Comprehensive energy system optimization scheduling method and system based on carbon emission factors and wind power prediction

Abstract

The invention relates to an integrated energy system optimal scheduling method and system based on a carbon emission factor and wind power prediction, and relates to the technical field of energy system scheduling, and the method comprises the steps: building a carbon emission metering model based on a dynamic carbon emission factor through metering the direct carbon emission and indirect carbon emission of an integrated energy system; meteorological features are extracted based on the improved wind power hybrid prediction model, and a wind power prediction interval is generated in combination with kernel density estimation; taking the sum of the energy purchase cost, the carbon transaction cost and the wind curtailment cost to be minimized as a target function, and constructing an economic dispatching model; and on the basis of the economic dispatching model and the target function, solving an optimal dispatching scheme through the dynamic carbon emission metering interval and the wind power prediction interval under the condition that a set constraint condition set is met. According to the method, the technical problems of large carbon emission accounting error and low wind power prediction precision in the optimal scheduling process of the existing integrated energy system are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of energy system scheduling, and in particular to a method and system for optimizing and scheduling an integrated energy system based on carbon emission factors and wind power forecasting. Background Art

[0002] As a key renewable energy source, wind power plays a crucial role in reducing reliance on fossil fuels and lowering carbon emissions. However, its power generation is significantly affected by changing natural conditions, exhibiting significant volatility and intermittency, posing challenges to the safe and stable operation of integrated energy systems. Selecting appropriate carbon emission measurement methods to calculate carbon emissions and accurately forecasting wind power output can accurately assess system carbon emissions. By developing appropriate scheduling strategies, such as power generation plans, we can achieve carbon reduction while ensuring safe and stable system operation.

[0003] Traditional carbon emission measurement methods mostly use static carbon emission factors to calculate carbon emissions. These methods cannot accurately measure the actual carbon emissions in the system and cannot effectively achieve carbon emission reduction in the system. Furthermore, traditional carbon emission measurement methods use point predictions, which cannot fully capture the dynamic characteristics of wind power and accurately reflect its randomness, making it difficult to effectively meet the system's scheduling decision-making needs. Summary of the Invention

[0004] (1) Technical issues to be resolved In view of the above-mentioned shortcomings and deficiencies of the prior art, the present invention provides a method and system for optimizing and scheduling an integrated energy system based on carbon emission factors and wind power forecasts, which solves the technical problems of large carbon emission calculation errors and low wind power forecast accuracy in the optimization and scheduling process of the existing integrated energy system.

[0005] (2) Technical solution In order to achieve the above objectives, the main technical solutions adopted by the present invention include: In a first aspect, an embodiment of the present invention provides a method for optimizing and scheduling an integrated energy system based on a carbon emission factor and wind power forecast, comprising: By measuring the direct carbon emissions of energy supply equipment and the indirect carbon emissions from electricity purchases from external power grids, a carbon emission measurement model based on dynamic carbon emission factors is established; Historical wind power data is cleaned and normalized using the quartile method. A convolutional neural network is used to extract meteorological features. A bidirectional gated recurrent unit is used to capture temporal dynamic features. An attention mechanism is introduced to enhance information. Kernel density estimation is then used to generate wind power forecast intervals that include the forecast error distribution. An economic dispatch model with interval variables is constructed, with the objective function minimizing the sum of energy purchase costs, carbon trading costs, and wind curtailment costs. The carbon trading costs are calculated based on the difference between the dynamic carbon emission measurement interval and the quota, while the wind curtailment costs are penalized based on the deviation between the wind power forecast interval and the actual output. Based on the economic dispatch model and objective function, the optimal dispatch scheme is solved through the dynamic carbon emission metering interval and wind power prediction interval, while satisfying the constraint condition set including the operating boundary of energy supply equipment, the charging and discharging range of the energy storage system, and the multi-energy balance of electricity, heat and gas.

[0006] Optionally, by measuring the direct carbon emissions of energy supply equipment and the indirect carbon emissions of electricity purchased from the external power grid, a carbon emission measurement model based on a dynamic carbon emission factor is established, including: Based on the carbon emission coefficient of natural gas, the carbon emissions are quantified by combining the energy output of gas boilers and cogeneration units in the energy supply equipment to obtain the direct carbon emissions of gas boilers and cogeneration units; Based on the wind power carbon emission coefficient of the wind turbine in the energy supply equipment, combined with the wind power generation, carbon emissions are quantified to obtain the wind power carbon emissions of the wind turbine; By measuring the total carbon emissions of direct carbon emissions, wind power carbon emissions, and indirect carbon emissions from external grid power purchases, and then comparing the total carbon emissions with the total energy output of the integrated energy system, the static carbon emission factor of the integrated energy system is obtained. According to the ratio of the total real-time carbon emissions of the integrated energy system to the non-new energy power generation, the static carbon emission factor is dynamically corrected in real time to obtain a dynamic carbon emission factor with adaptive carbon response; The real-time carbon emissions of gas boilers and electricity purchased from external power grids are calculated based on dynamic carbon emission factors, and the calculation results are measured against the real-time carbon emissions of the cogeneration unit to obtain a carbon emission measurement model. Among them, the expression of the carbon emission measurement model is: ; Where, E s represents the actual carbon emissions generated by the system, E CHP represents the actual carbon emissions of the cogeneration unit, E r,b Represents the actual carbon emissions from gas boilers and electricity purchased from external grids, T represents the scheduling period, E p represents the power generation of the cogeneration unit, K p Represents the unit power generation loss of natural gas during the power generation process of the cogeneration unit, lRepresents the carbon emission coefficient of natural gas ，P e,b ( t ) represents the amount of electricity purchased from the external power grid, P GB,h ( t ) represents a gas boiler t Output power during the time period, G lm Represents the dynamic carbon emission factor.

[0007] Optionally, after establishing a carbon emission measurement model based on a dynamic carbon emission factor by measuring the direct carbon emissions of energy supply equipment and the indirect carbon emissions of electricity purchased from the external power grid, the following is further included: The energy output range of the energy supply equipment and the electricity purchase range of the external power grid are input into the carbon emission measurement model to measure carbon emissions and obtain the carbon emission range of the integrated energy system; By measuring the carbon emission quotas of energy supply equipment and electricity purchased from external power grids, a carbon emission quota model for the integrated energy system is constructed; Based on the carbon emission quota range of the integrated energy system output by the carbon emission quota model, the carbon emission quota difference is calculated for the carbon emission range in combination with the carbon trading base price to obtain the carbon trading cost range of the integrated energy system; Among them, the expression of carbon trading cost range is: ; Where, represents the carbon trading cost range, represents the actual carbon emission range, represents the carbon emission quota range, β Represents the carbon trading base price, T represents the scheduling period, Represents the power generation range of the cogeneration unit, K p Represents the unit power generation loss of natural gas during the power generation process of the cogeneration unit, l represents the carbon emission coefficient of natural gas, represent t The period of external power grid purchase, G lm represents the dynamic carbon emission factor, represent t The output power range of the gas boiler during the period, l e represents the carbon emission quota consumed by the external grid, l g Represents the carbon emission quota per unit of electricity consumption of natural gas units, Representing a combined heat and power unit t The electric power output interval of the time period, Representing a combined heat and power unit t The thermal power range output during the time period.

[0008] Optionally, by measuring the carbon emission quotas of energy supply equipment and electricity purchased from external power grids, a carbon emission quota model for the integrated energy system is constructed, including: The external grid electricity purchase quota is constrained based on the carbon emission quota of the external grid electricity consumption, and the indirect carbon emission quota of the integrated energy system is obtained; The output power of energy supply equipment is constrained according to the carbon emission quota of the power consumption of energy supply equipment to obtain the direct carbon emission quota of the integrated energy system; By measuring indirect carbon emission quotas and direct carbon emission quotas, a carbon emission quota model for the integrated energy system is constructed; Among them, the external power grid is a coal-fired unit, and the energy supply equipment includes wind turbines and cogeneration units and gas boilers whose energy input is natural gas.

[0009] Optionally, the historical wind power data is cleaned and normalized using the quartile method, meteorological features are extracted using a convolutional neural network, temporal dynamic features are captured through a bidirectional gated recurrent unit, an attention mechanism is introduced to enhance information, and kernel density estimation is combined to generate a wind power forecast interval containing the forecast error distribution, including: Perform quartile cleaning and normalization on historical wind power data, and train a wind power hybrid prediction model based on the processed historical wind power data; According to the convolutional neural network layer, the meteorological feature vector is extracted from the wind farm data matrix obtained by the input layer transformation; A bidirectional gated recurrent unit layer is used to capture the dynamic characteristics of the time series in the meteorological feature vector; The hidden layer information of the bidirectional gated recurrent unit layer is enhanced according to the attention layer, and the optimal meteorological feature vector is output to the output layer for integration to obtain the wind power prediction value; The wind power prediction interval is obtained by fitting the error data of the wind power prediction value with a normal distribution using kernel density estimation and superimposing the generated random noise data with the wind power prediction value. Among them, the wind power hybrid prediction model includes: input layer, convolutional neural network layer, bidirectional gated recurrent unit layer, attention layer and output layer.

[0010] Optionally, performing quartile cleaning and normalization processing on historical wind power data includes: Obtain historical wind power data of wind turbines in the integrated energy system, including wind power data and meteorological characteristic data; Using the quartile method, outliers and / or missing values ​​in wind power data that exceed the set threshold are corrected to within the threshold range; The meteorological characteristic data and wind power data cleaned by the quartile method are normalized to obtain standardized historical wind power data.

[0011] Optionally, after performing quartile cleaning and normalization processing on historical wind power data, extracting meteorological features using a convolutional neural network, capturing temporal dynamic features through a bidirectional gated recurrent unit, introducing an attention mechanism to enhance information, and combining kernel density estimation to generate a wind power forecast interval containing a forecast error distribution, the following is also included: The wind power forecast interval is evaluated based on the set evaluation indicators including interval coverage, interval root mean square error and interval average width, and the optimal wind power forecast interval is obtained based on the evaluation results; Based on the wind curtailment penalty coefficient of the integrated energy system and the rated output range of wind power, the wind curtailment cost is quantified in combination with the optimal wind power prediction range to obtain the wind curtailment cost range of the integrated energy system.

[0012] Optionally, the wind power forecast interval is evaluated based on set evaluation indicators including interval coverage, interval root mean square error, and interval average width, and the optimal wind power forecast interval obtained based on the evaluation results includes: According to the position information of the wind power forecast value in the wind power forecast interval, the interval coverage rate of the current wind power forecast interval is calculated; Calculating the root mean square error of the wind power prediction interval based on the acquired real-time wind power to obtain the interval root mean square error of the current wind power prediction interval; Calculating the average width of the wind power prediction interval according to the obtained wind power rated output interval to obtain the average width of the current wind power prediction interval; When the interval coverage, interval root mean square error and interval average width all meet the set threshold conditions, the current wind power prediction interval is determined as the optimal wind power prediction interval.

[0013] Optionally, based on the economic dispatch model and objective function, and subject to a set of constraints including the operating boundaries of energy supply equipment, the charging and discharging intervals of the energy storage system, and the multi-energy balance of electricity, heat, and gas, an optimal dispatch solution is solved using the dynamic carbon emission metering interval and the wind power forecast interval. The optimized dispatch solution includes: Based on the topological structure of the integrated energy system, determine the operating characteristic parameters of each energy source in the integrated energy system; Based on the technical parameters of the energy supply equipment and the operating characteristic parameters of each energy source, a set of constraints is set, including the operating boundaries of the energy conversion equipment, the charging and discharging intervals of the energy storage system, and the multi-energy balance of electricity, heat and gas. When all constraints of the constraint set are met, the dynamic carbon emission measurement interval and wind power forecast interval are used as input to solve the economic dispatch model with the objective function as the core, and the optimal dispatch plan of the integrated energy system is obtained; The energy supply equipment operation boundaries include: wind power output constraints, cogeneration unit operation constraints, and gas boiler operation constraints; The charging and discharging range of the energy storage system includes: charging and discharging power limit range and energy storage capacity limit; The multi-energy balance of electricity, heat and gas includes: electric power balance, thermal power balance, gas power balance, electricity purchase limit and gas purchase limit.

[0014] In a second aspect, an embodiment of the present invention provides an integrated energy system, including: Energy demand side; The energy supply side connected to the energy demand side includes wind turbines, cogeneration units, gas boilers, and external power grids; The energy dispatching platform connected to the energy supply side is used to execute the above-mentioned steps of the integrated energy system optimization dispatching method based on carbon emission factors and wind power forecasts.

[0015] (3) Beneficial effects The beneficial effects of the present invention are as follows: the present invention uses a carbon emission measurement method based on a dynamic carbon emission factor to measure the actual carbon emissions generated in the system, and through the carbon emissions of the adaptive carbon response, it can formulate a targeted scheduling strategy to achieve the purpose of energy conservation and carbon reduction. At the same time, the present invention also uses a wind power hybrid prediction model improved by a convolutional neural network, a bidirectional gated recurrent unit, and an attention mechanism to perform more accurate wind power interval prediction, which provides a reliable basis for system scheduling, reduces the use of non-renewable energy by energy supply equipment, improves the utilization efficiency of renewable energy such as wind energy, and ensures the safe and stable operation of the integrated energy system. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 A flow chart of a method for optimizing and scheduling an integrated energy system based on carbon emission factors and wind power forecasting provided by one embodiment of the present invention; Figure 2 A diagram of an integrated energy system architecture provided in accordance with an embodiment of the present invention. DETAILED DESCRIPTION

[0017] In order to better explain the present invention and facilitate understanding, the present invention is described in detail below through specific implementation methods in conjunction with the accompanying drawings.

[0018] refer to Figure 1As shown, an embodiment of the present invention proposes a method for optimizing and dispatching an integrated energy system based on carbon emission factors and wind power forecasts, including: establishing a carbon emission measurement model based on dynamic carbon emission factors by measuring direct carbon emissions from energy supply equipment and indirect carbon emissions from electricity purchased from external power grids; performing quartile cleaning and normalization on historical wind power data, extracting meteorological features using a convolutional neural network, capturing temporal dynamic features using a bidirectional gated recurrent unit, introducing an attention mechanism to enhance information, and generating a wind power prediction interval containing a prediction error distribution using kernel density estimation; constructing an economic dispatch model containing interval variables with the objective function of minimizing the sum of energy purchase costs, carbon trading costs, and wind power curtailment costs; wherein the carbon trading costs are calculated based on the difference between the dynamic carbon emission measurement interval and the quota, and the wind power curtailment costs are penalized based on the deviation between the wind power prediction interval and the actual output; based on the economic dispatch model and the objective function, an optimized dispatch solution is solved using the dynamic carbon emission measurement interval and the wind power prediction interval, while satisfying a set of constraints including the operating boundaries of the energy supply equipment, the charging and discharging intervals of the energy storage system, and the multi-energy balance of electricity, heat, and gas.

[0019] This embodiment uses a carbon emission measurement method based on a dynamic carbon emission factor to measure the actual carbon emissions generated in the system. This allows for the development of targeted scheduling strategies based on the carbon emissions of an adaptive carbon response, achieving the goal of energy conservation and carbon reduction. Furthermore, this embodiment uses a hybrid wind power prediction model improved by a convolutional neural network, a bidirectional gated recurrent unit, and an attention mechanism to achieve higher-precision wind power range forecasts. This provides a reliable basis for system scheduling, reduces the use of non-renewable energy by energy supply equipment, improves the utilization efficiency of renewable energy such as wind energy, and ensures the safe and stable operation of the integrated energy system.

[0020] To better understand the above technical solutions, exemplary embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present invention are shown in the accompanying drawings, it should be understood that the present invention can be implemented in various forms and should not be limited by the embodiments described herein. Rather, these embodiments are provided to enable a clearer and more thorough understanding of the present invention and to fully convey the scope of the present invention to those skilled in the art.

[0021] Specifically, refer to Figure 1 As shown, this embodiment proposes a method for optimizing and scheduling an integrated energy system based on carbon emission factors and wind power forecasts, which includes: S100, establishing a carbon emission measurement model based on dynamic carbon emission factors by measuring the direct carbon emissions of energy supply equipment and the indirect carbon emissions of electricity purchased from the external power grid.

[0022] refer to Figure 2As shown, the optimized scheduling and operation of an integrated energy system primarily focuses on the energy supply and energy demand processes, without considering the energy transmission process. On the energy supply side, the integrated energy system purchases electricity and natural gas from the upstream power grid (external grid) and natural gas grid, while integrating its own wind turbines (WT), combined heat and power generation (CHP) units, and gas boilers (GB) to meet residents' energy needs. On the energy demand side, residents use electricity and heat generated or purchased from the energy supply side to meet their energy needs. Carbon emissions in the integrated energy system include direct carbon emissions from the combustion of natural gas in the CHP units and gas boilers, as well as indirect carbon emissions from the transmission of electricity from the external grid.

[0023] In this embodiment, step S100 may include the following sub-steps S110 to S150: S110. Based on the carbon emission coefficient of natural gas, carbon emissions are quantified in combination with the energy output of the gas boiler and the cogeneration unit in the energy supply equipment to obtain the direct carbon emissions of the gas boiler and the cogeneration unit.

[0024] Furthermore, the carbon emission coefficient of natural gas is obtained by converting the physical properties of carbon emission of natural gas in the natural gas grid into chemical characteristics. l The calculation formula is: (1) In formula (1), A p Represents the average lower calorific value of natural gas, C p represents the carbon content of natural gas, γ p Represents the oxidation rate of natural gas.

[0025] Direct carbon emissions from gas boilers and cogeneration units C The calculation formula of 1 is: (2) In formula (2), E p represents the power generation of the cogeneration unit, K p Represents the unit power generation loss of natural gas during the power generation process of the cogeneration unit, E b Represents the calorific value of the gas boiler. K b Represents the unit heat loss of natural gas during the heating process of a gas boiler.

[0026] S120. Carbon emissions are quantified based on the wind power carbon emission coefficient of the wind turbine in the energy supply equipment and combined with the wind power generation to obtain the wind power carbon emissions of the wind turbine.

[0027] S130. Obtain a static carbon emission factor for the integrated energy system by measuring the total carbon emissions of direct carbon emissions, wind power carbon emissions, and indirect carbon emissions from purchasing electricity from the external power grid, and performing ratio processing on the total carbon emissions and the total energy output of the integrated energy system.

[0028] Static carbon emission factor of the integrated energy system G The calculation formula of 1 is: (3) In formula (3), e r represents the carbon emission coefficient of the external power grid, P e,b ( t ) represents the amount of electricity purchased from the external power grid, E k represents wind power generation, e k Represents the carbon emission coefficient of wind power.

[0029] S140. Based on the ratio of the total real-time carbon emissions of the integrated energy system to the non-new energy power generation, the static carbon emission factor is dynamically corrected in real time to obtain a dynamic carbon emission factor with an adaptive carbon response.

[0030] Real-time total carbon emissions of the integrated energy system C s The calculation formula is: (4) In formula (4), e p represents the carbon emission coefficient of the combined heat and power unit, e b Represents the carbon emission coefficient of the gas boiler.

[0031] In this embodiment, non-renewable energy power generation (excluding the remaining power generation from renewable energy) is used to generate a dynamic carbon emission factor that responds to the system's carbon emissions in real time. The target carbon emissions of gas boilers and external power purchases are dynamically calculated based on the actual power supply from renewable energy, achieving precise control of carbon emissions and achieving the goal of carbon emission reduction. The dynamic carbon emission factor G lm The expression is: (5) In formula (5), E d Represents non-renewable energy generation.

[0032] S150. Calculate the real-time carbon emissions of the gas boiler and electricity purchased from the external power grid based on the dynamic carbon emission factor, and measure the calculation results with the real-time carbon emissions of the cogeneration unit to obtain a carbon emission measurement model.

[0033] The expression of the carbon emission measurement model is: (6) In formula (6), E s represents the actual carbon emissions generated by the system, E CHP represents the actual carbon emissions of the cogeneration unit, E r,b Represents the actual carbon emissions from gas boilers and electricity purchased from external grids, T represents the scheduling period, P GB,h ( t ) represents a gas boiler t Output power during a period of time.

[0034] In this embodiment, after step S100, the following sub-steps F100 to F300 are further included: F100. Input the energy output range of the energy supply equipment and the electricity purchase range of the external power grid into the carbon emission measurement model to measure carbon emissions and obtain the carbon emission range of the integrated energy system.

[0035] F200. By measuring the carbon emission quotas of energy supply equipment and electricity purchased from external power grids, a carbon emission quota model for the integrated energy system is constructed.

[0036] Furthermore, step F200 may include the following sub-steps F210 to F230: F210: Based on the carbon emission quota of electricity consumed by the external power grid, the power purchase quota of the external power grid is constrained to obtain the indirect carbon emission quota of the integrated energy system. The external power grid refers to the coal-fired unit.

[0037] F220: Based on the carbon emission quotas of the electricity consumed by energy supply equipment, quotas are imposed on the output power of energy supply equipment to obtain direct carbon emission quotas for the integrated energy system. Energy supply equipment includes wind turbines and combined heat and power units and gas boilers that use natural gas as energy input.

[0038] F230. By measuring indirect carbon emission quotas and direct carbon emission quotas, a carbon emission quota model for the integrated energy system is constructed.

[0039] The expression of the carbon emission quota model is: (7) In formula (7), Ds Represents the carbon emission quota of the integrated energy system, D e,b Represents the carbon emission quota for electricity purchased from the external power grid, D CHP Represents the carbon emission quota of the combined heat and power unit, D GB Represents the carbon emission quota of gas boilers, l e represents the carbon emission quota consumed by the external grid, l g Represents the carbon emission quota per unit of electricity consumption of natural gas units, P CHP,e ( t ) represents a combined heat and power unit t The output power of the time period represents, P CHP,h ( t ) represents a combined heat and power unit t Thermal power output during a period of time.

[0040] F300. Based on the carbon emission quota range of the integrated energy system output by the carbon emission quota model, the carbon emission quota difference of the carbon emission range is calculated in combination with the carbon trading base price to obtain the carbon trading cost range of the integrated energy system.

[0041] Among them, the expression of carbon trading cost range is: (8) In formula (8), represents the carbon trading cost range, represents the actual carbon emission range, represents the carbon emission quota range, β Represents the carbon trading base price, Represents the power generation range of the cogeneration unit, represent t The period of external power grid purchase, represent t The output power range of the gas boiler during the period, Representing a combined heat and power unit t The electric power output interval of the time period, Representing a combined heat and power unit t The thermal power range output during the time period.

[0042] S200. The historical wind power data is cleaned and normalized using the quartile method, meteorological features are extracted using a convolutional neural network, temporal dynamic features are captured through a bidirectional gated recurrent unit, an attention mechanism is introduced to enhance information, and kernel density estimation is combined to generate a wind power prediction interval containing the prediction error distribution.

[0043] In this embodiment, the hybrid wind power prediction model includes an input layer, a convolutional neural network layer, a bidirectional gated recurrent unit layer, an attention layer, and an output layer. The convolutional neural network layer extracts meteorological data features such as wind speed and direction, while the bidirectional gated recurrent unit layer processes time series information. An attention mechanism then enhances focus on key information, ultimately resulting in a wind power prediction result at the output layer. This wind power range prediction provides a reliable basis for system scheduling, reduces the use of non-renewable energy sources by energy supply equipment, improves the utilization efficiency of renewable energy sources such as wind power, and ensures the safe and stable operation of the integrated energy system.

[0044] In this embodiment, step S200 may include the following sub-steps S210 to S250: S210: Perform quartile cleaning and normalization processing on the historical wind power data, and train a wind power hybrid prediction model based on the processed historical wind power data.

[0045] Furthermore, step S210 may include the following sub-steps S211 to S213: S211. Obtain historical wind power data of wind turbines in the integrated energy system, where the historical wind power data includes wind power data and meteorological characteristic data.

[0046] S212: Using the quartile method, correct abnormal values ​​and / or missing values ​​in the wind power data that exceed a set threshold value to within a threshold range.

[0047] In this embodiment, the quartile method is used to process abnormal data of wind power. Before training the wind power prediction model, it is usually necessary to process the collected data containing abnormal values ​​and / or missing values. The range of normal wind power data obtained by processing abnormal data using the quartile method is: (9) In formula (9), l h 、 l l Respectively represent the upper and lower limits of normal wind power data, K HQ represents the interquartile range, H 1. H 3 represents the 1st and 3rd quantiles of wind power input data.

[0048] When wind power data is less than l l or greater than l hWhen , the wind power data will be judged as an outlier and needs to be processed, and the cleaned abnormal data and missing data need to be filled using interpolation method.

[0049] (10) In formula (10), l t-1 、 l t+1 Respectively represent t -1 period and t +1 period of wind power input data, l t represent t Wind power input data for the period.

[0050] S213. Normalize the meteorological characteristic data and the wind power data cleaned by the quartile method to obtain standardized historical wind power data.

[0051] In this embodiment, different meteorological characteristic data have different dimensions, which will have an adverse effect on the learning process of the wind power hybrid prediction model. In order to ensure that the model can effectively learn and improve the training speed and convergence, it is necessary to normalize the wind speed, wind direction, temperature, humidity and historical power data. The minimum-maximum normalization method is used to normalize the original data such as wind speed, wind direction, temperature, humidity, etc.

[0052] (11) In formula (11), x i-new Representative i The data output by the normalization operation, x i Represents the original data, x max Represents the maximum value of wind power data, x min Represents the minimum value of wind power data, N Represents the total amount of raw data.

[0053] S220. Extracting meteorological feature vectors from the wind farm data matrix obtained by transformation of the input layer according to the convolutional neural network layer.

[0054] In this embodiment, the input layer of the wind power hybrid prediction model preprocesses the input wind farm data to generate a wind farm data matrix X , mainly including wind speed, wind direction, temperature, humidity and wind power of the wind farm. Then, the convolutional neural network layer is used to extract the wind farm data matrix XExtract meteorological feature vectors. The convolutional neural network layer consists of a convolutional layer, a pooling layer, and a fully connected layer, which is used to extract features from the input wind farm data matrix. Formulas (12) to (14) present the meteorological feature vectors output by the convolutional neural network layer. H c process.

[0055] (12) In formula (12), C represents the output of the convolutional layer, f Represents the output solution function of the convolutional layer, X represents the wind farm data matrix, W 1 represents the weight matrix, b 1 represents bias, s Represents the sigmoid activation function.

[0056] (13) In formula (13), Pl represents the output of the pooling layer, C represents the output of the convolutional layer, max(·) represents the maximum value function, b 2 represents bias.

[0057] (14) In formula (14), W 2 represents the weight matrix, b 3 represents bias.

[0058] S230. Use a bidirectional gated recurrent unit layer to capture the dynamic characteristics of the time series in the meteorological feature vector.

[0059] In this embodiment, the bidirectional gated recurrent unit layer is mainly used to further mine the dynamic characteristics of the time series from the meteorological feature vector extracted from the convolutional neural network layer. t Dynamic characteristics of moment output y t The calculation process is shown in formula (15).

[0060] (15) In formula (15), BiGRU(·) represents the bidirectional gated recurrent unit that combines the forward and reverse hidden states to mine the dynamic characteristics of the time series. Represents the forward and backward hidden layer states of a bidirectional gated recurrent unit.

[0061] S240. Strengthen the hidden layer information of the bidirectional gated recurrent unit layer according to the attention layer, output the optimal meteorological feature vector to the output layer for integration, and obtain the wind power prediction value.

[0062] In this embodiment, the input vector of the attention layer is the hidden layer information of the bidirectional gated recurrent unit layer. The attention layer assigns different weights to reflect the importance of various feature information by calculating the probability distribution. The importance of different features at each moment is calculated by equations (16) and (17). Hidden layer unit h t ( t ∈[1, t ]) Scoring is performed through the softmax function, and the calculation process is shown in formula (18).

[0063] (16) In formula (16), Represents the "importance" of the current input to the output, v 、 w Represents the model parameters to be trained, tanh represents the hyperbolic tangent activation function, b Represents bias.

[0064] (17) In formula (17), α i represents the attention weights of the hidden layer output of the bidirectional gated recurrent unit layer, ( i ∈[1, t ]), e i represent i The features obtained at each moment are passed through the convolutional neural network layer and the bidirectional gated recurrent unit layer.

[0065] (18) In formula (18), S i represent i The output of the attention layer at any moment, h i Represents the output of the hidden layer of the bidirectional gated recurrent unit layer.

[0066] In this embodiment, the output layer is a fully connected layer that integrates the output of the attention layer to obtain the final prediction result. Y t , the calculation process is shown in formula (19).

[0067] (19) In formula (19), w o 、b o represents the weight bias, s Represents the sigmoid activation function.

[0068] S250. Perform normal distribution fitting on the error data of the wind power prediction value using kernel density estimation, superimpose the generated random noise data with the wind power prediction value, and obtain the wind power prediction interval.

[0069] In this embodiment, since the wind power prediction error of each period approximately obeys the normal distribution, the non-parametric kernel density estimation method can be used to construct the probability density function (20) of the wind power prediction error to fit the normal distribution parameters of the wind power prediction error data. The wind power prediction error data is fitted with a normal distribution, and the generated random noise is superimposed with the wind power prediction value to obtain a wind power interval input data set. The non-parametric kernel density estimation method is used to fit the normal distribution parameters of the N wind power samples in the data set. . t The probability density function of wind power forecast error at the moment can be expressed as: (20) In formula (20), F N,t represents the probability density function of wind power forecast error, x Represents the wind power output value, s represents the variance, P t,p Represents the predicted wind power output value.

[0070] (twenty one) In formula (21), represents the predicted output range of wind power, [·] represents the interval number form, |·| represents the absolute value, P t,up 、 P t,dw Represents the upper and lower limits of the wind power range.

[0071] In this embodiment, after step S200, the following steps G100 to G200 are further included: G100: Evaluate the wind power forecast interval based on the set evaluation indicators including interval coverage, interval root mean square error, and interval average width, and obtain the optimal wind power forecast interval based on the evaluation results.

[0072] In this embodiment, after obtaining the upper and lower limits of the wind power prediction interval, it is necessary to evaluate the wind power prediction interval. In this embodiment, interval coverage, interval root mean square error, and interval average width are used as evaluation indicators of interval prediction error to evaluate the wind power prediction interval. The specific evaluation process is as follows G110 to G140: G110. Calculate the interval coverage rate of the current wind power prediction interval based on the position information of the wind power prediction value in the wind power prediction interval.

[0073] Interval coverage I C The calculation formula is: (twenty two) (twenty three) In formula (22), s t Represents 0, 1 variables, N represents the number of samples, I C Larger values ​​indicate more reliable prediction intervals.

[0074] In formula (23), Represents the predicted output range of wind power, P t,rl represent t The actual value of wind power at the moment.

[0075] G120. Calculate the root mean square error of the wind power prediction interval based on the acquired real-time wind power to obtain the interval root mean square error of the current wind power prediction interval.

[0076] Interval root mean square error I RE The calculation formula is: (twenty four) G130. Calculate the average width of the wind power prediction interval based on the obtained wind power rated output interval to obtain the average width of the current wind power prediction interval.

[0077] Average interval width A WI The calculation formula is: (25) In formula (25), P t,p Represents the predicted value of wind power.

[0078] G140: When the interval coverage, interval root mean square error, and interval average width all meet the set threshold conditions, the current wind power forecast interval is determined as the optimal wind power forecast interval.

[0079] G200: Based on the wind curtailment penalty coefficient of the integrated energy system and the rated output range of wind power, combined with the optimal wind power prediction range, the wind curtailment cost is quantified to obtain the wind curtailment cost range of the integrated energy system.

[0080] Wind curtailment cost range The expression is: (26) In formula (26), W p represents the wind curtailment penalty coefficient of the integrated energy system, represent t Rated output range of wind power during the period.

[0081] S300, with the minimization of the sum of energy purchase cost, carbon trading cost and wind curtailment cost as the objective function, an economic dispatch model with interval variables is constructed.

[0082] In this embodiment, the calculation formula for the purchased energy cost is: (27) In formula (27), Represents the energy purchase cost range, c e,b ( t ), c g,b ( t ) represent the electricity purchase price and gas purchase price of the integrated energy system from the external power grid and gas grid, respectively. Respectively represent t The power range of electricity purchase and gas purchase in the time period.

[0083] In the embodiment, the objective function is to minimize the sum of energy purchase cost, carbon trading cost and wind curtailment cost, and the range of the objective function is: (28) S400. Based on the economic dispatch model and objective function, while satisfying the set of constraints including the operating boundaries of the energy supply equipment, the charging and discharging intervals of the energy storage system, and the multi-energy balance of electricity, heat and gas, the optimal dispatch plan is solved through the dynamic carbon emission metering interval and the wind power prediction interval.

[0084] In this embodiment, step S400 may include the following sub-steps S410 to S430: S410. Based on the topological structure of the integrated energy system, determine the operating characteristic parameters of each energy source in the integrated energy system.

[0085] S420. According to the technical parameters of the energy supply equipment and the operating characteristic parameters of each energy source, a set of constraints including the operating boundary of the energy conversion equipment, the charging and discharging range of the energy storage system, and the multi-energy balance of electricity, heat and gas is set.

[0086] In this embodiment, the operating boundaries of the energy conversion equipment include: Wind power output constraints: (29) Cogeneration unit operating constraints: (30) In formula (30), They represent the efficiency of converting electrical energy and thermal energy when the cogeneration unit is working. 、 Respectively represent the upper and lower limits of the ramp power of the cogeneration unit, Represents the natural gas power range of the combined heat and power unit, 、 They represent the upper and lower limits of the heat-to-electricity ratio of the cogeneration unit, 、 They represent the upper and lower limits of the input power of the cogeneration unit respectively.

[0087] Gas boiler operation constraints: (31) In formula (31), Represents the heat production efficiency of the gas boiler, Represents the range of natural gas input power of gas boiler, Respectively represent the upper and lower limits of the gas boiler's ramp power. Respectively represent the upper and lower limits of the gas boiler output.

[0088] In this embodiment, the charging and discharging intervals of the energy storage system are: (32) In formula (32), Represents the charging power range of the energy storage device, Represents the discharge power range of the energy storage device, Represents 0, 1 variables, Represents the maximum charging and discharging power of the energy storage device, Representative n Energy storage device t The final output power of the time period, Represent the charging and discharging efficiency of the energy storage device, Represents the output power range of the energy storage device, Representative n The capacity range of energy storage devices, Represents the rated capacity of the energy storage device, Respectively represent the upper and lower limits of the capacity of the energy storage device, Represents the initial moment n The capacity of the energy storage device, Represents the final moment n The capacity of an energy storage device.

[0089] The multi-energy balance constraints of electricity, heat and gas are: (33) (34) (35) In formulas (33), (34), and (35), Represents integrated energy systems t Electricity load, heat load and gas load of the time period, Represent the power range of input electricity, heat and gas energy storage respectively, They represent the electricity and gas purchase limits for each time period respectively.

[0090] S430. When all constraints of the constraint set are met, the dynamic carbon emission measurement interval and the wind power prediction interval are used as input to solve the economic dispatch model with the objective function as the core, and the optimal dispatch plan of the integrated energy system is obtained.

[0091] On the other hand, reference Figure 2 As shown, this embodiment also provides an integrated energy system, comprising: an energy demand side; an energy supply side connected to the energy demand side, including a wind turbine, a cogeneration unit, a gas boiler, and an external power grid; and an energy scheduling platform connected to the energy supply side, configured to execute the above-described integrated energy system optimization scheduling method based on carbon emission factors and wind power forecasts.

[0092] In summary, the present embodiment proposes a method and system for optimizing and scheduling an integrated energy system based on carbon emission factors and wind power forecasts, which can achieve the goal of carbon emission reduction while ensuring the safe and stable operation of the system. This embodiment accurately measures the carbon emissions generated in the system through a carbon emission measurement method based on a dynamic carbon emission factor, and performs wind power forecasting based on an improved wind power hybrid forecasting model comprising a convolutional neural network, a bidirectional gated recurrent unit, and an attention mechanism. It also uses a non-parametric kernel density estimation method to combine point forecasts and forecast errors to generate interval forecast results, thereby providing a reliable basis for the optimal scheduling of the system, reducing the use of fossil energy, maximizing the use of renewable energy such as wind energy, and achieving the goal of energy conservation and emission reduction while ensuring the safe and stable operation of the system.

[0093] Since the systems / devices described in the above embodiments of the present invention are systems / devices used to implement the methods of the above embodiments of the present invention, those skilled in the art will be able to understand the specific structures and variations of these systems / devices based on the methods described in the above embodiments of the present invention, and thus will not be described in detail here. All systems / devices used in the methods of the above embodiments of the present invention are within the scope of protection of the present invention.

[0094] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROMs, optical storage, etc.) containing computer-usable program code.

[0095] The present invention is described with reference to flowcharts and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions.

[0096] It should be noted that, in the description of the present invention, the word "a" or "an" preceding a component does not exclude the presence of a plurality of such components. The present invention can be implemented by means of hardware comprising several distinct components and by means of a suitably programmed computer. The use of the words first, second, third, etc., is merely for convenience and does not imply any order. These words should be understood as part of the component name.

[0097] In addition, it should be noted that, in the description of this specification, the description of the terms "one embodiment", "some embodiments", "embodiment", "example", "specific example" or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art can combine and combine different embodiments or examples described in this specification and the features of different embodiments or examples, unless they are contradictory.

[0098] Although preferred embodiments of the present invention have been described, those skilled in the art will be able to make additional changes and modifications to these embodiments after obtaining the basic inventive concepts.

[0099] Obviously, those skilled in the art can make various modifications and variations to the present invention without departing from the spirit and scope of the invention.

Claims

1. A method for optimizing the scheduling of an integrated energy system based on carbon emission factors and wind power forecasts, characterized in that: include: By measuring the direct carbon emissions of energy supply equipment and the indirect carbon emissions from electricity purchases from external power grids, a carbon emission measurement model based on dynamic carbon emission factors is established; Historical wind power data is cleaned and normalized using the quartile method. A convolutional neural network is used to extract meteorological features. A bidirectional gated recurrent unit is used to capture temporal dynamic features. An attention mechanism is introduced to enhance information. Kernel density estimation is then used to generate wind power forecast intervals that include the forecast error distribution. An economic dispatch model with interval variables is constructed, with the objective function minimizing the sum of energy purchase costs, carbon trading costs, and wind curtailment costs. The carbon trading costs are calculated based on the difference between the dynamic carbon emission measurement interval and the quota, while the wind curtailment costs are penalized based on the deviation between the wind power forecast interval and the actual output. Based on the economic dispatch model and objective function, the optimal dispatch scheme is solved through the dynamic carbon emission metering interval and wind power prediction interval, while satisfying the constraint condition set including the operating boundary of energy supply equipment, the charging and discharging range of the energy storage system, and the multi-energy balance of electricity, heat and gas.

2. The method according to claim 1, wherein By measuring the direct carbon emissions of energy supply equipment and the indirect carbon emissions of electricity purchased from the external power grid, a carbon emission measurement model based on dynamic carbon emission factors is established, including: Based on the carbon emission coefficient of natural gas, the carbon emissions are quantified by combining the energy output of gas boilers and cogeneration units in the energy supply equipment to obtain the direct carbon emissions of gas boilers and cogeneration units; Based on the wind power carbon emission coefficient of the wind turbine in the energy supply equipment, combined with the wind power generation, carbon emissions are quantified to obtain the wind power carbon emissions of the wind turbine; By measuring the total carbon emissions of direct carbon emissions, wind power carbon emissions, and indirect carbon emissions from external grid power purchases, and then comparing the total carbon emissions with the total energy output of the integrated energy system, the static carbon emission factor of the integrated energy system is obtained. According to the ratio of the total real-time carbon emissions of the integrated energy system to the non-new energy power generation, the static carbon emission factor is dynamically corrected in real time to obtain a dynamic carbon emission factor with adaptive carbon response; The real-time carbon emissions of gas boilers and electricity purchased from external power grids are calculated based on dynamic carbon emission factors, and the calculation results are measured against the real-time carbon emissions of the cogeneration unit to obtain a carbon emission measurement model. Among them, the expression of the carbon emission measurement model is: ; Where, E s represents the actual carbon emissions generated by the system, E CHP represents the actual carbon emissions of the cogeneration unit, E r,b Represents the actual carbon emissions from gas boilers and electricity purchased from external grids, T represents the scheduling period, E p represents the power generation of the cogeneration unit, K p Represents the unit power generation loss of natural gas during the power generation process of the cogeneration unit, λ Represents the carbon emission coefficient of natural gas ， P e,b ( t ) represents the amount of electricity purchased from the external power grid, P GB,h ( t ) represents a gas boiler t Output power during the time period, G lm Represents the dynamic carbon emission factor.

3. The method according to claim 1, wherein After establishing a carbon emission measurement model based on dynamic carbon emission factors by measuring the direct carbon emissions of energy supply equipment and the indirect carbon emissions of electricity purchased from the external grid, it also includes: The energy output range of the energy supply equipment and the electricity purchase range of the external power grid are input into the carbon emission measurement model to measure carbon emissions and obtain the carbon emission range of the integrated energy system; By measuring the carbon emission quotas of energy supply equipment and electricity purchased from external power grids, a carbon emission quota model for the integrated energy system is constructed; Based on the carbon emission quota range of the integrated energy system output by the carbon emission quota model, the carbon emission quota difference is calculated for the carbon emission range in combination with the carbon trading base price to obtain the carbon trading cost range of the integrated energy system; Among them, the expression of carbon trading cost range is: ; Where, represents the carbon trading cost range, represents the actual carbon emission range, represents the carbon emission quota range, β Represents the carbon trading base price, T represents the scheduling period, Represents the power generation range of the cogeneration unit, K p Represents the unit power generation loss of natural gas during the power generation process of the cogeneration unit, λ represents the carbon emission coefficient of natural gas, represent t The period of external power grid purchase, G lm represents the dynamic carbon emission factor, represent t The output power range of the gas boiler during the period, λ e represents the carbon emission quota consumed by the external grid, λ g Represents the carbon emission quota per unit of electricity consumption of natural gas units, Representing a combined heat and power unit t The electric power output interval of the time period, Representing a combined heat and power unit t The thermal power range output during the time period.

4. The method according to claim 3, wherein By measuring the carbon emission quotas of energy supply equipment and electricity purchased from external power grids, a carbon emission quota model for the integrated energy system is constructed, including: The external grid electricity purchase quota is constrained based on the carbon emission quota of the external grid electricity consumption, and the indirect carbon emission quota of the integrated energy system is obtained; The output power of energy supply equipment is constrained according to the carbon emission quota of the power consumption of energy supply equipment to obtain the direct carbon emission quota of the integrated energy system; By measuring indirect carbon emission quotas and direct carbon emission quotas, a carbon emission quota model for the integrated energy system is constructed; Among them, the external power grid is a coal-fired unit, and the energy supply equipment includes wind turbines and cogeneration units and gas boilers whose energy input is natural gas.

5. The method according to claim 1, wherein The historical wind power data is cleaned and normalized using the quartile method. A convolutional neural network is used to extract meteorological features. A bidirectional gated recurrent unit is used to capture temporal dynamic features. An attention mechanism is introduced to enhance information. Combined with kernel density estimation, a wind power forecast interval containing the forecast error distribution is generated. The following are included: Perform quartile cleaning and normalization on historical wind power data, and train a wind power hybrid prediction model based on the processed historical wind power data; According to the convolutional neural network layer, the meteorological feature vector is extracted from the wind farm data matrix obtained by the input layer transformation; A bidirectional gated recurrent unit layer is used to capture the dynamic characteristics of the time series in the meteorological feature vector; The hidden layer information of the bidirectional gated recurrent unit layer is enhanced according to the attention layer, and the optimal meteorological feature vector is output to the output layer for integration to obtain the wind power prediction value; The wind power prediction interval is obtained by fitting the error data of the wind power prediction value with a normal distribution using kernel density estimation and superimposing the generated random noise data with the wind power prediction value. Among them, the wind power hybrid prediction model includes: input layer, convolutional neural network layer, bidirectional gated recurrent unit layer, attention layer and output layer.

6. The method according to claim 5, wherein The quartile cleaning and normalization of historical wind power data includes: Obtain historical wind power data of wind turbines in the integrated energy system, including wind power data and meteorological characteristic data; Using the quartile method, outliers and / or missing values ​​in wind power data that exceed the set threshold are corrected to within the threshold range; The meteorological characteristic data and wind power data cleaned by the quartile method are normalized to obtain standardized historical wind power data.

7. The method according to claim 1, wherein After quartile cleaning and normalization of historical wind power data, using a convolutional neural network to extract meteorological features, capturing temporal dynamic features through a bidirectional gated recurrent unit, introducing an attention mechanism to enhance information, and combining kernel density estimation to generate wind power forecast intervals containing forecast error distributions, the following also applies: The wind power forecast interval is evaluated based on the set evaluation indicators including interval coverage, interval root mean square error and interval average width, and the optimal wind power forecast interval is obtained based on the evaluation results; Based on the wind curtailment penalty coefficient of the integrated energy system and the rated output range of wind power, the wind curtailment cost is quantified in combination with the optimal wind power prediction range to obtain the wind curtailment cost range of the integrated energy system.

8. The method according to claim 7, wherein The wind power forecast interval is evaluated based on the set evaluation indicators including interval coverage, interval root mean square error, and interval average width. The optimal wind power forecast interval obtained based on the evaluation results includes: According to the position information of the wind power forecast value in the wind power forecast interval, the interval coverage rate of the current wind power forecast interval is calculated; Calculating the root mean square error of the wind power prediction interval based on the acquired real-time wind power to obtain the interval root mean square error of the current wind power prediction interval; Calculating the average width of the wind power prediction interval according to the obtained wind power rated output interval to obtain the average width of the current wind power prediction interval; When the interval coverage, interval root mean square error and interval average width all meet the set threshold conditions, the current wind power prediction interval is determined as the optimal wind power prediction interval.

9. The method according to claim 1, wherein Based on the economic dispatch model and objective function, and under the constraints of the energy supply equipment operating boundaries, the energy storage system charging and discharging intervals, and the multi-energy balance of electricity, heat, and gas, the optimal dispatch scheme is solved through the dynamic carbon emission measurement interval and wind power forecast interval. The scheme includes: Based on the topological structure of the integrated energy system, determine the operating characteristic parameters of each energy source in the integrated energy system; Based on the technical parameters of the energy supply equipment and the operating characteristic parameters of each energy source, a set of constraints is set, including the operating boundaries of the energy conversion equipment, the charging and discharging intervals of the energy storage system, and the multi-energy balance of electricity, heat and gas. When all constraints of the constraint set are met, the dynamic carbon emission measurement interval and wind power forecast interval are used as input to solve the economic dispatch model with the objective function as the core, and the optimal dispatch plan of the integrated energy system is obtained; The energy supply equipment operation boundaries include: wind power output constraints, cogeneration unit operation constraints, and gas boiler operation constraints; The charging and discharging range of the energy storage system includes: charging and discharging power limit range and energy storage capacity limit; The multi-energy balance of electricity, heat and gas includes: electric power balance, thermal power balance, gas power balance, electricity purchase limit and gas purchase limit.

10. An integrated energy system, characterized in that: include: Energy demand side; The energy supply side connected to the energy demand side includes wind turbines, cogeneration units, gas boilers, and external power grids; An energy dispatching platform connected to the energy supply end is used to execute the steps of the integrated energy system optimization dispatching method based on carbon emission factors and wind power forecasts as described in any one of claims 1 to 9.

Images