Green hydrogen flexible scheduling method and system considering wind and light output uncertainty

Abstract

The invention discloses a green hydrogen flexible scheduling method and system considering wind and light output uncertainty, and relates to the technical field of energy power, and the method comprises the steps: calculating a prediction deviation according to a historical wind and light output actual value and a prediction value; carrying out normal distribution fitting on the prediction deviation; correcting the standard deviation; reconstructing normal distribution and a probability density function of the prediction deviation; selecting a plurality of sampling points; calculating corrected wind and light output prediction; calculating the weight of each sampling point; calculating a production objective function value; according to the production objective function value and the weight of the corresponding sampling point, performing calculation to obtain multiple groups of objective function expectations under corrected wind and light output prediction; and optimizing the target function expectation, and taking the decision variable corresponding to the minimum target function expectation as an optimization result of the decision variable. According to the method, correction and uncertainty processing are carried out on wind and light output prediction, so that the influence of wind and light output uncertainty on operation of a scheduling system is reduced.

Description

Technical Field

[0001] The present invention relates to the field of energy and power technology, and in particular to a green hydrogen flexible scheduling method and system that takes into account the uncertainty of wind and solar power output. Background Art

[0002] "Green chemistry" is a trend for sustainable development of the chemical industry in the future. It uses "green electricity" produced by renewable energy sources such as wind power and photovoltaics to input into hydrogen production equipment such as electrolyzers to produce "green hydrogen" through water electrolysis. Hydrogen, as a clean energy carrier, can not only replace traditional fossil fuels and reduce greenhouse gas emissions, but also be used as a chemical raw material to promote the low-carbon transformation of the entire industry.

[0003] However, the output of renewable energy such as wind power and photovoltaics is greatly affected by natural factors such as wind speed and light intensity, and has strong inherent intermittency and uncertainty. The output fluctuates greatly, which brings great challenges to the subsequent production of "green hydrogen". The scheduling and control of hydrogen production equipment such as electrolyzers need to fully consider the volatility and uncertainty of wind and solar output forecasts in order to quickly respond to changes in power supply and ensure the continuous and stable production of hydrogen. Therefore, full consideration of the volatility and uncertainty of wind and solar output is a necessary condition to ensure the safe and stable production and economic and reliable operation of subsequent "green hydrogen". Summary of the invention

[0004] The present invention provides a green hydrogen flexible scheduling method and system taking into account the uncertainty of wind and solar power output, so as to overcome at least one technical problem existing in the prior art.

[0005] On the one hand, an embodiment of the present invention provides a green hydrogen flexible scheduling method considering the uncertainty of wind and solar power output, including:

[0006] Obtain the actual and predicted values ​​of wind and solar power output at corresponding moments in the historical period;

[0007] Calculating a prediction deviation based on the actual value and the predicted value;

[0008] Performing normal distribution fitting on the prediction deviation to obtain the mean and standard deviation of the prediction deviation;

[0009] Obtain the wind and solar power output forecast for the future predetermined time period provided by the manufacturer;

[0010] Correcting the standard deviation according to the time difference between the provision time of the wind and solar power output forecast provided by the manufacturer and the current time;

[0011] Reconstructing a normal distribution and a probability density function of the prediction deviation according to the mean and the corrected standard deviation;

[0012] Based on the reconstructed normal distribution, multiple sampling points are selected;

[0013] Calculate the sum of the wind and light power output predictions provided by the manufacturer and each of the said sampling points as the corresponding corrected wind and light power output prediction.

[0014] Calculate the weight of each of the said sampling points according to the probability density function and the sampling point position.

[0015] Establish the mathematical model and production objective function of the scheduling algorithm.

[0016] Substitute each of the said corrected wind and light power output predictions into the mathematical model to calculate the corresponding production objective function value.

[0017] Calculate the expected value of the objective function under multiple groups of corrected wind and light power output predictions according to the production objective function value and the weight of the corresponding sampling point.

[0018] Optimize the expected value of the objective function, and take the decision variable corresponding to the minimum value of the expected value of the objective function as the optimization result of the decision variable.

[0019] Optionally, the wind and light power output includes wind power output and photovoltaic power output; calculating the prediction deviation includes calculating the wind power output prediction deviation and the photovoltaic power output prediction deviation; wherein,

[0020] The wind power output prediction deviation is expressed as indicates the value-taking position, indicates the predicted value of the historical wind power output, indicates the actual value of the historical wind power output, ΔP wind,i indicates the deviation between the predicted value and the actual value of the wind power output.

[0021] The photovoltaic power output prediction deviation is expressed as indicates the value-taking position, indicates the predicted value of the historical photovoltaic power output, indicates the actual value of the historical photovoltaic power output, sol,i indicates the deviation between the predicted value and the actual value of the photovoltaic power output.

[0022] Optionally, the mean and standard deviation of the prediction deviation include the mean μ wind and standard deviation σ wind of the wind power output prediction deviation, and the mean μ sol and standard deviation σ sol of the photovoltaic power output prediction deviation;

[0023] Correct the standard deviation, including correcting the standard deviation σ wind of the wind power output prediction deviation and correcting the standard deviation σ sol of the photovoltaic power output prediction deviation; wherein,

[0024] The standard deviation of the wind power output forecast deviation is corrected, specifically:

[0025] The standard deviation correction coefficient of the wind power output forecast deviation is set as

[0026] and

[0027] They represent the upper and lower limits of the standard deviation correction coefficient of wind power output forecast deviation, Δt represents the time difference, and Δt max and Δt min Respectively represent the upper and lower limits of the time difference;

[0028] According to the standard deviation correction coefficient of the wind power output forecast deviation, the standard deviation of the wind power output forecast deviation after correction is calculated as σ wind,c =λ wind σ wind ;

[0029] The standard deviation of the photovoltaic output prediction deviation is corrected, specifically:

[0030] The standard deviation correction coefficient of the photovoltaic output prediction deviation is set to

[0031] and

[0032] They represent the upper and lower limits of the standard deviation correction coefficient of photovoltaic output prediction deviation respectively;

[0033] According to the standard deviation correction coefficient of the photovoltaic output prediction deviation, the standard deviation of the photovoltaic output prediction deviation after correction is calculated as σ sol,c =λ sol σ sol .

[0034] Optionally, the sampling points include wind power output prediction deviation sampling points and photovoltaic output prediction deviation sampling points;

[0035] The wind power output prediction deviation sampling point is expressed as The photovoltaic output prediction deviation sampling point is expressed as Among them, 2N-1 represents the number of deviation sampling points, x wind,j represents the wind power output prediction deviation of the jth sampling point, x sol,j Represents the photovoltaic output prediction deviation of the jth sampling point.

[0036] Optionally, the probability density function includes a probability density function of wind power output prediction deviation and a probability density function of photovoltaic output prediction deviation; wherein,

[0037] The probability density function of the wind power output forecast deviation is expressed as x wind represents the independent variable of wind power output forecast deviation;

[0038] The probability density function of the photovoltaic output prediction deviation is expressed as x sol Represents the independent variable of photovoltaic output prediction deviation.

[0039] Optionally, the weight of each sampling point is calculated according to the probability density function and the sampling point position, specifically:

[0040] According to the probability density function of the wind power output forecast deviation and the sampling point position, the weight of the wind power output forecast deviation sampling point is calculated as:

[0041] According to the probability density function of the photovoltaic output prediction deviation and the sampling point position, the weight of the photovoltaic output prediction deviation sampling point is calculated as:

[0042] Optionally, the objective function under the corrected wind and solar output forecast is expected to be expressed as Among them, Obj j represents the production objective function value,

[0043] Optionally, the objective function expectation is optimized, and the decision variable corresponding to the minimum objective function expectation is taken as the optimization result of the decision variable, specifically:

[0044] Obtaining the initial value and optimization interval of the decision variable;

[0045] Based on the initial value and the optimization interval, the objective function is expected to be optimized by Bayesian iteration, and the Bayesian optimization set is Obj des,p ; Wherein, p = 1, 2..., M, M represents the number of cycles;

[0046] Take the Bayesian optimization set Obj des,p The decision variable corresponding to the minimum value is taken as the optimization result of the decision variable.

[0047] On the other hand, the present invention also provides a green hydrogen flexible scheduling system considering the uncertainty of wind and solar power output, including:

[0048] The first acquisition module is used to obtain the actual value and predicted value of wind and solar power output at the corresponding time in the historical period;

[0049] A first calculation module, used for calculating a prediction deviation according to the actual value and the predicted value;

[0050] A fitting module, used to perform normal distribution fitting on the prediction deviation to obtain a mean and a standard deviation of the prediction deviation;

[0051] The second acquisition module is used to obtain the wind and solar power output forecast within a future predetermined time period provided by the manufacturer;

[0052] A correction module, used to correct the standard deviation according to the time difference between the provision time of the wind and solar power output forecast provided by the manufacturer and the current time;

[0053] A reconstruction module, used for reconstructing the normal distribution and probability density function of the prediction deviation according to the mean and the corrected standard deviation;

[0054] A selection module, used for selecting multiple sampling points based on the reconstructed normal distribution;

[0055] The second calculation module is used to calculate the sum of the wind and solar power output prediction provided by the manufacturer and each of the sampling points as the corresponding corrected wind and solar power output prediction;

[0056] A third calculation module, used to calculate the weight of each sampling point according to the probability density function and the sampling point position;

[0057] Establishing modules for establishing mathematical models of scheduling algorithms and production objective functions;

[0058] A fourth calculation module, used for substituting each of the corrected wind and solar power output predictions into the mathematical model to calculate a corresponding production objective function value;

[0059] A fifth calculation module, configured to calculate the expected objective functions under multiple groups of corrected wind and solar power output predictions according to the production objective function value and the weights of the corresponding sampling points;

[0060] The optimization module is used to optimize the expected value of the objective function and take the decision variables corresponding to the minimum expected value of the objective function as the optimization result of the decision variables.

[0061] Optionally, the mean and standard deviation of the forecast deviation include the mean μ of the wind power output forecast deviation wind and standard deviation σ wind , and the mean value μ of the photovoltaic output prediction deviation sol and standard deviation σ sol ;

[0062] The correction module is specifically used to: correct the standard deviation σ of the wind power output prediction deviation wind Correction and standard deviation of the photovoltaic output prediction deviation σ sol Make corrections; where

[0063] The standard deviation of the wind power output forecast deviation is corrected, specifically:

[0064] The standard deviation correction coefficient of the wind power output forecast deviation is set as

[0065] and

[0066] They represent the upper and lower limits of the standard deviation correction coefficient of wind power output forecast deviation, Δt represents the time difference, and Δt max and Δt min Respectively represent the upper and lower limits of the time difference;

[0067] According to the standard deviation correction coefficient of the wind power output forecast deviation, the standard deviation of the wind power output forecast deviation after correction is calculated as σ wind,c =λ wind σ wind ;

[0068] The standard deviation of the photovoltaic output prediction deviation is corrected, specifically:

[0069] The standard deviation correction coefficient of the photovoltaic output prediction deviation is set to

[0070] and

[0071] They represent the upper and lower limits of the standard deviation correction coefficient of photovoltaic output prediction deviation respectively;

[0072] According to the standard deviation correction coefficient of the photovoltaic output prediction deviation, the standard deviation of the photovoltaic output prediction deviation after correction is calculated as σ sol,c =λ sol σ sol .

[0073] The innovative features of the embodiments of the present invention include:

[0074] 1. In this embodiment, the uncertainty of wind and solar power output prediction is taken into account by reconstructing the normal distribution to obtain the uncertainty interval, thereby coping with the uncertainty prediction deviation of wind and solar power output and completing the correction of wind and solar power output prediction, which is one of the innovative points of the embodiment of the present invention.

[0075] 2. In this embodiment, by reconstructing the probability density function of the wind and solar power output prediction deviation, different probability weights are assigned to different deviation sampling points, and by optimizing the objective function expectation, the uncertainty processing of the wind and solar power output prediction is achieved, thereby reducing the impact of the wind and solar power output uncertainty on the operation of the scheduling system. This is one of the innovative points of the embodiment of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0077] Figure 1 A flow chart of a scheduling method provided by an embodiment of the present invention;

[0078] Figure 2 A flow chart for correcting the standard deviation of wind power output prediction deviation provided by an embodiment of the present invention;

[0079] Figure 3 A flow chart for correcting the standard deviation of photovoltaic output prediction deviation provided by an embodiment of the present invention;

[0080] Figure 4 A flow chart for optimizing the objective function expectation provided by an embodiment of the present invention;

[0081] Figure 5 A schematic diagram of the structure of a scheduling system provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0082] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0083] It should be noted that the terms "including" and "having" and any variations thereof in the embodiments of the present invention and the accompanying drawings are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device including a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products or devices.

[0084] The embodiment of the present invention discloses a green hydrogen flexible scheduling method considering the uncertainty of wind and solar power output. The following is a detailed description of each method.

[0085] Figure 1 For a flowchart of the scheduling method provided by an embodiment of the present invention, please refer to Figure 1 , the green hydrogen flexible scheduling method considering the uncertainty of wind and solar power output provided by the embodiment of the present invention includes:

[0086] Step 1: Obtain the actual value and predicted value of wind and solar power output at the corresponding time in the historical period;

[0087] Step 2: Calculate the forecast deviation based on the actual value and the forecast value;

[0088] Step 3: Fit the forecast deviation with a normal distribution to obtain the mean and standard deviation of the forecast deviation;

[0089] Step 4: Obtain the wind and solar power output forecast for the future predetermined time period provided by the manufacturer;

[0090] Step 5: Correct the standard deviation based on the time difference between the wind and solar power output forecast time provided by the manufacturer and the current time;

[0091] Step 6: Reconstruct the normal distribution and probability density function of the prediction deviation based on the mean and the corrected standard deviation;

[0092] Step 7: Based on the reconstructed normal distribution, select multiple sampling points;

[0093] Step 8: Calculate the sum of the wind and solar power output forecast provided by the manufacturer and each sampling point as the corresponding corrected wind and solar power output forecast;

[0094] Step 9: Calculate the weight of each sampling point based on the probability density function and the sampling point location;

[0095] Step 10: Establish the mathematical model of the scheduling algorithm and the production objective function;

[0096] Step 11: Substitute each corrected wind and solar output forecast into the mathematical model to calculate the corresponding production objective function value;

[0097] Step 12: According to the production objective function value and the weight of the corresponding sampling point, the objective function expectation under multiple groups of corrected wind and solar output forecasts is calculated;

[0098] Step 13: Optimize the expectation of the objective function and take the decision variable corresponding to the minimum expectation of the objective function as the optimization result of the decision variable.

[0099] Specifically, please refer to Figure 1The green hydrogen flexible scheduling method considering the uncertainty of wind and solar power output provided by the embodiment of the present invention first collects the actual value and predicted value of wind and solar power output in the past period of time through step 1. Wind and solar power output includes wind power output and photovoltaic output. Therefore, it is necessary to obtain the actual value of wind power output separately. Predicted value And the actual value of photovoltaic output Predicted value

[0100] After obtaining the above values, in step 2, the wind power output forecast deviation and photovoltaic output forecast deviation can be calculated according to the actual value and forecast value at the corresponding time. and actual value The deviation between them is the wind power output prediction deviation ΔP wind,i Therefore, the wind power output forecast deviation can be expressed as Here, i represents the value position, i = 1, 2, .... Similarly, the historical photovoltaic output forecast value and actual value The deviation between them is the photovoltaic output prediction deviation ΔP sol,i Therefore, the photovoltaic output prediction deviation can be expressed as Among them, i represents the value position, i=1,2,...

[0101] After obtaining the prediction deviation, in step 3, the prediction deviation is fitted with a normal distribution according to the time series. For example, the wind power output prediction deviation is fitted with a normal distribution to obtain the mean value μ of the wind power output prediction deviation. wind and standard deviation σ wind At this time, the probability density function of the predicted deviation of wind power output forecast value relative to the actual value is Among them, x wind Represents the independent variable of wind power output forecast deviation. The photovoltaic output forecast deviation is fitted with a normal distribution to obtain the mean value μ of the photovoltaic output forecast deviation sol and standard deviation σ sol At this time, the probability density function of the predicted deviation of the photovoltaic output forecast value relative to the actual value is Among them, x sol Represents the independent variable of photovoltaic output prediction deviation.

[0102] Obtain the wind and light power output predictions provided by the manufacturer for the future predetermined time period through Step 4, namely the wind power output prediction and the photovoltaic power output prediction. Usually, the manufacturer will regularly provide the wind and light power output predictions for a future period of time. The prediction accuracy is related to the time difference. The closer the time is, the higher the prediction accuracy. Conversely, the farther the time is, the lower the prediction accuracy. Therefore, in Step 5 of the present invention, according to the time difference between the provided time of the wind and light power output predictions provided by the manufacturer and the current time, the standard deviation σ wind of the wind power output prediction deviation and the standard deviation σ sol of the photovoltaic power output prediction deviation are corrected respectively, and the corrected standard deviation of the wind power output prediction deviation is σ wind,c , and the corrected standard deviation of the photovoltaic power output prediction deviation is σ so l ,c .

[0103] After correcting the standard deviation, in Step 6, using the mean value obtained by fitting in Step 3 and the standard deviation corrected in Step 5, the normal distribution and probability density function of the prediction deviation are reconstructed. When reconstructing, using the mean value μ wind of the wind power output prediction deviation and the corrected standard deviation σ wind,c of the wind power output prediction deviation, the reconstructed probability density function of the wind power output prediction deviation is x wind represents the independent variable of the wind power output prediction deviation. Using the mean value μ sol of the photovoltaic power output prediction deviation and the corrected standard deviation σ sol,c of the photovoltaic power output prediction deviation, the reconstructed probability density function of the photovoltaic power output prediction deviation is x sol represents the independent variable of the photovoltaic power output prediction deviation.

[0104] After reconstructing the normal distribution, in Step 7, a plurality of sampling points are respectively selected on the reconstructed normal distribution of the wind power output prediction deviation and the normal distribution of the photovoltaic power output prediction deviation to obtain the wind power output prediction deviation sampling points and the photovoltaic power output prediction deviation sampling points.

[0105] In this embodiment, when selecting the sampling points, in order to cover the complete normal distribution interval, the wind power output prediction deviation sampling points are expressed as The photovoltaic power output prediction deviation sampling points are expressed as where 2N - 1 represents the number of deviation sampling points, x wind,j represents the wind power output prediction deviation of the j-th sampling point, and x sol,j represents the photovoltaic power output prediction deviation of the j-th sampling point.

[0106] After obtaining the prediction deviation of each sampling point, in step 8, according to the predicted wind and light power provided by the manufacturer and the prediction deviation of each sampling point, the corrected predicted wind and light power corresponding to each sampling point can be calculated. In this embodiment, the sum of the predicted wind and light power provided by the manufacturer and the prediction deviation of each sampling point is used as the corrected predicted wind and light power of the corresponding sampling point.

[0107] Therefore, the corrected predicted wind power output can be obtained as Wherein, represents the predicted wind power output provided by the manufacturer, is the corrected predicted wind power output, that is, the predicted wind power output finally participating in the scheduling algorithm. Similarly, the corrected predicted photovoltaic power output can be obtained as Wherein, represents the predicted photovoltaic power output provided by the manufacturer, is the corrected predicted photovoltaic power output, that is, the predicted photovoltaic power output finally participating in the scheduling algorithm.

[0108] After obtaining the probability density function and sampling points, in step 9, the weights of each sampling point can be calculated. The weight of the sampling point of the predicted wind power output deviation is

[0109] The weight of the sampling point of the predicted photovoltaic power output deviation is

[0110] After obtaining the above data, they can be substituted into the model for calculation. Therefore, through step 10 of the present invention, a mathematical model and a production objective function of the scheduling algorithm are established. Here, the mathematical model and the production objective function can be the mathematical model and the production objective function of the green hydrogen to ammonia system, or they can also be the mathematical model and the production objective function of the green hydrogen to methanol system. Specifically, according to the actual situation, the corresponding mathematical model and production objective function can be established, and the present invention does not make specific limitations on this.

[0111] Then in step 11, each corrected predicted wind and light power output ( and ) is substituted into the mathematical model to calculate the production objective function, and the production objective function value Obj j can be obtained. In step 12, according to the production objective function value Obj j and the weights of the corresponding sampling points (W wind (x wind,j ) and W sol (x sol,j ))), the expected value of the objective function under multiple groups of corrected predicted wind and light power outputs can be calculated as Wherein,

[0112] After obtaining the expected objective function, in step 13, to minimize Obj des As the optimization target, the Bayesian iterative optimization method is used to optimize the objective function expectation. des When it is minimum, the corresponding decision variable is the optimization result of the decision variable.

[0113] The green hydrogen flexible scheduling method considering the uncertainty of wind and solar power output provided by the present invention considers the uncertainty of wind and solar power output prediction by reconstructing the normal distribution, and obtains the uncertainty interval, thereby coping with the uncertainty prediction deviation of wind and solar power output and completing the correction of wind and solar power output prediction. In addition, the present invention reconstructs the probability density function, assigns different probability weights to different deviation sampling points in combination with the sampling point position, and optimizes the expectation of the objective function to achieve uncertainty processing of wind and solar power output prediction, thereby reducing the impact of wind and solar power output uncertainty on the operation of the scheduling system.

[0114] Optionally, Figure 2 A flow chart of correcting the standard deviation of wind power output prediction deviation provided by an embodiment of the present invention, Figure 3 A flow chart of correcting the standard deviation of photovoltaic power output prediction deviation provided by an embodiment of the present invention. Correcting the standard deviation includes correcting the standard deviation σ of wind power output prediction deviation. wind Correction and standard deviation of photovoltaic output prediction deviation σ sol For calibration, refer to Figure 2 , the standard deviation of the wind power output forecast deviation is corrected, specifically: Step 511, set the standard deviation correction coefficient of the wind power output forecast deviation to and They represent the upper and lower limits of the standard deviation correction coefficient of wind power output forecast deviation, Δt represents the time difference, and Δt max and Δt min Respectively represent the upper and lower limits of the time difference; Step 512, according to the standard deviation correction coefficient of the wind power output prediction deviation, the standard deviation of the corrected wind power output prediction deviation is calculated as σ wind,c =λ wind σ wind .

[0115] Please refer to Figure 3 , the standard deviation of the photovoltaic output prediction deviation is corrected, specifically: Step 521, set the standard deviation correction coefficient of the photovoltaic output prediction deviation to and They represent the upper and lower limits of the standard deviation correction coefficient of the photovoltaic output prediction deviation respectively; Step 522, according to the standard deviation correction coefficient of the photovoltaic output prediction deviation, the standard deviation of the corrected photovoltaic output prediction deviation is calculated as σ sol,c=λ sol σ sol .

[0116] Specifically, the standard deviation is corrected, including the standard deviation of wind power output forecast deviation σ wind Correction and standard deviation of photovoltaic output prediction deviation σ sol Please refer to Figure 2 When correcting the standard deviation of the wind power output prediction deviation, firstly, in step 511, the standard deviation correction coefficient of the wind power output prediction deviation is set. In this embodiment, according to the time difference Δt and the upper and lower limits of the time difference (Δt max and Δt min ), the standard deviation correction coefficient of wind power output forecast deviation is set as in, and They represent the upper and lower limits of the standard deviation correction coefficient of the wind power output forecast deviation. It should be noted that the time difference Δt here is the time difference between the provision time of the wind and solar power output forecast provided by the manufacturer and the current time.

[0117] After the correction coefficient is obtained, in step 512, the standard deviation correction coefficient λ of the wind power output prediction deviation is calculated. wind And the standard deviation of wind power output forecast deviation σ wind , the standard deviation of the wind power output forecast deviation after correction can be calculated as σ wind,c =λ wind σ wind .

[0118] Please refer to Figure 3 When correcting the standard deviation of the photovoltaic output prediction deviation, firstly, in step 521, the standard deviation correction coefficient of the photovoltaic output prediction deviation is set. Similarly, according to the time difference Δt and the upper and lower limits of the time difference (Δt max and Δt min ), the standard deviation correction coefficient of photovoltaic output prediction deviation is set as in, and They respectively represent the upper and lower limits of the standard deviation correction coefficient of the photovoltaic output prediction deviation.

[0119] After the correction coefficient is obtained, in step 522, the correction coefficient λ is calculated based on the standard deviation of the photovoltaic output prediction deviation. sol And the standard deviation of photovoltaic output prediction deviation σ sol , the standard deviation of the corrected photovoltaic output prediction deviation can be calculated as σ sol,c =λ sol σ sol .

[0120] Optionally, Figure 4 For a flow chart of optimizing the objective function expectation provided by an embodiment of the present invention, please refer to Figure 1 and Figure 4 In step 13, the objective function expectation is optimized, and the decision variable corresponding to the minimum objective function expectation is taken as the optimization result of the decision variable, specifically: step 131, obtain the initial value and optimization interval of the decision variable; step 132, based on the initial value and optimization interval, perform Bayesian iterative optimization on the objective function expectation, and obtain the Bayesian optimization set Obj des,p ; Where p = 1, 2..., M, M represents the number of cycles; Step 133, take the Bayesian optimization set Obj des,p The decision variable corresponding to the minimum value is taken as the optimization result of the decision variable.

[0121] Specifically, please refer to Figure 4 When optimizing the objective function expectation, in step 131, the initial value and optimization interval of the decision variable are determined according to the system state and the constraint conditions. In step 132, the objective function expectation is iteratively optimized according to the initial value and optimization interval of the decision variable. Since a large number of random deviations of wind power output prediction and photovoltaic output prediction are generated each time, the amount of calculation for each optimization is large. Therefore, this embodiment adopts the Bayesian optimization algorithm with higher optimization efficiency to perform parameter optimization. The specific algorithm of Bayes can refer to the prior art and will not be described here.

[0122] Assuming the number of loops is M, the Bayesian optimization set obtained at the end of the loop is Obj des,p , where p = 1, 2, ..., M. In step 133, in the Bayesian optimization set Obj des,p Find the minimum value in , and take the decision variable corresponding to the minimum value as the optimization result of the decision variable.

[0123] Based on the same inventive concept, the present invention also provides a green hydrogen flexible scheduling system that takes into account the uncertainty of wind and solar power output. Figure 5 For a schematic diagram of a scheduling system provided by an embodiment of the present invention, please refer to Figure 5 The green hydrogen flexible dispatching system 100 considering the uncertainty of wind and solar power output provided by the present invention includes:

[0124] The first acquisition module is used to obtain the actual value and predicted value of wind and solar power output at the corresponding time in the historical period;

[0125] A first calculation module, used for calculating the prediction deviation according to the actual value and the predicted value;

[0126] A fitting module is used to fit the prediction deviation to a normal distribution and obtain the mean and standard deviation of the prediction deviation;

[0127] The second acquisition module is used to obtain the wind and solar power output forecast within a future predetermined time period provided by the manufacturer;

[0128] A correction module is used to correct the standard deviation according to the time difference between the wind and solar power output forecast provided by the manufacturer and the current time;

[0129] A reconstruction module, used to reconstruct the normal distribution and probability density function of the prediction deviation based on the mean and the corrected standard deviation;

[0130] A selection module, used for selecting multiple sampling points based on the reconstructed normal distribution;

[0131] The second calculation module is used to calculate the sum of the wind and solar power output forecast provided by the manufacturer and each sampling point as the corresponding corrected wind and solar power output forecast;

[0132] A third calculation module is used to calculate the weight of each sampling point according to the probability density function and the sampling point position;

[0133] Establishing modules for establishing mathematical models of scheduling algorithms and production objective functions;

[0134] The fourth calculation module is used to substitute each corrected wind and solar power output forecast into the mathematical model to calculate the corresponding production objective function value;

[0135] A fifth calculation module is used to calculate the expected objective functions under multiple groups of corrected wind and solar power output forecasts according to the production objective function value and the weights of the corresponding sampling points;

[0136] The optimization module is used to optimize the expectation of the objective function and take the decision variable corresponding to the minimum expectation of the objective function as the optimization result of the decision variable.

[0137] Specifically, please refer to Figure 5 The green hydrogen flexible dispatching system 100 considering the uncertainty of wind and solar power output provided by the embodiment of the present invention first collects the actual value and predicted value of wind and solar power output in the past period of time through the first acquisition module. Wind and solar power output includes wind power output and photovoltaic output. Therefore, it is necessary to obtain the actual value of wind power output separately. Predicted value And the actual value of photovoltaic output Predicted value

[0138] After obtaining the above values, the first calculation module can calculate the wind power output prediction deviation and the photovoltaic output prediction deviation respectively according to the actual value and the predicted value at the corresponding time. and actual value The deviation between them is the wind power output prediction deviation ΔPwind,i Therefore, the wind power output forecast deviation can be expressed as Here, i represents the value position, i = 1, 2, .... Similarly, the historical photovoltaic output forecast value and actual value The deviation between them is the photovoltaic output prediction deviation ΔP sol,i Therefore, the photovoltaic output prediction deviation can be expressed as Among them, i represents the value position, i=1,2,...

[0139] After obtaining the prediction deviation, the fitting module performs normal distribution fitting on the prediction deviation according to the time series. For example, the wind power output prediction deviation is fitted with a normal distribution to obtain the mean value μ of the wind power output prediction deviation. wind and standard deviation σ wind At this time, the probability density function of the predicted deviation of wind power output forecast value relative to the actual value is Among them, x wind Represents the independent variable of wind power output forecast deviation. The photovoltaic output forecast deviation is fitted with a normal distribution to obtain the mean value μ of the photovoltaic output forecast deviation sol and standard deviation σ sol At this time, the probability density function of the predicted deviation of the photovoltaic output forecast value relative to the actual value is Among them, x sol Represents the independent variable of photovoltaic output prediction deviation.

[0140] The second acquisition module collects the wind and solar power output forecasts provided by the manufacturer within the future scheduled time period, which are wind power output forecasts and photovoltaic output forecasts. Usually, the manufacturer will regularly provide wind and solar power output forecasts for a period of time in the future. The accuracy of the forecast is related to the time difference. The closer the time, the higher the forecast accuracy. Conversely, the farther the time, the lower the forecast accuracy. Therefore, the correction module calculates the standard deviation σ of the wind power output forecast deviation based on the time difference between the wind and solar power output forecast provided by the manufacturer and the current time. wind and the standard deviation of photovoltaic output forecast deviation σ sol After correction, the standard deviation of wind power output forecast deviation after correction is σ wind,c The standard deviation of the corrected photovoltaic output prediction deviation is σ sol,c .

[0141] After correcting the standard deviation, the reconstruction module uses the fitted mean and the corrected standard deviation to reconstruct the normal distribution and probability density function of the forecast deviation. During reconstruction, the mean μ of the wind power output forecast deviation is used. wind and the standard deviation of the wind power output forecast deviation after correction σ wind,c , the probability density function of wind power output forecast deviation is reconstructed as xwind represents the independent variable of wind power output forecast deviation. The mean value μ of photovoltaic output forecast deviation is used sol and the standard deviation of the corrected photovoltaic output prediction deviation σ sol,c , the probability density function of the photovoltaic output prediction deviation is reconstructed as x sol Represents the independent variable of photovoltaic output prediction deviation.

[0142] After reconstructing the normal distribution, the selection module selects multiple sampling points on the reconstructed normal distribution of wind power output prediction deviation and the normal distribution of photovoltaic output prediction deviation to obtain wind power output prediction deviation sampling points and photovoltaic output prediction deviation sampling points.

[0143] In this embodiment, when selecting sampling points, in order to cover the complete normal distribution interval, the wind power output prediction deviation sampling points are expressed as The photovoltaic output prediction deviation sampling point is expressed as Among them, 2N-1 represents the number of deviation sampling points, x wind,j represents the wind power output prediction deviation of the jth sampling point, x sol,j Represents the photovoltaic output prediction deviation of the jth sampling point.

[0144] After obtaining the prediction deviation of each sampling point, the second calculation module can calculate the corrected wind-solar output prediction corresponding to each sampling point based on the wind-solar output prediction provided by the manufacturer and the prediction deviation of each sampling point. In this embodiment, the sum of the wind-solar output prediction provided by the manufacturer and the prediction deviation of each sampling point is used as the corrected wind-solar output prediction of the corresponding sampling point.

[0145] Therefore, the corrected wind power output forecast can be obtained as in, represents the wind power output forecast provided by the manufacturer, is the forecast of wind power output after correction, which is also the forecast of wind power output that finally participates in the dispatching algorithm. Similarly, the forecast of photovoltaic output after correction can be obtained as in, It indicates the photovoltaic output forecast provided by the manufacturer. It is the corrected photovoltaic output forecast, that is, the photovoltaic output forecast that finally participates in the scheduling algorithm.

[0146] After obtaining the probability density function and sampling points, the weights of the corresponding sampling points are calculated, and the weights of the wind power output prediction deviation sampling points are obtained as follows:

[0147] The weight of the photovoltaic output prediction deviation sampling point is

[0148] After obtaining the above data, it is necessary to substitute it into the model for calculation. Therefore, the present invention establishes a mathematical model of the scheduling algorithm and a production objective function through a building module. Here, the mathematical model and the production objective function can be the mathematical model and the production objective function of the green hydrogen ammonia production system, or they can also be the mathematical model and the production objective function of the green hydrogen methanol production system. Specifically, corresponding mathematical models and production objective functions can be established according to the actual situation, and the present invention does not make specific limitations on this.

[0149] Then, the fourth calculation module substitutes each corrected wind and solar power output prediction ( and ) into the mathematical model to calculate the production objective function, and the production objective function value Obj under the group of wind and solar power output predictions can be obtained. j . The fifth calculation module can calculate the expected value of the objective function under multiple groups of corrected wind and solar power output predictions according to the production objective function value Obj j and the weights of the corresponding sampling points (W wind (x wind,j ) and W sol (x sol,j )) as where,

[0150] After obtaining the expected value of the objective function, the optimization module takes minimizing Obj des as the optimization objective, and uses the Bayesian iterative optimization method to optimize the expected value of the objective function. When the expected value of the objective function Obj des is the smallest, the corresponding decision variable is the optimization result of the decision variable.

[0151] The green hydrogen flexible scheduling system considering the uncertainty of wind and solar power output provided by the present invention considers the uncertainty of wind and solar power output prediction by reconstructing the normal distribution to obtain the uncertainty interval, so as to cope with the uncertainty prediction deviation of wind and solar power output and complete the correction of wind and solar power output prediction. In addition, the present invention reconstructs the probability density function, assigns different probability weights to different deviation sampling points, and realizes the uncertainty processing of wind and solar power output prediction by optimizing the expected value of the objective function, reducing the impact of wind and solar power output uncertainty on the operation of the scheduling system.

[0152] Optionally, the correction module corrects the standard deviation, including correcting the standard deviation σ wind of the wind power output prediction deviation and correcting the standard deviation σ sol of the photovoltaic power output prediction deviation; the correction of the standard deviation of the wind power output prediction deviation is specifically: setting the standard deviation correction coefficient of the wind power output prediction deviation as and They represent the upper and lower limits of the standard deviation correction coefficient of wind power output forecast deviation, Δt represents the time difference, and Δt max and Δt min Respectively represent the upper and lower limits of the time difference; according to the standard deviation correction coefficient of wind power output forecast deviation, the standard deviation of wind power output forecast deviation after correction is calculated as σ wind,c =λ wind σ wind .

[0153] Correct the standard deviation of the photovoltaic output prediction deviation. Specifically, set the standard deviation correction coefficient of the photovoltaic output prediction deviation to and They represent the upper and lower limits of the standard deviation correction coefficient of the photovoltaic output prediction deviation respectively; according to the standard deviation correction coefficient of the photovoltaic output prediction deviation, the standard deviation of the photovoltaic output prediction deviation after correction is calculated as σ sol,c =λ sol σ sol .

[0154] Specifically, the correction module corrects the standard deviation, including the standard deviation σ of the wind power output prediction deviation. wind Correction and standard deviation of photovoltaic output prediction deviation σ sol When correcting the standard deviation of the wind power output forecast deviation, the correction module is first used to set the standard deviation correction coefficient of the wind power output forecast deviation. In this embodiment, the correction coefficient is calculated based on the time difference Δt and the upper and lower limits of the time difference (Δt max and Δt min ), the standard deviation correction coefficient of wind power output forecast deviation is set as in, and They represent the upper and lower limits of the standard deviation correction coefficient of the wind power output forecast deviation. It should be noted that the time difference Δt here is the time difference between the provision time of the wind and solar power output forecast provided by the manufacturer and the current time.

[0155] After obtaining the correction coefficient, the standard deviation correction coefficient λ of the wind power output forecast deviation is used wind And the standard deviation of wind power output forecast deviation σ wind , the standard deviation of the corrected wind power output forecast deviation can be calculated as σ wind,c =λ wind σ wind .

[0156] When correcting the standard deviation of the photovoltaic output prediction deviation, the correction module is used to set the standard deviation correction coefficient of the photovoltaic output prediction deviation. Similarly, according to the time difference Δt and the upper and lower limits of the time difference (Δt max and Δt min), the standard deviation correction coefficient of photovoltaic output prediction deviation is set as in, and They respectively represent the upper and lower limits of the standard deviation correction coefficient of the photovoltaic output prediction deviation.

[0157] After obtaining the correction coefficient, the correction coefficient λ is calculated based on the standard deviation of the photovoltaic output prediction deviation. sol And the standard deviation of photovoltaic output prediction deviation σ sol , the standard deviation of the corrected photovoltaic output prediction deviation can be calculated as σ sol,c =λ sol σ sol .

[0158] Those skilled in the art can understand that the accompanying drawings are only schematic diagrams of an embodiment, and the modules or processes in the accompanying drawings are not necessarily required to implement the present invention.

[0159] Those skilled in the art can understand that the modules in the device in the embodiment can be distributed in the device in the embodiment according to the description of the embodiment, or can be changed accordingly and located in one or more devices different from the embodiment. The modules in the above embodiment can be combined into one module, or can be further divided into multiple sub-modules.

[0160] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A green hydrogen flexible scheduling method considering the uncertainty of wind and solar power output, characterized in that: include: Obtain the actual and predicted values ​​of wind and solar power output at corresponding moments in the historical period; Calculating a prediction deviation based on the actual value and the predicted value; Performing normal distribution fitting on the prediction deviation to obtain the mean and standard deviation of the prediction deviation; Obtain the wind and solar power output forecast for the future predetermined time period provided by the manufacturer; Correcting the standard deviation according to the time difference between the provision time of the wind and solar power output forecast provided by the manufacturer and the current time; Reconstructing a normal distribution and a probability density function of the prediction deviation according to the mean and the corrected standard deviation; Based on the reconstructed normal distribution, multiple sampling points are selected; Calculate the sum of the wind and solar power output forecast provided by the manufacturer and each of the sampling points as the corresponding corrected wind and solar power output forecast; Calculate the weight of each sampling point according to the probability density function and the sampling point position; Establish mathematical models of scheduling algorithms and production objective functions; Substituting each of the corrected wind and solar output predictions into the mathematical model to calculate the corresponding production objective function value; According to the production objective function value and the weight of the corresponding sampling point, the objective function expectation under the multiple groups of corrected wind and solar output prediction is calculated; The objective function expectation is optimized, and the decision variable corresponding to the minimum objective function expectation is taken as the optimization result of the decision variable.

2. The green hydrogen flexible scheduling method considering the uncertainty of wind and solar power output according to claim 1 is characterized in that: The wind and solar power output includes wind power output and photovoltaic power output; the calculation of prediction deviation includes the calculation of wind power output prediction deviation and photovoltaic output prediction deviation; among which, The wind power output forecast deviation is expressed as Indicates the value location, represents the predicted value of historical wind power output, Indicates the actual value of historical wind power output, ΔP wind,i Indicates the deviation between the predicted value and the actual value of wind power output; Photovoltaic output prediction deviation is expressed as Indicates the value location, represents the predicted value of historical photovoltaic output, Indicates the actual value of historical photovoltaic output, ΔP sol,i Indicates the deviation between the predicted photovoltaic output and the actual value.

3. The green hydrogen flexible scheduling method considering the uncertainty of wind and solar power output according to claim 2 is characterized in that: The mean and standard deviation of the forecast deviation, including the mean μ of the wind power output forecast deviation wind and standard deviation σ wind , and the mean value μ of the photovoltaic output prediction deviation sol and standard deviation σ sol ; The standard deviation is corrected, including the standard deviation σ of the wind power output forecast deviation wind Correction and standard deviation of the photovoltaic output prediction deviation σ sol Make corrections; where The standard deviation of the wind power output forecast deviation is corrected, specifically: The standard deviation correction coefficient of the wind power output forecast deviation is set as and They represent the upper and lower limits of the standard deviation correction coefficient of wind power output forecast deviation, Δt represents the time difference, Δt max and Δt min Respectively represent the upper and lower limits of the time difference; According to the standard deviation correction coefficient of the wind power output forecast deviation, the standard deviation of the wind power output forecast deviation after correction is calculated as σ wind,c =λ wind σ wind ; The standard deviation of the photovoltaic output prediction deviation is corrected, specifically: The standard deviation correction coefficient of the photovoltaic output prediction deviation is set to and They represent the upper and lower limits of the standard deviation correction coefficient of photovoltaic output prediction deviation respectively; According to the standard deviation correction coefficient of the photovoltaic output prediction deviation, the standard deviation of the photovoltaic output prediction deviation after correction is calculated as σ sol,c =λ sol σ sol .

4. The green hydrogen flexible scheduling method considering the uncertainty of wind and solar power output according to claim 3 is characterized in that: The sampling points include wind power output prediction deviation sampling points and photovoltaic output prediction deviation sampling points; The wind power output prediction deviation sampling point is expressed as The photovoltaic output prediction deviation sampling point is expressed as Among them, 2N-1 represents the number of deviation sampling points, x wind,j represents the wind power output prediction deviation of the jth sampling point, x sol,j Represents the photovoltaic output prediction deviation of the jth sampling point.

5. The green hydrogen flexible scheduling method considering the uncertainty of wind and solar power output according to claim 4 is characterized in that: The probability density function includes the probability density function of wind power output prediction deviation and the probability density function of photovoltaic output prediction deviation; wherein, The probability density function of the wind power output forecast deviation is expressed as x wind represents the independent variable of wind power output forecast deviation; The probability density function of the photovoltaic output prediction deviation is expressed as x sol Represents the independent variable of photovoltaic output prediction deviation.

6. The green hydrogen flexible scheduling method considering the uncertainty of wind and solar power output according to claim 5 is characterized in that: According to the probability density function and the sampling point position, the weight of each sampling point is calculated, specifically: According to the probability density function of the wind power output forecast deviation and the sampling point position, the weight of the wind power output forecast deviation sampling point is calculated as: According to the probability density function of the photovoltaic output prediction deviation and the sampling point position, the weight of the photovoltaic output prediction deviation sampling point is calculated as:

7. The green hydrogen flexible scheduling method considering the uncertainty of wind and solar power output according to claim 6 is characterized in that: The expected objective function of the corrected wind and solar power output forecast is expressed as Among them, Obj j represents the production objective function value, 8. The green hydrogen flexible scheduling method considering the uncertainty of wind and solar power output according to claim 7 is characterized in that: The objective function expectation is optimized, and the decision variable corresponding to the minimum objective function expectation is taken as the optimization result of the decision variable, which is specifically: Obtaining the initial value and optimization interval of the decision variable; Based on the initial value and the optimization interval, the objective function is expected to be optimized by Bayesian iteration, and the Bayesian optimization set is Obj des,p ; Where, p = 1, 2, ..., M, M represents the number of cycles; Take the Bayesian optimization set Obj des,p The decision variable corresponding to the minimum value is taken as the optimization result of the decision variable.

9. A green hydrogen flexible dispatching system considering the uncertainty of wind and solar power output, characterized in that: include: The first acquisition module is used to obtain the actual value and predicted value of wind and solar power output at the corresponding time in the historical period; A first calculation module, used for calculating a prediction deviation according to the actual value and the predicted value; A fitting module, used to perform normal distribution fitting on the prediction deviation to obtain a mean and a standard deviation of the prediction deviation; The second acquisition module is used to obtain the wind and solar power output forecast within a future predetermined time period provided by the manufacturer; A correction module, used to correct the standard deviation according to the time difference between the provision time of the wind and solar power output forecast provided by the manufacturer and the current time; A reconstruction module, used for reconstructing the normal distribution and probability density function of the prediction deviation according to the mean and the corrected standard deviation; A selection module, used for selecting multiple sampling points based on the reconstructed normal distribution; The second calculation module is used to calculate the sum of the wind and solar power output prediction provided by the manufacturer and each of the sampling points as the corresponding corrected wind and solar power output prediction; A third calculation module, used to calculate the weight of each sampling point according to the probability density function and the sampling point position; Establishing modules for establishing mathematical models of scheduling algorithms and production objective functions; A fourth calculation module, used for substituting each of the corrected wind and solar power output predictions into the mathematical model to calculate a corresponding production objective function value; A fifth calculation module, configured to calculate the expected objective functions under multiple groups of corrected wind and solar power output predictions according to the production objective function value and the weights of the corresponding sampling points; The optimization module is used to optimize the expected value of the objective function and take the decision variables corresponding to the minimum expected value of the objective function as the optimization result of the decision variables.

10. The green hydrogen flexible dispatching system considering the uncertainty of wind and solar power output according to claim 9 is characterized in that: The mean and standard deviation of the forecast deviation, including the mean μ of the wind power output forecast deviation wind and standard deviation σ wind , and the mean value μ of the photovoltaic output prediction deviation sol and standard deviation σ sol ; The correction module is specifically used to: correct the standard deviation σ of the wind power output prediction deviation wind Correction and standard deviation of the photovoltaic output prediction deviation σ sol Make corrections; where The standard deviation of the wind power output forecast deviation is corrected, specifically: The standard deviation correction coefficient of the wind power output forecast deviation is set as and They represent the upper and lower limits of the standard deviation correction coefficient of wind power output forecast deviation, Δt represents the time difference, Δt max and Δt min Respectively represent the upper and lower limits of the time difference; According to the standard deviation correction coefficient of the wind power output forecast deviation, the standard deviation of the wind power output forecast deviation after correction is calculated as σ wind,c =λ wind σ wind ; The standard deviation of the photovoltaic output prediction deviation is corrected, specifically: The standard deviation correction coefficient of the photovoltaic output prediction deviation is set to and They represent the upper and lower limits of the standard deviation correction coefficient of photovoltaic output prediction deviation respectively; According to the standard deviation correction coefficient of the photovoltaic output prediction deviation, the standard deviation of the photovoltaic output prediction deviation after correction is calculated as σ sol,c =λ sol σ sol .

Images