ReLU activation function is used, and the number of hidden layer neurons is 128 and 64 respectively;

[0119] Output layer: predicted electricity price

[0120] The stochastic gradient descent (SGD) algorithm is used to train the DNN model, and the loss function uses the mean square error (MSE):

[0121]

[0122] Among them, P i is the real electricity price, To predict electricity prices, n is the number of samples;

[0123] Use Bayesian optimization to adjust the hyperparameters of the DNN model, including the learning rate η and the batch size B. The goal of Bayesian optimization is to minimize the MSE on the validation set.

[0124] S4: Construct a Lasso estimated autoregressive (LEAR) model to handle the autocorrelation and linear trend of electricity price data. The formula of the LEAR model is:

[0125]

[0126] in:

[0127] is the electricity price predicted by the LEAR model; β0 is the intercept term; β i is the autoregressive coefficient; γ j is the moving average coefficient; ∈ t is a white noise sequence; p is the autoregressive order; q is the moving average order;

[0128] The model parameters β0, β i , γ j , and use the Akaike Information Criterion (AIC) to determine the optimal p and q;

[0129] S5: Perform weighted fusion on the outputs of the DNN model and the LEAR model to obtain a comprehensive prediction result:

[0130]

[0131] Among them, the weight α is obtained through historical data training, and the calculation method of the weight α is:

[0132]

[0133] Use the gradient descent algorithm to optimize α to minimize the MSE on the validation set;

[0134] S6: Introduce the reinforcement learning module, use the Q-learning algorithm to dynamically adjust the weight α, and define the state St as:

[0135] S t =[P t ,∈ t , ΔP t ] (9)

[0136] Among them, ∈ t is the prediction error:

[0137]

[0138] ΔP t The rate of change of electricity price:

[0139] ΔP t =P t -P t-1 (11)

[0140] The action At is the change Δα of the adjustment weight α;

[0141] The reward Rt is the negative of the prediction error, that is:

[0142] R t =|∈ t | (12)

[0143] The goal of Q-learning is to maximize the cumulative reward, and the update formula of Q value is:

[0144] Q(S t ,A t )=Q(S t ,A t )+α(R t +γmax a Q(S t+1 ,α)-Q(S t ,A t )) (13)

[0145] Among them, α is the learning rate and γ is the discount factor;

[0146] By simulating market operations, dynamically adjusting weight α, and optimizing forecasting strategies;

[0147] S7: Use mean square error (MSE), mean absolute error (MAE), and mean absolute percentage error (MAPE) to evaluate model performance. The calculation formula is as follows:

[0148]

[0149] Quantitative analysis of new energy base transmission based on the CFD mechanism: Build a quantitative analysis model for new energy base transmission based on the CFD mechanism, and quantitatively analyze the relationship between mechanism electricity quantity and mechanism electricity price under the full participation of new energy in market-based electricity trading prices;

[0150] Within the framework of the CFD mechanism, there is currently a lack of theoretical methods for large-scale renewable energy base operators to conduct quantitative analysis of mechanism electricity and mechanism electricity prices. In view of this, this paper specifically proposes a quantitative analysis method for renewable energy base outbound transmission based on the CFD mechanism.

[0151] The quantitative analysis model for new energy base transmission based on the CFD mechanism is as follows:

[0152] The income from wind and solar power generation is as follows:

[0153]

[0154] Among them, I n , I ni , I no They are renewable energy power generation income, on-site income, and off-site income;

[0155] Q m，t 、 They are the amount of renewable energy transmitted and the market transaction price within a certain period of time;

[0156] Q j C j They are the mechanism electricity quantity and mechanism electricity price included in the renewable energy transmission electricity quantity;

[0157] C ma The average market transaction price is determined by the weighted average price of similar projects in the monthly real-time market on the power generation side in areas where the power spot market operates continuously. The average market transaction price is determined by the weighted average price of similar projects in the medium- and long-term transactions on the power generation side during the active trading period in areas where the power spot market does not operate continuously.

[0158]

[0159] Among them, I t , I N They are the electricity export income of flexible power sources thermal power and nuclear power in the new energy base, C T 、C N The on-grid electricity price of thermal power and nuclear power in the location of the new energy base;

[0160] By combining formula (16) and formula (17), the relationship between the mechanism electricity quantity included in the renewable energy transmission electricity quantity and the mechanism electricity price is quantified;

[0161] Formulate a transmission strategy for new energy bases: Based on the quantitatively analyzed mechanism electricity volume and mechanism electricity price, formulate a transmission strategy that is suitable for the new energy base itself and competitive to the receiving end.

[0162] Another aspect of the present invention provides a new energy base delivery quantitative analysis system based on a contract for difference mechanism, comprising:

[0163] The new energy base receiving terminal selection unit is used to obtain the power and lighting requirements of the receiving terminal and the coal benchmark price of the power receiving terminal;

[0164] The new energy base outbound electricity price calculation unit is used to calculate the upper limit of the comprehensive on-grid electricity price of the new energy base;

[0165] The new energy base's outbound power transmission composition analysis unit is used to calculate the power of different types of power sources such as wind, solar, thermal, and nuclear at the new energy base;

[0166] Prediction of new energy market-based transaction electricity price unit, used to predict wind and solar power grid prices in new energy bases;

[0167] A quantitative analysis unit for outbound transmission from new energy bases based on the CFD mechanism is used to quantitatively analyze the relationship between mechanism electricity quantity and mechanism electricity price under the electricity price of full participation of new energy in market-based transactions;

[0168] The new energy base delivery strategy formulation unit is used to formulate a delivery strategy that is suitable for the new energy base itself and competitive to the receiving end.

[0169] Other embodiments of the present invention will readily occur to those skilled in the art after considering the specification and practicing the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of the present invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein. The description and examples are to be considered as exemplary only, with the true scope and spirit of the invention being indicated by the claims.

[0170] It should be understood that the present invention is not limited to the precise construction shown in the above description and that various modifications and changes can be made without departing from the scope thereof. The scope of the present invention is limited only by the appended claims.

Claims

1. A quantitative analysis method for new energy base outbound delivery based on the CFD mechanism, characterized by: The steps include: New energy base receiving terminal selection: Select the new energy base power receiving terminal target, obtain the receiving terminal's power and lighting requirements, and the power receiving terminal coal benchmark price; Calculation of electricity price for new energy bases: Based on the benchmark price of coal at the receiving end and the channel costs between the transmission and reception, the upper limit of the comprehensive on-grid electricity price of the new energy base is calculated; Analysis of the composition of electricity output from new energy bases: Based on the power and electricity demand of the receiving end, calculate the electricity output of different types of power sources such as wind, solar, thermal, and nuclear at the new energy bases; Predicting the electricity price of renewable energy market transactions: Build a new energy market transaction price prediction model based on hybrid deep learning and reinforcement learning to predict the on-grid electricity price of wind and solar renewable energy in new energy bases; Quantitative analysis of new energy base transmission based on the CFD mechanism: Build a quantitative analysis model for new energy base transmission based on the CFD mechanism, and quantitatively analyze the relationship between mechanism electricity quantity and mechanism electricity price under the full participation of new energy in market-based electricity trading prices; Formulate a transmission strategy for new energy bases: Based on the quantitatively analyzed mechanism electricity volume and mechanism electricity price, formulate a transmission strategy that is suitable for the new energy base itself and competitive to the receiving end.

2. The quantitative analysis method for new energy base outbound delivery based on the CFD mechanism according to claim 1 is characterized in that: The power receiving object is R, and the power demand of the receiving object is represented by the vector P R Indicates that the power demand in a certain period is expressed as Q R Indicates that the coal base price at the power receiving end is C R .

3. The quantitative analysis method for new energy base outbound delivery based on the CFD mechanism according to claim 1 is characterized in that: The upper limit of the comprehensive grid-connected electricity price of the new energy base can be expressed as follows: C up =C R -C C (1) Among them, C up is the upper limit of the comprehensive on-grid electricity price of the new energy base; C R C is the coal benchmark price at the power receiving end; C It is the cost of the power grid channel between the sending and receiving ends.

4. The quantitative analysis method for new energy base outbound delivery based on the CFD mechanism according to claim 1 is characterized in that: The relationship between the power of different power types is shown in the following formula: Q R =Q W +Q S +Q T +Q N (2) Among them, Q R Q is the power demand of the receiving power grid in a certain period of time, W is the wind power generation of the new energy base in a certain period of time, Q S is the photovoltaic power generation of the new energy base in a certain period of time, Q T is the thermal power generation of the new energy base in a certain period of time, Q N It refers to the amount of nuclear power generated by the new energy base during a certain period of time.

5. The quantitative analysis method for new energy base outbound delivery based on the CFD mechanism according to claim 1 is characterized in that: The construction of the new energy market-based transaction electricity price prediction model includes the following steps: S1: Collect historical electricity price data, meteorological data, and market supply and demand data of new energy market transactions, clean and standardize the data, remove outliers and missing values, and standardize the data using the Z-score standardization method: Where X is the original data, μ is the mean, σ is the standard deviation, and the data is divided into training set and test set with a ratio of 80% and 20%; S2: Extract key features, including seasonality, trend, and correlation between meteorological data and electricity prices; Among them, seasonal features: use one-hot encoding to convert season and week into binary feature vectors; Trend characteristics: Use a rolling window to calculate the moving average and moving standard deviation of historical electricity prices. The rolling average over the past 7 days is: Among them, P t-i is the electricity price on day ti; Meteorological characteristics: Process meteorological data, including wind speed, light intensity, and temperature, use linear interpolation to fill missing values, and perform standardization; Construct feature vector X t , contains all the above features; S3: Build a deep neural network (DNN) model to capture the nonlinear characteristics and complex patterns in electricity price data. The structure of the DNN model is as follows: Input layer: feature vector Xt; Hidden layer: ReLU activation function is used, and the number of hidden layer neurons is 128 and 64 respectively; Output layer: predicted electricity price The stochastic gradient descent (SGD) algorithm is used to train the DNN model, and the loss function uses the mean square error (MSE): Among them, P i is the real electricity price, To predict electricity prices, n is the number of samples; Use Bayesian optimization to adjust the hyperparameters of the DNN model, including the learning rate η and the batch size B. The goal of Bayesian optimization is to minimize the MSE on the validation set. S4: Construct a Lasso estimated autoregressive (LEAR) model to handle the autocorrelation and linear trend of electricity price data. The formula of the LEAR model is: in: is the electricity price predicted by the LEAR model; β0 is the intercept term; β i is the autoregressive coefficient; γ j is the moving average coefficient; ∈ t is a white noise sequence; p is the autoregressive order; q is the moving average order; The model parameters β0, β i , γ j , and use the Akaike Information Criterion (AIC) to determine the optimal p and q; S5: Perform weighted fusion on the outputs of the DNN model and the LEAR model to obtain a comprehensive prediction result: Among them, the weight α is obtained through historical data training, and the calculation method of the weight α is: Use the gradient descent algorithm to optimize α to minimize the MSE on the validation set; S6: Introduce the reinforcement learning module, use the Q-learning algorithm to dynamically adjust the weight α, and define the state St as: S t =[P t ,∈ t ,ΔP t ] (9) Among them, ∈ t is the prediction error: ΔP t The rate of change of electricity price: ΔP t =P t -P t-1 (11) The action At is the change Δα of the adjustment weight α; The reward Rt is the negative of the prediction error, that is: R t ＝-|∈ t | (12) The goal of Q-learning is to maximize the cumulative reward, and the update formula of Q value is: Q(S t ,A t )=Q(S t ,A t )+α(R t +γmax a Q(S t+1 ,a)-Q(S t ,A t )) (13) Among them, α is the learning rate and γ is the discount factor; By simulating market operations, dynamically adjusting weight α, and optimizing forecasting strategies; S7: Use mean square error (MSE), mean absolute error (MAE), and mean absolute percentage error (MAPE) to evaluate model performance. The calculation formula is as follows:

6. The quantitative analysis method for new energy base outbound delivery based on the CFD mechanism according to claim 1 is characterized in that: The quantitative analysis model for new energy base transmission based on the CFD mechanism is as follows: The income from wind and solar power generation is as follows: Among them, I n , I ni , I no They are renewable energy power generation income, on-site income, and off-site income; Q m，t 、 They are the amount of renewable energy transmitted and the market transaction price within a certain period of time; Q j C j They are the mechanism electricity quantity and mechanism electricity price included in the renewable energy transmission electricity quantity; C ma The average market transaction price is determined by the weighted average price of similar projects in the monthly real-time market on the power generation side in areas where the power spot market operates continuously. The average market transaction price is determined by the weighted average price of similar projects in the medium- and long-term transactions on the power generation side during the active trading period in areas where the power spot market does not operate continuously. The quantitative analysis model for new energy base transmission based on the CFD mechanism is as follows: Among them, I t , I N They are the electricity transmission income of flexible power sources thermal power and nuclear power in the new energy base, C T 、C N The on-grid electricity price of thermal power and nuclear power in the location of the new energy base; By combining formula (16) and formula (17), the relationship between the mechanism electricity quantity included in the renewable energy transmission electricity quantity and the mechanism electricity price is quantified.

7. The new energy base delivery quantitative analysis system based on the CFD mechanism is characterized by: include: The new energy base receiving terminal selection unit is used to obtain the power and lighting requirements of the receiving terminal and the coal benchmark price of the power receiving terminal; The new energy base outbound electricity price calculation unit is used to calculate the upper limit of the comprehensive on-grid electricity price of the new energy base; The new energy base's outbound power supply composition analysis unit is used to calculate the power of different types of power sources such as wind, solar, thermal, and nuclear at the new energy base; Prediction of new energy market-based transaction electricity price unit, used to predict wind and solar power grid prices in new energy bases; A quantitative analysis unit for outbound transmission from new energy bases based on the CFD mechanism is used to quantitatively analyze the relationship between mechanism electricity quantity and mechanism electricity price under the electricity price of full participation of new energy in market-based transactions; The new energy base delivery strategy formulation unit is used to formulate a delivery strategy that is suitable for the new energy base itself and competitive to the receiving end.

Images