A renewable energy benefit risk analysis method and system based on environmental data

By acquiring and evaluating the economic and environmental information of the target location, combining numerical weather forecast and Copula model, analyzing the relationship between renewable energy output and transmission power, the high cost, high technical threshold and environmental analysis of renewable energy benefits in the existing technology is solved. The problem of changing factors has achieved efficient and accurate benefits and risk assessment and promoted sustainable development.

CN118365124BActive Publication Date: 2024-12-03NORTH CHINA ELECTRIC POWER UNIV +2
View PDF 0 Cites 2 Cited by

Patent Information

Application Number
CN202410507616.1
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-04-25
Publication Date
2024-12-03
Estimated Expiration
2044-04-25

AI Technical Summary

Technical Problem

The existing renewable energy benefit analysis methods have problems such as high data collection and processing costs, high technical thresholds, data quality affects analysis results, and changing environmental factors, resulting in low credibility in risk analysis.

Method used

By obtaining economic and environmental information of the target location, using numerical weather forecasting methods for climate assessment, combined with Shannon entropy, non-noise ratio, renewable energy enrichment, renewable energy matching, extreme event value and topographic limitation, the Copula model is adopted Analyze the relationship between renewable energy output and transmission power and build a planning model to evaluate renewable energy benefits risks.

Benefits of technology

Real-time and accurate renewable energy project benefits and risk assessments have been achieved, the analysis costs have been reduced, the analysis is improved integrity and flexibility, and the sustainable development has been promoted.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118365124B_ABST
    Figure CN118365124B_ABST
Patent Text Reader

Abstract

The present invention provides a renewable energy benefit risk analysis method and system based on environmental data, which belongs to the field of risk analysis and solves the problem of low credibility of risk analysis. The method is specifically as follows: Step 1: Obtain basic information of a target location; Construct a planning model; Step 2: Use a numerical weather forecast method to perform climate assessment; Define and calculate evaluation indicators of an information value quantitative evaluation system; Define and calculate specific values ​​of evaluation indicators based on basic information; Step 3: Determine the weights of evaluation indicators and calculate composite information value indicators; Analyze the relationship between renewable energy output and transmission power; Obtain analysis results; Step 4: Summarize planning models and analysis results; Obtain missing values ​​of constraint functions and complete analysis of renewable energy benefit risks. The present invention analyzes the renewable energy benefits of a target location, and adjusts and optimizes them according to different project requirements and environmental changes to achieve accurate evaluation of benefits and risks.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention discloses a renewable energy benefit risk analysis method and system based on environmental data, and relates to the field of risk analysis. Background Art

[0002] The existing analysis methods or systems for renewable energy benefits have the following deficiencies:

[0003] 1. High cost of data collection and processing: Analysis methods based on environmental data require a large amount of environmental data as support, which involves multiple links such as data collection, processing and analysis, and may require a lot of manpower, material and financial resources; for some projects with limited resources, this may be a considerable burden.

[0004] 2. High technical threshold: This method requires the use of advanced data analysis technology and algorithms, and requires high professional knowledge and skills of analysts; if the project team lacks relevant talent and technical support, it may affect the accuracy and effectiveness of the analysis.

[0005] 3. Data quality affects analysis results: The accuracy and completeness of environmental data have a crucial impact on the analysis results; if there are quality problems in the data, such as missing, incorrect or inconsistent data, it will lead to deviations in the analysis results and even mislead the decision.

[0006] 4. Environmental factors are variable: Environmental factors are complex and variable, such as climate change and natural disasters. Changes in these factors may have a significant impact on the benefits and risks of renewable energy projects; therefore, analysis methods based on environmental data need to be constantly updated and adjusted to adapt to the challenges brought about by environmental changes. Summary of the invention

[0007] In view of the shortcomings of the prior art, the purpose of the present invention is to provide a renewable energy benefit risk analysis method and system based on environmental data, aiming to solve the problem of low credibility of risk analysis.

[0008] In order to achieve the above object, the present invention is implemented through the following technical solutions: A renewable energy benefit risk analysis method based on environmental data includes:

[0009] Step 1: Obtain economic and environmental information of the target site to obtain basic information; based on the wide-area planning process, integrate construction capacity, ecological benefits and penalty factors related to extreme events; integrate grid security, renewable energy utilization and external transmission channel utilization; build a planning model to complete the primary processing of basic information;

[0010] Step 2: Use numerical weather forecasting methods to conduct climate assessment; use Shannon entropy, non-noise ratio, renewable energy abundance, renewable energy matching, extreme event value and terrain restrictions as evaluation indicators of the information value quantitative evaluation system; define and calculate the specific values ​​of the evaluation indicators based on the basic information, and complete the secondary processing of the basic information;

[0011] Step 3: Determine the weights of the evaluation indicators, calculate the composite information value indicators, and measure the information benefits; use the Copula model to analyze the relationship between renewable energy output and transmission power; complete the three-step processing of basic information;

[0012] Step 4: Summarize the analysis and processing results of steps 1 to 3; obtain the missing values ​​of the constraint function and complete the analysis of renewable energy benefit risks.

[0013] Furthermore, the specific steps of step 1 are as follows:

[0014] Step 11: The objective function of constructing the planning model is as follows:

[0015]

[0016] Formula (a1) where Φ represents a set consisting of the selected wind turbine generator sets, photovoltaic generator sets, energy storage devices, thermal generator sets, and various random scenarios that need to be considered;

[0017] denote the discounted value of the investment costs associated with wind turbine installation, photovoltaic power generation, energy storage equipment, and thermal power generation units, respectively; is a binary variable, which is used as a decision factor to determine whether to invest;

[0018] γ z The probability distribution corresponding to scenario z; represents the operating cost within this specific scenario; σ D It represents the differences between ecological gas regulation, climate regulation, water conservation, soil shaping and protection and the corresponding economic indicators of the target site;

[0019] Represent the construction capacity of wind power and photovoltaic power generation units respectively; the last item is regarded as a penalty item, which will impose penalties on extreme events; γ e,t Takes a binary value, indicating whether an extreme event occurs or not. Represents the annualized loss due to extreme events;

[0020] Step 12: The constraints are as follows:

[0021]

[0022]

[0023]

[0024]

[0025]

[0026]

[0027]

[0028]

[0029]

[0030] W SOC,min ≤W t SOC ≤W SOC,max (a11);

[0031]

[0032]

[0033] Step 13: Meaning of the parameters in formulas a2 to a13:

[0034] in, represents the power generation of the thermal power generating unit at time t, represents the power generation of the photovoltaic unit at time t, represents the power generated by the wind turbine at time t, represents the power demand of the load at time t;

[0035] Represent the charging power and discharging power of the energy storage device respectively;

[0036] The minimum and maximum active power output of each node of the thermal power generating unit is expressed as P G,min and P G,max , in kilowatts (kW); is a 0-1 variable indicating the start and stop of the thermal power generating unit at each node;

[0037] σ GIV Indicates the upper limit of wind and solar power curtailment; Respectively represent the energy storage state at time t and t-1, η c , η d Indicates the unit charge and discharge efficiency, η SOC Indicates the self-damage rate of the energy storage device;

[0038] They refer to the charging and discharging power of the energy storage device in cycle t; C SOC Represents the capacity of the energy storage device; γ c , γ d They represent the charge and discharge rates of the energy storage device respectively;

[0039] Respectively represent the charge and discharge states of the energy storage device, but the constraint is that both cannot be 1 at the same time; W SOC,max , W SOC,min Indicates the maximum and minimum values ​​of the energy storage device’s charging state, Refers to the charging state at time t;

[0040] represents the capacity of the energy storage device, τ3 represents the safety reserve capacity required by the device; Inequality (10) represents the ramp rate per unit time; τ1 represents the maximum ramp rate, C G Indicates the capacity of thermal power generation equipment;

[0041] Step 14: Analyze the constraints:

[0042] Formula (a2) represents the power balance constraint, which ensures that the total power provided by different sources matches the load demand;

[0043] Formulas (a3) ​​to (a5) impose constraints on the output of thermal power generating units;

[0044] Formula (a6) specifies the minimum and maximum active power output of thermal power generating units at each node;

[0045] Equations (a7) and (a8) describe the power generated by the wind turbine and photovoltaic system, respectively;

[0046] Formulas (a6) to (a11) give constraints related to energy storage capacity and operation, which ensure that the charging and discharging operations of the energy storage device meet certain conditions; Formula (a8) sets the charging power limit of the energy storage device within time period t, while (a9) sets the discharge power limit;

[0047] Formula (a12) represents the safety standby constraint, which ensures the reserve capacity of the equipment to maintain safe operation; finally, formula (a13) imposes constraints on the system's ramp-up capability, limiting the maximum growth rate of power output per unit time.

[0048] Furthermore, the specific working steps of step 2 are as follows:

[0049] Step 21: Use numerical weather prediction methods to conduct climate assessment;

[0050] The FourCastNet model uses an adaptive Fourier neural operator to convert the mixed operation process into a continuous global convolution, and uses thousands of integrated nodes for efficient large-scale integrated prediction and high-confidence evaluation of prediction errors under extreme conditions; the standard self-attention calculation formula is as follows:

[0051]

[0052] Where X represents input, XW q represents the query matrix of X, XW k represents the bond matrix of X, XW v represents the value matrix of X;

[0053] Step 211: Obtain a meteorological image of a region as input X; retrieve regions with high light radiation intensity and wind speed in the image and mark them;

[0054] Step 212: Set <·> represents the dot product operation, which is used to measure the similarity between two vectors;

[0055] Given any input image X, by calculating the attention value between the sth token and the tth token of image X, formula (b1) can be rewritten as:

[0056] Att(X,s,t)=K[s,t]X[t]W v (b2);

[0057] Where K[s,t] represents the similarity comparison between the sth tag and the tth tag; X[t]W v is the value of the tth token;

[0058] Step 213: Define a kernel κ[s,t]=K[s,t]·W v , swap the order of K[s,t] and X[t], and we get:

[0059] Att(X,s,t)=X[t]κ[s,t] (b3);

[0060] By summing the above equations (b2) and (b3) over t, we get the operator form of self-attention:

[0061]

[0062] Converting equation (b4) into a continuous integral yields:

[0063]

[0064] Among them, ∫k(st)X(t)dt=(κ*X)(s) is the convolution formula;

[0065] Step 214: According to the fast Fourier transform, the convolution in the time-space domain is equal to the product in the frequency domain; the Fourier neural operator can be written as:

[0066]

[0067] where · represents matrix multiplication; represents the continuous Fourier transform and its inverse transform;

[0068] Step 216: The input data is converted to the frequency domain via FFT, multiplied by a learnable weight, and then converted back to the spatial domain via inverse fast Fourier transform; after discretization, becomes:

[0069] z m,n =[DFT(X)] m,n (b7);

[0070] After discretization, becomes:

[0071]

[0072] After discretization, becomes:

[0073]

[0074] Furthermore, the subsequent working steps of step 216 are as follows:

[0075] Step 217: Equation (b9) is in the form of discrete FNO; the parameter size of discrete FNO is O(Nd 2 ), AFNO will reweight Decomposed into a set of shared weights, each of dimension Rewrite equation (b7) as:

[0076]

[0077] Definition of AFNO:

[0078]

[0079] Among them, the multilayer perceptron (MLP) is a two-layer feed-forward network shared among all tokens; W1, W1 is also shared among all tokens;

[0080] Step 218: AFNO utilizes soft thresholding and shrinkage operations to promote sparsity of labels:

[0081]

[0082] Step 219: In the context of meteorological data, multiple meteorological elements are integrated into a time series and organized in the AFNO format of time, variable, latitude, and longitude;

[0083] FourCastNet maps the input variables to two dimensions (h×w) in a 720×1440 (latitude-longitude) grid, with each region represented as a d-dimensional label;

[0084] The generated patch sequence is combined with the positional encoding and input into a series of AFNO layers, which are described as tokens in a hybrid module; the hybrid module combines the AFNO learning method with the visual transformer architecture;

[0085] The AFNO method utilizes a two-layer MLP structure for computation, converts the input information into the Fourier domain through 2D discrete Fourier transform, applies information weighting in the Fourier domain to promote sparsity, converts the information back to the patch domain using IFFT, and adds residual connections;

[0086] The result is fed to a linear decoder to obtain a final output; the final output result of step 219 is used as the climate benefit;

[0087] Step 22: Define and calculate the evaluation indicators of the information value quantitative evaluation system.

[0088] Furthermore, the specific working steps of step 22 are as follows:

[0089] Step 221: Shannon entropy is expressed as follows:

[0090]

[0091] X represents the information matrix, x ij represents an element in the matrix, and the probability of occurrence of this element is given by p(x ij ), which is calculated using the logarithm with base b; usually b is 2, indicating that the unit of measurement of information is 1 bit;

[0092] Step 222: The non-noise ratio can be expressed as follows:

[0093]

[0094] Where N(X) represents the total number of elements in the sample, and Z(X) represents the number of problematic elements in the sample;

[0095] Step 223: The renewable energy abundance is expressed as follows:

[0096]

[0097] in,, Indicates the wind speed frequency within the optimal range, Indicates the light irradiance frequency within the optimal range; W v and S s Respectively represent the total amount of wind speed information and light irradiance information; γ1 and γ2 are coefficients;

[0098] Step 224: The renewable energy matching degree is expressed as follows:

[0099]

[0100] Among them, N V , N S denote the number of wind turbines and PV arrays, respectively; represents the planned output of the wind turbine and PV array at time t, and represents the theoretical output of the wind turbine and photovoltaic array at time t; the variable t max represents the last moment considered by the model;

[0101] Step 225: The extreme event value is expressed as follows:

[0102]

[0103]

[0104]

[0105] Among them, the evaluation index of extreme event value is defined using the information matrix X; Indicates the maximum wind speed at which the fan can operate, usually set to 25m / s;

[0106] Variable μ ij is a binary variable; if x ij Greater than If yes, it is set to 1, otherwise it is set to 0;

[0107] M represents the frequency of extreme events, and m represents the frequency threshold of extreme events;

[0108] Y is a segmentation coefficient that changes with the value of M. When M does not reach the threshold m, Y shows an upward trend; when M exceeds the threshold m, Y drops sharply and even becomes negative.

[0109] Step 226: Terrain restrictions;

[0110] Using Blender modeling technology, create terrain maps to evaluate suitable planning areas; by analyzing the terrain maps, select relatively flat areas in the target location as candidate areas for wide-area planning.

[0111] Furthermore, the specific working steps of step 3 are as follows:

[0112] Step 31: Determine the weight of the evaluation index and calculate the composite information value index;

[0113] The composite information value indicator q, which is the product of Shannon entropy h(X) and non-noise ratio τ(X), plus renewable energy abundance f(X), minus renewable energy matching m(X) and adding extreme event value j(X):

[0114] The calculation of the composite information value index a(X) is as follows:

[0115] q(X)=w1·h(X)·τ(X)+w2·f(X)-w3·m(X)+w4·j(X);

[0116] Among them, q(X) is the total value of information, w1, w2, w3, w4 are the value coefficients of the corresponding indicators;

[0117] Step 32: Indicator evaluation method;

[0118] Step 321: Determine the subjective weight based on the globally optimized hierarchical analysis process;

[0119] Step 322: Determine a hybrid enhanced artificial intelligence method for target weights;

[0120] The probability of each indicator is shown in the following equation:

[0121]

[0122] The output information entropy of the indicator is expressed as follows:

[0123]

[0124] When p ij = 0, it is stipulated that p ij lnp ij =0, then 0≤h i ≤1;

[0125] The weights of each index are calculated as follows:

[0126]

[0127] Furthermore, the subsequent steps of step 32 are as follows:

[0128] Step 33: Use the Copula model to analyze the relationship between renewable energy output and transmission power;

[0129] Step 331: For the joint distribution H(x1, x2, ..., x n) related marginal distribution functions F1(x1), F2(x2), ... F n (x n ), there exists a Copula function C that satisfies the following relationship:

[0130] H(x1,,x2,…,x n )=C(F1(x1), F2(x2),…,F n (x n ));

[0131] When the marginal distribution function is continuous, the Copula function is unique;

[0132] Step 332: Create a new Copula function by convex combination of multiple Copula functions; the obtained mixed Copula function is as follows:

[0133]

[0134] Among them, C i Represents the i-th Copula function involved in the mixed Copula combination;

[0135] γ i Represents the same as C i The relevant weight parameters, where all γ i is greater than or equal to zero, and all γ i The sum of the values ​​is equal to 1;

[0136] u1, u2 correspond to the marginal distribution functions of the first and second random variables respectively; δ i Represents C i The shape parameters of θ represent the set of all parameters involved.

[0137] Step 333: The Euclidean distance method involves calculating the mean square error between the Copula model and the empirical Copula at a specific sample point; the random variable samples x1, x2 are denoted as {x 1,i , x 2,i}(i=1,…,n), the related edge distribution is defined as F1(x1), F2(x2), and the empirical Copula can be expressed as follows:

[0138]

[0139] The empirical copula can be represented by a function I(·) with an indicator function. When the conditions in the brackets are met, the value of this function is 1, and when the conditions are not met, the value is 0.

[0140] Step 334: The Euclidean distance is defined as follows:

[0141]

[0142] The smaller the value of the Euclidean distance, the closer the alignment between the selected Copula model and the empirical Copula is, and the better the fitting effect is;

[0143] Step 335: A typical daily K-means aggregation method is used to extract representative scenarios and the absolute error is calculated in combination with the transmission curve; the matching level of renewable energy sources, as the value of m(X) decreases, the matching degree of renewable energy sources increases:

[0144]

[0145] Where N represents the number of representative scenes. represents the value coefficient; Corresponding to the predicted power of new energy generation, P S Refers to output power in watts.

[0146] A renewable energy benefit risk analysis system based on environmental data includes:

[0147] Data acquisition module: used to obtain economic and environmental information of the target location and obtain basic information; create constraint functions based on RPGS planning to complete the primary processing of basic information;

[0148] Primary analysis module: used to conduct climate assessment using numerical weather forecasting methods; define and calculate Shannon entropy, noise-free ratio, renewable energy enrichment, renewable energy matching, extreme event value and terrain restrictions, conduct comprehensive indicator evaluation and information benefit measurement on basic information, and obtain the initial analysis results;

[0149] Secondary analysis module: used to analyze the relationship between renewable energy output and transmission power using the Copula model; use relevant indicators to quantify the degree of alignment between the power generation curve and the dispatch power curve; establish a mapping relationship between the recalibrated data and the original NWP data, and correct the loss functions of wind, photovoltaic and thermal power generating units; complete the confirmation and correction of the initial analysis results;

[0150] Real-time monitoring module: used to summarize the analysis and processing results of the data acquisition module, primary analysis module, and secondary analysis module; obtain the missing values ​​of the constraint function, and generate analysis feedback of renewable energy benefit risks to the user.

[0151] Compared with the prior art, the present invention has the following beneficial effects:

[0152] 1. High accuracy: The present invention can obtain various environmental data in real time or near real time, such as wind speed, sunshine duration, temperature, etc., so as to accurately evaluate the benefits and risks of renewable energy projects. This real-time performance ensures the accuracy of the analysis and helps project decision makers make more reasonable choices.

[0153] 2. Comprehensive consideration: The present invention not only takes into account the technical and economic factors of the renewable energy project itself, but also comprehensively considers environmental factors, such as geographical location, climatic conditions, etc. This comprehensive consideration makes the analysis more comprehensive and helps to discover potential risks and opportunities.

[0154] 3. Strong flexibility: The present invention can be adjusted and optimized according to different project requirements and environmental changes; whether it is a large wind farm or a small solar power generation system, the method can be flexibly applied according to actual conditions to achieve accurate assessment of benefits and risks.

[0155] 4. Promote sustainable development: The present invention deeply analyzes the relationship between environmental data and renewable energy benefit risks. This method helps promote the rational planning and layout of renewable energy projects, thereby promoting the optimization of energy structure and sustainable development.

[0156] 5. Improve the reference value of planning: The present invention uses information evaluation method to directly replace the traditional complex planning process by scoring information, and finds the planning point with the best comprehensive benefit. It does not need to carry out complex planning process for all alternative points one by one, which greatly shortens the planning time. BRIEF DESCRIPTION OF THE DRAWINGS

[0157] Other features, objects and advantages of the present invention will become more apparent from the detailed description of non-limiting embodiments made with reference to the following drawings:

[0158] Figure 1 It is a schematic diagram of the analysis method of the present invention;

[0159] Figure 2 It is a schematic diagram of a specific evaluation method of the present invention;

[0160] Figure 3 This is a schematic diagram of the FourCastNet architecture of the present invention;

[0161] Figure 4 1 is a step diagram of the analysis system of the present invention. DETAILED DESCRIPTION

[0162] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and understandable, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0163] Embodiment 1

[0164] See also Figure 1 and Figure 2 , a renewable energy benefit risk analysis method based on environmental data includes:

[0165] It should be noted that, in the present invention, “RPGS” refers to “Renewable Energy Power Generation System (RPGS)”; the target location refers to the area where renewable energy is planned to be developed;

[0166] It is important to note that a large number of data simulation calculations are designed in the present invention, and there are a lot of noun duplications between the data acquisition part and the data processing and analysis part, so the data acquisition part is omitted in the present invention; the specific data appearing in steps 1 to 3 are acquired by default.

[0167] Step 1: Obtain economic and environmental information of the target site to obtain basic information; based on the wide-area planning process, integrate construction capacity, ecological benefits and penalty factors related to extreme events; integrate grid security, renewable energy utilization and external transmission channel utilization; build a planning model to complete the primary processing of basic information;

[0168] The specific steps of step 1 are as follows:

[0169] Step 1: Evaluate the value of information in wide-area planning;

[0170] Obtain relevant economic parameters and equipment parameters in RPGS planning;

[0171] Step 11: In the process of wide-area planning, it is necessary to comprehensively consider construction capacity, ecological benefits and penalty factors related to extreme events, as well as constraints such as grid security, renewable energy utilization and external transmission channel utilization; the objective function of the constructed planning model is as follows:

[0172]

[0173] Formula (a1) where Φ represents a set consisting of the selected wind turbine generator sets, photovoltaic generator sets, energy storage devices, thermal generator sets, and various random scenarios that need to be considered;

[0174] denote the discounted value of the investment costs associated with wind turbine installation, photovoltaic power generation, energy storage equipment, and thermal power generation units, respectively; is a binary variable, which is used as a decision factor to determine whether to invest;

[0175] γ z The probability distribution corresponding to scenario z; represents the operating cost within this specific scenario; σ DIt represents the differences between ecological gas regulation, climate regulation, water conservation, soil shaping and protection and the corresponding economic indicators of the target site;

[0176] It should be noted that the above-mentioned “scene z” and “within the specific scene” both refer to the special ecological climate or special natural environment of the target location.

[0177] Represents the construction capacity of wind and photovoltaic power generation units respectively; is a penalty term, which will impose penalties on extreme events; γ e,t Takes a binary value, indicating whether an extreme event occurs or not. Represents the annualized loss due to extreme events;

[0178] Step 12: The constraints are as follows:

[0179]

[0180]

[0181]

[0182]

[0183]

[0184]

[0185]

[0186]

[0187]

[0188] W SOC,min ≤W t SOC ≤W SOC,max (a11);

[0189]

[0190]

[0191] Step 13: Meaning of the parameters in formulas a2 to a13:

[0192] represents the power generation of the thermal power generating unit at time t, represents the power generation of the photovoltaic unit at time t, represents the power generated by the wind turbine at time t, represents the power demand of the load at time t;

[0193] Represent the charging power and discharging power of the energy storage device respectively;

[0194] Represent the charging power and discharging power of the energy storage device respectively;

[0195] The minimum and maximum active power output of each node of the thermal power generating unit is expressed as P G,min and P G,max , in kilowatts (kW); is a 0-1 variable indicating the start and stop of the thermal power generating unit at each node;

[0196] σ GIV Indicates the upper limit of wind and solar power curtailment; Respectively represent the energy storage state at time t and t-1, η c , η d Indicates the unit charge and discharge efficiency, η SOC Indicates the self-damage rate of the energy storage device;

[0197] They refer to the charging and discharging power of the energy storage device in cycle t; C SOC Represents the capacity of the energy storage device; γ c , γ d They represent the charge and discharge rates of the energy storage device respectively;

[0198] Respectively represent the charge and discharge states of the energy storage device, but the constraint is that both cannot be 1 at the same time; W SOC,max , W SOC,min Indicates the maximum and minimum values ​​of the energy storage device’s charging state, Refers to the charging state at time t;

[0199] represents the capacity of the energy storage device, τ3 represents the safety reserve capacity required by the device; Inequality (10) represents the ramp rate per unit time; τ1 represents the maximum ramp rate, C G Indicates the capacity of thermal power generation equipment;

[0200] Step 14: Analyze the constraints:

[0201] Formula (a2) represents the power balance constraint, which ensures that the total power provided by different sources matches the load demand;

[0202] Formulas (a3) ​​to (a5) impose constraints on the output of thermal power generating units;

[0203] Formula (a6) specifies the minimum and maximum active power output of thermal power generating units at each node;

[0204] Equations (a7) and (a8) describe the power generated by the wind turbine and photovoltaic system, respectively;

[0205] Formulas (a6) to (a11) give constraints related to energy storage capacity and operation, which ensure that the charging and discharging operations of the energy storage device meet certain conditions; Formula (a8) sets the charging power limit of the energy storage device within time period t, while (a9) sets the discharge power limit;

[0206] Formula (a12) represents the safety standby constraint, which ensures the reserve capacity of the equipment to maintain safe operation. Finally, formula (a13) imposes constraints on the system's ramp-up capability, limiting the maximum growth rate of power output per unit time. Incorporating these constraints into the planning model can achieve balanced and efficient operation of the power system.

[0207] The complexity and precision of the information, combined with economic and safety considerations, pose challenges in determining the best planning solution that maximizes overall benefits; moreover, the planning process is highly susceptible to the quality of the information used, as any deviation may lead to suboptimal results;

[0208] Considering information such as wind speed, solar irradiance, and geographical location, the information values ​​such as Shannon entropy, noise-free ratio, renewable energy enrichment, renewable energy matching, and extreme events were analyzed. The Fourier Cast Net neural network algorithm was used to evaluate the value of information, replacing the complex planning process and obtaining the optimal planning result with the best comprehensive benefits.

[0209] Step 2: Use numerical weather forecasting methods to conduct climate assessment; use Shannon entropy, non-noise ratio, renewable energy abundance, renewable energy matching, extreme event value and terrain restrictions as evaluation indicators of the information value quantitative evaluation system; define and calculate the specific values ​​of the evaluation indicators based on the basic information, and complete the secondary processing of the basic information;

[0210] Step 21: Use numerical weather prediction methods to conduct climate assessment;

[0211] The FourCastNet model uses the Adaptive Fourier Neural Operator (AFNO) to convert the mixed operation process into a continuous global convolution, and uses thousands of integrated nodes for efficient large-scale integrated prediction and high-confidence evaluation of prediction errors under extreme conditions; the standard self-attention calculation formula is as follows:

[0212]

[0213] Where X represents input, XW q represents the query matrix of X, XW krepresents the bond matrix of X, XW v represents the value matrix of X;

[0214] Step 211: Obtain a meteorological image of a region as input X; retrieve regions with high light radiation intensity and wind speed in the image and mark them;

[0215] Step 212: Set <·> represents the dot product operation, which is used to measure the similarity between two vectors;

[0216] Given any input image X, by calculating the attention value between the sth token and the tth token of image X, formula (b1) can be rewritten as:

[0217] Att(X,s,t)=K[s,t]X[t]W v (b2);

[0218] Where K[s,t] represents the similarity comparison between the sth tag and the tth tag; X[t]W v is the value of the tth token;

[0219] Step 213: Define a kernel κ[s,t]=K[s,t]·W v , swap the order of K[s,t] and X[t], and we get:

[0220] Att(X,s,t)=X[t]κ[s,t] (b3);

[0221] By summing the above equations (b2) and (b3) over t, we get the operator form of self-attention:

[0222]

[0223] Converting equation (b4) into a continuous integral yields:

[0224]

[0225] Among them, ∫κ(st)X(t)dt=(κ*X)(s) is the convolution formula;

[0226] Step 214: According to Fast Fourier Transform (FFT), the convolution in the time-space domain is equal to the product in the frequency domain; the Fourier Neural Operator (FNO) can be written as:

[0227]

[0228] where · represents matrix multiplication; represents the continuous Fourier transform and its inverse transform;

[0229] Step 216: The input data is converted to the frequency domain via FFT, multiplied by a learnable weight, and then converted back to the spatial domain via Inverse Fast Fourier Transform (IFFT); after discretization, becomes:

[0230] z m,n =[DFT(X)] m,n (b7);

[0231] After discretization, becomes:

[0232]

[0233] After discretization, becomes:

[0234]

[0235] Step 217: Equation (b9) is in the form of discrete FNO, but the parameter W m,n It is still fixed, which makes the model unable to flexibly adapt to changes in input resolution. Therefore, the AFNO proposed in this paper aims to solve this problem.

[0236] The parameter size of discrete FNO is O(Nd 2 ), and AFNO will weight Decomposed into a set of shared weights, each of dimension Therefore, equation (b7) can be rewritten as:

[0237]

[0238] The definition of AFNO is:

[0239]

[0240] The Multilayer Perceptron (MLP) is a two-layer feedforward network shared among all tokens. W1 and W2 are also shared among all tokens. This feedforward network can be made large enough to ensure effective interaction between tokens.

[0241] Step 218: Inspired by the Least Absolute Shrinkage and Selection Operator (LASSO) method, AFNO uses soft thresholding and shrinkage operations to promote label sparsity:

[0242]

[0243] Step 219: See Figure 3 ,In the context of meteorological data, multiple meteorological elements are integrated into a time series and organized in the AFNO time,variable, latitude, longitude format; specifically, FourCastNet maps the input variables to a 720×1440 (latitude-longitude) grid with two dimensions (h×w), and each region is represented as a d-dimensional label;

[0244] The generated patch sequence is combined with the positional encoding and input into a series of AFNO layers, which are described as tokens in a hybrid module; the hybrid module combines the AFNO learning method with the Vision Transformer (ViT) architecture;

[0245] The AFNO method utilizes a two-layer MLP structure for calculation, converts the input information into the Fourier domain through a 2D discrete Fourier transform, applies information weighting in the Fourier domain to promote sparsity, and finally uses IFFT to convert the information back to the patch domain and add residual connections; the result is then fed into a linear decoder to obtain the final output; the final output result of step 219 is used as the climate benefit.

[0246] It should be noted that FourCastNet is a computational method that uses current meteorological data as initial conditions to predict the spatial distribution of meteorological variables in a specific geographical area, thereby constructing a comprehensive meteorological matrix; this method is very suitable for providing meteorological information related to the Shagohuang New Energy Base; therefore, this paper combines existing observational data with real-time weather forecasts to explore the possibility of designing a new energy base based on the FourCastNet model.

[0247] It should be noted that climate benefit refers to the corrected values ​​of relevant climate factors such as wind speed, solar radiation intensity, topography, rainfall, etc. at the target location using numerical weather forecasting methods.

[0248] Step 22: Comprehensive information evaluation method;

[0249] Step 221: Shannon entropy;

[0250] The value of information is measured by considering the size of the information; however, this approach cannot effectively eliminate redundant information; Shannon entropy provides a solution to this problem;

[0251] Shannon entropy can quantify the value of information without being affected by redundant interference; Shannon entropy is used as an indicator to measure the value of information;

[0252] Shannon entropy is expressed as follows:

[0253]

[0254] Among them, X represents the information matrix, x ij represents an element in the matrix, and the probability of occurrence of this element is given by p(x ij ), which is calculated using the logarithm with base b; usually b is 2, indicating that the unit of measurement of information is 1 bit;

[0255] Step 222: non-noise ratio;

[0256] Noise refers to inaccurate or erroneous data in a data set, while the non-noise ratio represents the proportion of correct data relative to the entire data set; the non-noise ratio can be expressed as follows:

[0257]

[0258] Where N(X) represents the total number of elements in the sample, and Z(X) represents the number of problematic elements in the sample;

[0259] Step 223: Renewable energy abundance;

[0260] The abundance of renewable energy is measured by the frequency of the best wind and solar information, reflecting the abundance of wind and solar resources; the abundance of renewable energy is expressed as follows:

[0261]

[0262] in,, Indicates the wind speed frequency within the optimal range, Indicates the light irradiance frequency within the optimal range; W v and S s Respectively represent the total amount of wind speed information and light irradiance information; γ1 and γ2 are coefficients;

[0263] Step 224: Renewable energy matching degree;

[0264] Renewable energy matching is a quantitative assessment based on the consumption of renewable energy, aiming to maximize the consumption rate; it is expressed as follows:

[0265]

[0266] Among them, N V , N S denote the number of wind turbines and PV arrays, respectively; represents the planned output of the wind turbine and PV array at time t, and represents the theoretical output of the wind turbine and photovoltaic array at time t; the variable t maxrepresents the last moment considered by the model;

[0267] Step 225: Extreme event value;

[0268] Reference

[22] emphasizes that meteorological information should include extreme meteorological conditions to ensure the integrity of the data; therefore, this paper proposes an evaluation index for extreme event values, which is expressed as follows:

[0269]

[0270]

[0271]

[0272] Among them, the evaluation index of extreme event value is defined using the information matrix X; Indicates the maximum wind speed at which the fan can operate, usually set to 25m / s;

[0273] Variable μ ij is a binary variable; if x ij Greater than If yes, it is set to 1, otherwise it is set to 0;

[0274] M represents the frequency of extreme events, and m represents the frequency threshold of extreme events;

[0275] Y is a segmentation coefficient that changes with the value of M. When M does not reach the threshold m, Y shows an upward trend; when M exceeds the threshold m, Y drops sharply and even becomes negative.

[0276] Therefore, when the frequency does not reach the threshold, the extreme event value shows a nonlinear growth trend, which can be used to evaluate the completeness and accuracy of information; when the extreme event value exceeds the threshold, the corresponding information value decreases rapidly, reflecting the negative impact of frequent extreme events on planning;

[0277] Step 226: Terrain restrictions;

[0278] By using Blender modeling technology, a terrain map is created to evaluate the suitable area for planning; by analyzing the terrain map, relatively flat areas in the target location are selected as candidate areas for wide-area planning; in this section, only the evaluation criteria are provided, without a detailed description of various terrain features.

[0279] Step 3: Determine the weights of the evaluation indicators, calculate the composite information value index, and measure the information benefits; use the Copula model to analyze the relationship between renewable energy output and transmission power; establish a mapping relationship between the recalibrated data and the original NWP data, and correct the loss function of wind and photovoltaic generators; complete the three-step processing of basic information;

[0280] The specific steps of step 3 are as follows:

[0281] Step 31: Determine the weights of the evaluation indicators, calculate the composite information value index, and measure the information benefits;

[0282] Calculate composite information value index: Shannon entropy is an information value evaluation index to measure data diversity, and non-noise ratio can reflect the accuracy of information. These two indicators reflect the credibility of information;

[0283] The abundance of renewable energy reflects the superiority of wind and solar resource conditions; the renewable energy matching degree provides a measure of the value of information based on the degree of matching between the wind power generation curve and the wind transmission curve; the extreme event value can be used to evaluate the completeness of information and consider the impact of extreme events in wide-area planning; these three indicators reflect the value of information in a wide area, that is, whether the information has the best planning benefits, while also taking into account terrain restrictions;

[0284] However, the assessment based on a single dimension may be inaccurate; to address this problem, we propose a composite information value indicator a, which is the product of Shannon entropy h(X) and non-noise ratio τ(X), plus renewable energy abundance f(X), minus renewable energy matching m(X) and adding extreme event value j(X):

[0285] The calculation of the composite information value index a(X) is as follows:

[0286] a(X)=w1·h(X)·τ(X)+w2·f(X)-w3·m(X)+w4·j(X);

[0287] Among them, a(X) is the total value of information, w1, w2, w3, w4 are the value coefficients of the corresponding indicators;

[0288] Step 32: Indicator evaluation method;

[0289] There are two main methods for multi-criteria decision analysis: subjective method and objective method. In order to achieve the objectivity of indicators as much as possible, while considering the subjective preference of indicators, a hybrid enhanced artificial intelligence method is used to solve the weighting problem in the indicator evaluation system.

[0290] Step 321: Determine the subjective weight based on the globally optimized hierarchical analysis process;

[0291] The analytic hierarchy process (AHP) quantifies the differences between the subjective judgment of the subject and the indicators, thus avoiding disputes caused by too many human judgment factors;

[0292] Step 322: Determine a hybrid enhanced artificial intelligence method for target weights;

[0293] By using preset weights for training, the distance between the information evaluation results and the actual planning results is continuously adjusted according to the confidence level;

[0294] The probability of each indicator is shown in the following equation:

[0295]

[0296] The output information entropy of the indicator is expressed as follows:

[0297]

[0298] When p ij = 0, it is stipulated that p ij lnp ij =0, then 0≤h i ≤1;

[0299] The weights of each index are calculated as follows:

[0300]

[0301] Step 33: Use the Copula model to analyze the relationship between renewable energy output and transmission power;

[0302] In order to evaluate the energy consumption of new energy bases, this paper uses the Copula model to analyze the relationship between renewable energy output and transmission power; uses relevant indicators to quantify the degree of alignment between the power generation curve and the dispatch power curve; and uses the Euclidean distance method

[21] to evaluate the goodness of fit of the Copula model;

[0303] Step 331: For the joint distribution H(x1, x2, ..., x n ) related marginal distribution functions F1(x1), F2(x2), ... F n (x n ), there exists a Copula function C that satisfies the following relationship:

[0304] H(x1,,x2,…,x n )=C(F1(x1), F2(x2),…,F n (x n ));

[0305] When the marginal distribution function is continuous, the Copula function is unique;

[0306] Step 332: In order to improve the accuracy of the model, a new Copula function can be created by a convex combination of multiple Copula functions; the obtained mixed Copula function is as follows:

[0307]

[0308] Among them, C i Represents the i-th Copula function involved in the mixed Copula combination;

[0309] γ i Represents the same as C i The relevant weight parameters, where all γ i is greater than or equal to zero, and all γ i The sum of the values ​​is equal to 1;

[0310] u1, u2 correspond to the marginal distribution functions of the first and second random variables respectively; δ i Represents C i The shape parameters of θ represent the set of all parameters involved.

[0311] Step 333: The Euclidean distance method involves calculating the mean square error between the Copula model and the empirical Copula at a specific sample point; the random variable samples x1, x2 are denoted as {x 1,i , x 2,i}(i=1,…,n), the related edge distribution is defined as F1(x1), F2(x2), and the empirical Copula can be expressed as follows:

[0312]

[0313] The empirical copula can be represented by a function I(·) with an indicator function. When the conditions in the brackets are met, the value of this function is 1, and when the conditions are not met, the value is 0.

[0314] Step 334: The Euclidean distance is defined as follows:

[0315]

[0316] The smaller the value of the Euclidean distance, the closer the alignment between the selected Copula model and the empirical Copula is, and the better the fitting effect is;

[0317] Step 335: A typical daily K-means aggregation method is used to extract representative scenarios and the absolute error is calculated in combination with the transmission curve; the following equation illustrates the matching level of renewable energy sources, and as the value of m(X) decreases, the matching degree of renewable energy increases:

[0318]

[0319] Where N represents the number of representative scenes. represents the value coefficient; Corresponding to the predicted power of new energy generation, P S Refers to output power in watts;

[0320] Step 34: for numerical weather prediction methods, information value enhancement methods;

[0321] Step 341: To address the inaccuracy problem caused by NWP data errors, a benchmark efficacy calculation method based on NWP forecast correction is proposed to obtain accurate effectiveness estimates; this method actively selects samples with high uncertainty for labeling and calibration;

[0322] It should be noted that NWP stands for numerical weather forecasting method; raw NWP data represents climate benefits; and calibrated data represents data with high uncertainty in the raw NWP data.

[0323] Step 342: Establish a mapping relationship between the calibration data and the original NWP data; Penalize according to the applicability of different NWP predictions to obtain higher value NWP information;

[0324] Step 343: Use a multi-layer perceptron to establish a mapping relationship between the NWP data and the correction data;

[0325] Step 344: Taking wind speed as an example, the general form of the loss function is as follows:

[0326]

[0327] in, Indicates the correction value of wind speed in the original NWP data;

[0328] v i Indicates the measured value of wind speed;

[0329] N represents the sample size;

[0330] It should be noted that step 343 does not consider the nonlinear relationship between wind speed and power; in order to solve this problem, it is necessary to consider the rate of change of power relative to wind speed, which can be calculated by wind speed-power L cor The first order difference of the curve is used for correction;

[0331] Step 345: Wind speed-power L cor The first-order difference of the curve is used for correction:

[0332]

[0333] Among them, the wind speed prediction penalty function (expressed as g(v i )) is derived from the wind speed-power function; it captures the relationship between wind speed and the penalty incurred for inaccurate forecasts:

[0334] g(v)=αf′(v)+β;

[0335] Where f′(v) represents the derivative function of the wind speed-power function f(v); the function f(v) is fitted by polynomial regression using historical wind speed and power data; the coefficients α and β are introduced and adjusted to keep the penalty function greater than or equal to zero;

[0336] Step 346: Calculate the reference wind power corresponding to the numerical weather forecast using the corrected wind speed and wind speed-power function f(v), as shown below:

[0337]

[0338] in, represents the reference wind force;

[0339] v c represents the corrected predicted wind speed;

[0340] Step 347: The final predicted power value is determined based on the minimum Euclidean distance to the reference power and the strongest continuity with the previous day's power data among the generated wind power data; this is calculated using the following formula:

[0341]

[0342] Where T represents the total number of time points in a given sample;

[0343] x m,n Indicates the power forecast value of the generated wind power at the current moment;

[0344] Indicates the corresponding value in the reference wind power data;

[0345] x0 represents the power value at the current predicted moment;

[0346] Step 348: Call the method of determining the wind speed loss function in steps 344 to 347 to determine the loss function of the photovoltaic generator set.

[0347] Step 4: Summarize the analysis and processing results of steps 1 to 3; obtain the missing values ​​of the constraint function and complete the analysis of renewable energy benefit risks;

[0348] The specific steps of step 4 are as follows:

[0349] Step 41: Summarize the analysis and processing results of steps 1 to 3;

[0350] Get constraint function x op Substitute the missing values ​​into equation (1);

[0351] Calculate the value of the constraint function; determine whether the value of the constraint function is a positive number;

[0352] Step 42: If the value of the constraint function is a positive number, it is determined whether the terrain restriction is 1; if it is 1, the target location can develop renewable energy; if it is 0, the target location does not develop renewable energy;

[0353] If the value of the constraint function is negative, the target site will not develop renewable energy;

[0354] Step 43: Develop renewable energy at the target location; take "100" as the input object, calculate and compare the loss function values ​​of wind power, photovoltaic and thermal power generating units;

[0355] If the loss function value of the wind turbine generator set is the smallest, then the target location will develop wind power generation;

[0356] If the loss function value of the photovoltaic power generation unit is the smallest, then photovoltaic power generation will be developed at the target location;

[0357] If the loss function value of the thermal power generating unit is the smallest, then the target location will develop thermal power generation;

[0358] Step 44: The analysis results of the above steps 1 to 43 are used as a renewable energy benefit risk analysis report to complete the analysis of renewable energy benefit risks.

[0359] Embodiment 2

[0360] See also Figure 4 A renewable energy benefit risk analysis system based on environmental data includes: a data acquisition module, a primary analysis module, a secondary analysis module, a real-time monitoring module, a database and a controller; wherein the data acquisition module, the primary analysis module, the secondary analysis module and the real-time monitoring module are connected to the database and the server respectively.

[0361] Data acquisition module: used to obtain economic and environmental information of the target location and obtain basic information; create constraint functions based on RPGS planning to complete the primary processing of basic information;

[0362] Primary analysis module: used to conduct climate assessment using numerical weather forecasting methods; define and calculate Shannon entropy, noise-free ratio, renewable energy enrichment, renewable energy matching, extreme event value and terrain restrictions, conduct comprehensive indicator evaluation and information benefit measurement on basic information, and obtain the initial analysis results;

[0363] Secondary analysis module: used to analyze the relationship between renewable energy output and transmission power using the Copula model; use relevant indicators to quantify the degree of alignment between the power generation curve and the dispatch power curve; establish a mapping relationship between the recalibrated data and the original NWP data, and correct the loss functions of wind, photovoltaic and thermal power generating units; complete the confirmation and correction of the initial analysis results;

[0364] Real-time monitoring module: used to summarize the analysis and processing results of the data acquisition module, primary analysis module, and secondary analysis module; obtain the missing values ​​of the constraint function, and generate analysis feedback of renewable energy benefit risks to the user.

[0365] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technical personnel in this field according to actual conditions. For example, if there are weight coefficients and proportional coefficients, their set sizes are to quantify each parameter to obtain a specific value, which is convenient for subsequent comparison. Regarding the size of the weight coefficient and the proportional coefficient, it is sufficient as long as it does not affect the proportional relationship between the parameter and the quantized value.

[0366] Finally, it should be noted that the above-described embodiments are only specific implementations of the present invention, which are used to illustrate the technical solutions of the present invention, rather than to limit them. The protection scope of the present invention is not limited thereto. Although the present invention is described in detail with reference to the above-described embodiments, ordinary technicians in the field should understand that any technician familiar with the technical field can still modify the technical solutions recorded in the above-described embodiments within the technical scope disclosed by the present invention, or can easily think of changes, or make equivalent replacements for some of the technical features therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A renewable energy benefit risk analysis method based on environmental data, characterized in that: The analytical method comprises the following steps: Step 1: Obtain economic and environmental information of the target site to obtain basic information; Based on the wide-area planning process, integrate construction capacity, ecological benefits and penalty factors related to extreme events; Integrate grid security, renewable energy utilization and external transmission channel utilization; Build a planning model to complete the primary processing of basic information; Step 2: Use numerical weather forecasting methods to conduct climate assessment; use Shannon entropy, non-noise ratio, renewable energy abundance, renewable energy matching, extreme event value and terrain restrictions as evaluation indicators of the information value quantitative evaluation system; define and calculate the specific values ​​of the evaluation indicators based on the basic information, and complete the secondary processing of the basic information; Step 3: Determine the weights of the evaluation indicators, calculate the composite information value indicators, and measure the information benefits; use the Copula model to analyze the relationship between renewable energy output and transmission power; complete the three-step processing of basic information; Step 4: Summarize the analysis and processing results of steps 1 to 3; obtain the missing values ​​of the constraint function and complete the analysis of renewable energy benefit risks; The specific steps of step 1 are as follows: Step 11: The objective function of constructing the planning model is as follows: Among them, Φ represents the set consisting of the selected wind turbines, photovoltaic turbines, energy storage devices, thermal turbines, and various random scenarios that need to be considered; denote the discounted value of the investment costs associated with wind turbine installation, photovoltaic power generation, energy storage equipment, and thermal power generation units, respectively; is a binary variable, which is used as a decision factor to determine whether to invest; γ z The probability distribution corresponding to scenario z; represents the operating cost within this specific scenario; σ D It represents the differences between ecological gas regulation, climate regulation, water conservation, soil shaping and protection and the corresponding economic indicators of the target site; Represent the construction capacity of wind power and photovoltaic power generation units respectively; the last item is regarded as a penalty item, which will impose penalties on extreme events; γ e,t Takes a binary value, indicating whether an extreme event occurs or not. Represents the annualized loss due to extreme events; Step 12: The first constraint is as follows: in, represents the power generation of the thermal power generating unit at time t, represents the power generation of the photovoltaic unit at time t, represents the power generated by the wind turbine at time t, represents the power demand of the load at time t; Represent the charging power and discharging power of the energy storage device respectively; The second constraint is as follows: Among them, P G,min and P G,max Respectively represent the minimum and maximum active power output of each node of the thermal power generating unit; A 0-1 variable indicating the start and stop of the thermal power generating unit at each node of the thermal power generating unit; The third constraint is as follows: They represent the charging and discharging power of the energy storage device in cycle t respectively; C SOC Represents the capacity of the energy storage device; γ c , γ d They represent the charge and discharge rates of the energy storage device respectively; Step 14: Analyze the constraints: Obtaining the corresponding parameters of the wind turbine generator set, the thermal power generator set and the energy storage device in step 12 in real time, and substituting them into the calculation formula of the constraint condition; judging whether the constraint condition is satisfied; If the first constraint condition is not met, it is judged that the output power of thermal power, photovoltaic and wind power generating units at the target location does not match the load, and thermal power, photovoltaic and wind power generation are not suitable for the target location; If the second constraint condition is not met, it is judged that the target location is not suitable for thermal power generation; If the third constraint condition is not met, it is judged that the energy storage device capacity at the target location is small; If none of the constraints in step 12 are satisfied, it is determined that the target location is not suitable for renewable energy generation; The specific working steps of step 2 are as follows: Step 21: Use numerical weather prediction methods to conduct climate assessment; The FourCastNet model uses an adaptive Fourier neural operator to convert the mixed operation process into a continuous global convolution, and uses thousands of integrated nodes for efficient large-scale integrated prediction and high-confidence evaluation of prediction errors under extreme conditions; the standard self-attention calculation formula is as follows: Where X represents input, XW q represents the query matrix of X, XW k represents the bond matrix of X, XW v represents the value matrix of X; Step 211: Obtain a meteorological image of a region as input X; retrieve regions with high light radiation intensity and wind speed in the image and mark them; Step 212: Set <·> represents the dot product operation, which is used to measure the similarity between two vectors; Given any input image X, by calculating the attention value between the sth token and the tth token of image X, we get: Att(X,s,t)=K[s,t]X[t]W v ; Where K[s,t] represents the similarity comparison between the sth tag and the tth tag; X[t]W v is the value of the tth token; Step 213: Define a kernel κ[s,t]=K[s,t]·W v , swap the order of K[s,t] and X[t], and we get: Att(X,s,t)=X[t]κ[s,t]; The equations in the above steps 212 and 213 are summed over t to obtain the operator form of self-attention: Converting equation (b4) into a continuous integral yields: Among them, ∫κ(st)X(t)dt=(κ*X)(s) is the convolution formula; Step 214: According to the fast Fourier transform, the convolution in the time-space domain is equal to the product in the frequency domain; the Fourier neural operator is written as: where · represents matrix multiplication; represents the continuous Fourier transform and its inverse transform; Step 216: The input data is converted to the frequency domain via FFT, multiplied by a learnable weight, and then converted back to the spatial domain via inverse fast Fourier transform; after discretization, becomes: z m,n =[DFT(X)] m,n ; After discretization, becomes: After discretization, becomes: The subsequent working steps of step 216 are as follows: Step 217: The equation in step 216 is in the form of a discrete FNO; the parameter size of the discrete FNO is O(Nd 2 ), AFNO will reweight Decomposed into a set of shared weights, each of dimension get: Definition of AFNO: Among them, MLP is a two-layer feed-forward network shared among all tokens; W1 and W2 are also shared among all tokens; Step 218: AFNO utilizes soft thresholding and shrinkage operations to promote sparsity of labels: Step 219: In the context of meteorological data, multiple meteorological elements are integrated into a time series and organized in the AFNO format of time, variable, latitude, and longitude; FourCastNet maps the input variables to two dimensions in a 720×1440 grid, with each region represented as a d-dimensional label; The generated patch sequence is combined with the positional encoding and input into a series of AFNO layers, which are described as tokens in a hybrid module; the hybrid module combines the AFNO learning method with the visual transformer architecture; The AFNO method utilizes a two-layer MLP structure for computation, converts the input information into the Fourier domain through 2D discrete Fourier transform, applies information weighting in the Fourier domain to promote sparsity, converts the information back to the patch domain using IFFT, and adds residual connections; The result is fed to a linear decoder to obtain a final output; the final output result of step 219 is used as the climate benefit; Step 22: Define and calculate the evaluation indicators of the information value quantitative evaluation system; The evaluation indicators in step 22 include Shannon entropy, non-noise ratio, renewable energy abundance, renewable energy matching, extreme event value and terrain restriction: Step 221: Shannon entropy is expressed as follows: Among them, X represents the information matrix, x ij represents an element in the matrix, and the probability of occurrence of this element is given by p(x ij ), which is calculated using the logarithm with base b; b is 2, indicating that the unit of measurement of information is 1 bit; Step 222: The non-noise ratio is expressed as follows: Where N(X) represents the total number of elements in the sample, and Z(X) represents the number of problematic elements in the sample; Step 223: The renewable energy abundance is expressed as follows: in, Indicates the wind speed frequency within the optimal range, Indicates the light irradiance frequency within the optimal range; W v and S s Respectively represent the total amount of wind speed information and light irradiance information; γ1 and γ2 are coefficients; Step 224: The renewable energy matching degree is expressed as follows: Among them, N V , N S denote the number of wind turbines and PV arrays, respectively; represents the planned output of the wind turbine and PV array at time t, and represents the theoretical output of the wind turbine and photovoltaic array at time t; the variable t max represents the last moment considered by the model; Step 225: The extreme event value is expressed as follows: Among them, the evaluation index of extreme event value is defined using the information matrix X; Indicates the maximum wind speed at which the fan can operate, set to 25m / s; Variable μ ij is a binary variable; if x ij Greater than If yes, it is set to 1, otherwise it is set to 0; M represents the frequency of extreme events, and m represents the frequency threshold of extreme events; Y is a segmentation coefficient that changes with the value of M. When M does not reach the threshold m, Y shows an upward trend; when M exceeds the threshold m, Y drops sharply and even becomes negative. Step 226: Terrain constraints; Use Blender modeling technology to create terrain maps to evaluate suitable planning areas; by analyzing the terrain maps, select relatively flat areas in the target location as candidate areas for wide-area planning; The specific working steps of step 3 are as follows: Step 31: Determine the weight of the evaluation index and calculate the composite information value index; The composite information value indicator q, which is the product of Shannon entropy h(X) and non-noise ratio τ(X), plus renewable energy abundance f(X), minus renewable energy matching m(X) and adding extreme event value j(X): The calculation of the composite information value index q(X) is as follows: q(X)=w1·h(X)·τ(X)+w2·f(X)-w3·m(X)+w4·j(X); Among them, q(X) is the total value of information, w1, w2, w3, w4 are the value coefficients of the corresponding indicators; Step 32: Indicator evaluation method; Step 321: Determine the subjective weight based on the globally optimized hierarchical analysis process; Step 322: Determine a hybrid enhanced artificial intelligence method for target weights; The probability of each indicator is shown in the following equation: The output information entropy of the indicator is expressed as follows: When p ij = 0, it is stipulated that p ij lnp ij =0, then 0≤h i ≤1; The weights of each index are calculated as follows: The subsequent steps of step 32 are as follows: Step 33: Use the Copula model to analyze the relationship between renewable energy output and transmission power; Step 331: For the joint distribution H(x1,x2,…,x n ) related marginal distribution functions F1(x1), F2(x2), ... F n (x n ), there exists a Copula function C that satisfies the following relationship: H(x1,x2,…,x n )=C(F1(x1),F2(x2),…,F n (x n )); When the marginal distribution function is continuous, the Copula function is unique; Step 332: Create a new Copula function by convex combination of multiple Copula functions; the obtained mixed Copula function is as follows: Among them, C i Represents the i-th Copula function involved in the mixed Copula combination; γ i Represents the same as C i The relevant weight parameters, where all γ i is greater than or equal to zero, and all γ i The sum of the values ​​is equal to 1; u1, u2 correspond to the marginal distribution functions of the first and second random variables respectively; δ i Represents C i The shape parameters of θ represent the set of all parameters involved. Step 333: The Euclidean distance method involves calculating the mean square error between the Copula model and the empirical Copula at a specific sample point; the random variable samples x1, x2 are denoted as {x 1,i , x 2,i }(i=1,…,n), the relevant edge distribution is defined as F1(x1), F2(x2), and the empirical Copula is expressed as follows: The empirical copula is represented by a function I(·) with an indicator function. When the conditions in the brackets are met, the value of this function is 1, and when the conditions are not met, the value is 0. Step 334: The Euclidean distance is defined as follows: The smaller the value of the Euclidean distance, the closer the alignment between the selected Copula model and the empirical Copula is, and the better the fitting effect is; Step 335: A typical daily K-means aggregation method is used to extract representative scenarios and the absolute error is calculated in combination with the transmission curve; the matching level of renewable energy sources, as the value of m(X) decreases, the matching degree of renewable energy sources increases: Where N represents the number of representative scenes. represents the value coefficient; Corresponding to the predicted power of new energy generation, P S Refers to output power; The specific steps of step 4 are as follows: Step 41: Summarize the analysis and processing results of steps 1 to 3; Get constraint function x op Substitute the missing values ​​into the objective function of the planning model; Calculate the value of the constraint function; determine whether the value of the constraint function is a positive number; Step 42: If the value of the constraint function is a positive number, it is determined whether the terrain restriction is 1; if it is 1, the target location can develop renewable energy; if it is 0, the target location does not develop renewable energy; If the value of the constraint function is negative, the target site will not develop renewable energy; Step 43: Develop renewable energy at the target location; take "100" as the input object, calculate and compare the loss function values ​​of wind power, photovoltaic and thermal power generating units; If the loss function value of the wind turbine generator set is the smallest, then the target location will develop wind power generation; If the loss function value of the photovoltaic power generation unit is the smallest, then photovoltaic power generation will be developed at the target location; If the loss function value of the thermal power generating unit is the smallest, then the target location will develop thermal power generation; Step 44: The analysis results of the above steps 1 to 43 are used as a renewable energy benefit risk analysis report to complete the analysis of renewable energy benefit risks.

2. A renewable energy benefit risk analysis system based on environmental data, applicable to the renewable energy benefit risk analysis method based on environmental data according to claim 1, characterized in that: The analysis system comprises: Data acquisition module: used to obtain economic and environmental information of the target location and obtain basic information; create constraint functions based on RPGS planning to complete the primary processing of basic information; Primary analysis module: used to conduct climate assessment using numerical weather forecasting methods; define and calculate Shannon entropy, noise-free ratio, renewable energy enrichment, renewable energy matching, extreme event value and terrain restrictions, conduct comprehensive indicator evaluation and information benefit measurement on basic information, and obtain the initial analysis results; Secondary analysis module: used to analyze the relationship between renewable energy output and transmission power using the Copula model; use relevant indicators to quantify the degree of alignment between the power generation curve and the dispatch power curve; establish a mapping relationship between the recalibrated data and the original NWP data, and correct the loss functions of wind, photovoltaic and thermal power generating units; complete the confirmation and correction of the initial analysis results; Real-time monitoring module: used to summarize the analysis and processing results of the data acquisition module, primary analysis module, and secondary analysis module; obtain the missing values ​​of the constraint function, and generate analysis feedback of renewable energy benefit risks to the user.

Citation Information

Cited By

  • Distributed photovoltaic cluster operation risk assessment method based on multi-dimensional observability

    CN121436666A

  • A Multidimensional Observational-Based Risk Assessment Method for Distributed Photovoltaic Clusters

    CN121436666B