Regional pumped storage development capability multi-dimensional evaluation system and method based on data feature modeling and Beta distribution

Abstract

The invention relates to the technical field of hydroelectric power generation, in particular to a regional pumped storage development capability multi-dimensional evaluation system and method based on data feature modeling and Beta distribution. The evaluation method comprises the steps of constructing a Beta distribution model of a multilayer structure, introducing an AHP score, an FCE score and a TOPSIS score as original data, and optimizing the original data, so that a more objective evaluation report is obtained, the evaluation system is applied to the evaluation method, and the evaluation efficiency can be improved by combining reasonable statistical calibration in a data driving mode. The objectivity, the dynamism and the accuracy of the evaluation method are improved, and more scientific and reasonable support can be provided in complex decisions.

Description

Technical Field

[0001] The present invention relates to the technical field of hydropower generation, and more specifically, to a multi-dimensional evaluation system and method for the development capacity of regional pumped storage based on data feature modeling and Beta distribution. Background Art

[0002] Against the backdrop of intensifying global climate change and growing environmental protection awareness, the adjustment of the energy structure has become an important issue that countries are competing to focus on. The application of renewable energy continues to expand, and pumped storage, as a key energy storage technology, plays an indispensable role in balancing power supply and demand, ensuring grid security, and promoting the effective utilization of renewable energy. Pumped storage power stations can store energy during low-demand periods and release it during peak-demand periods, thereby improving the overall efficiency and flexibility of the power system.

[0003] However, traditional evaluation methods often rely on expert experience and intuition. Although this method can reflect project characteristics to a certain extent, due to the lack of sufficient data support, the evaluation results are often affected by uncertainty and subjectivity. In addition, with the development of society and economy and the progress of technology, the original evaluation results may no longer be suitable for the current situation, and it is difficult for traditional methods to dynamically adjust to adapt to new data.

[0004] The implementation of pumped storage projects needs to comprehensively consider multiple factors, including the suitability of geographical location, the convenience of grid connection conditions, the improvement of transportation infrastructure, the supporting ability of economic development level, the payment ability of residents, and the possible space limitations and population density. There are complex interaction relationships among these factors, and traditional single-dimensional evaluation methods are difficult to comprehensively reflect the impact of these interactions on project feasibility.

[0005] It is worth noting that the construction and operation of pumped storage power stations are long-term investments, so it is necessary to accurately predict the energy development trends in the next few decades. Traditional evaluation methods often ignore the long-term potential and sustainability of projects and focus too much on short-term economic benefits, which may affect the long-term planning and development of projects.

[0006] Therefore, there is an urgent need to develop a new evaluation method that can not only integrate multiple data sources but also dynamically adjust weights according to historical data to adapt to the changing external environment and development needs. Summary of the Invention

[0007] The present invention provides a multi-dimensional evaluation system and method for the development capacity of regional pumped storage based on data feature modeling and Beta distribution, which can solve the problems of strong subjectivity and difficulty in dynamic updating in traditional evaluation methods.

[0008] The multi-dimensional evaluation method for the development capacity of regional pumped storage based on data feature modeling and Beta distribution according to the present invention includes:

[0009] Construct the score input layer of the scoring method, collect various indicators related to the development capacity of regional pumped storage, and use the AHP method, FCE method, and TOPSIS method to score each indicator to obtain the AHP score, FCE score, and TOPSIS score;

[0010] Construct the theoretical distribution construction layer, set specific distribution parameters for the AHP method, FCE method, and TOPSIS method, construct the Beta distribution density function, and generate quantiles;

[0011] Construct the absolute level evaluation layer, based on the AHP score, FCE score, and TOPSIS score, construct a score position calculation model to obtain the score position and the quantified rating score, and based on the score position, construct a development potential calculation model to obtain the development potential score;

[0012] Construct the comprehensive ability evaluation layer, and calculate the average value of the development potential scores of each indicator under the AHP method, FCE method, and TOPSIS method;

[0013] Construct the comprehensive evaluation layer, based on the quantified rating scores and development potential scores of each indicator under the AHP method, FCE method, and TOPSIS method, construct a basic index calculation model to obtain the basic ability score, development potential score, and balance score, and perform weighted calculation on them to obtain the comprehensive score;

[0014] Construct the feature analysis layer, based on the quantified rating scores of each indicator under the AHP method, FCE method, and TOPSIS method, construct a score feature analysis model to obtain the score dispersion and coefficient of variation, and based on the development potential scores of each indicator under the AHP method, FCE method, and TOPSIS method, construct a feature analysis model to obtain the potential mean, potential dispersion, and dominant development direction.

[0015] Preferably, each indicator in the score input layer of the scoring method includes the development demand degree (A1), resource endowment degree (A2), economic bearing capacity (A3), and construction constraint degree (A4).

[0016] Preferably, constructing the Beta distribution density function in the theoretical distribution construction layer includes:

[0017] Construct the Beta distribution, and its formula is:

[0018]

[0019] Among them, x is one of the AHP score, FCE score, and TOPSIS score, x ∈ [0, 1], α and β are distribution parameters,

[0020]

[0021] Cumulative Beta distribution, and its formula is:

[0022]

[0023] Calculate the quantile, and its formula is:

[0024] Q(p) = F -1 (p; α, β) = {x|F(x; α, β) = p} (3)

[0025] Preferably, constructing the scoring position calculation model in the absolute level evaluation layer includes:

[0026] Probability density discretization, and its formula is:

[0027]

[0028] Calculate the density value sequence, and its formula is:

[0029] Y = {y i = f(x i ; α, β)|i = 0, 1,..., N - 1} (5)

[0030] Calculate the scoring position, and its formula is:

[0031]

[0032] Calculate the quantization rating score, and its formula is:

[0033]

[0034] Preferably, constructing the development potential calculation model includes:

[0035] Current position calibration, and its formula is:

[0036] p c = P(x) (8)

[0037] In the formula, p c is the current position;

[0038] Target threshold determination, and its formula is:

[0039] T(p c ) = min{p ∈ {0.9, 0.75, 0.5, 0.25}|p > p c} (9)

[0040] In the formula, T(p c ) is the target threshold;

[0041] Development distance calculation, and its formula is:

[0042]

[0043] In the formula, Δ(x) is the development distance;

[0044] Development potential score calculation, and its formula is:

[0045]

[0046] In the formula, D(x) is the development potential score.

[0047] Preferably, the construction of the basic index calculation model in the comprehensive evaluation layer includes:

[0048] Calculating the basic ability score:

[0049]

[0050] Among them, C b is the basic ability score, M = 3, s i = L(x i ) score is the quantitative rating score of the i-th method, and score is one of the AHP method, FCE method, and TOPSIS method;

[0051] Calculating the development potential score:

[0052]

[0053] Among them, C p is the development potential score, M = 3, D(x i ) is the development potential score at x i location.

[0054] Calculating the balance score:

[0055]

[0056] Among them, C e is the balance score,

[0057] Preferably, the formula for the comprehensive score in the comprehensive evaluation layer is:

[0058] S = w 1 C b + w 2 C p + w 3 C e (15)

[0059] Among them, w1 = 0.5, w 2 = 0.3, w 3 = 0.2, C b is the basic ability score, C p is the development potential score, C e is the balance score.

[0060] Preferably, constructing the score feature analysis model in the feature analysis layer includes:

[0061] Calculate the separation divergence, and its formula is:

[0062]

[0063] In the formula, σ s is the separation divergence;

[0064] Calculate the coefficient of variation, and its formula is:

[0065]

[0066] In the formula, CV s is the coefficient of variation.

[0067] Preferably, constructing the feature analysis model in the feature analysis layer includes:

[0068] Calculate the potential mean, and its formula is:

[0069]

[0070] In the formula, is the potential mean;

[0071] Calculate the potential dispersion, and its formula is:

[0072]

[0073] In the formula, σ D is the potential dispersion;

[0074] Calculate the leading development direction, and its formula is:

[0075] d * = argmax i {D(x i )} (20)

[0076] In the formula, d * is the leading direction.

[0077] The multi-dimensional evaluation system for the development capacity of regional pumped storage based on data feature modeling and Beta distribution according to the present invention includes: a data input module, a data preprocessing module, a feature extraction and weight assignment module, a comprehensive evaluation model construction module, an intelligent grading and result output module, a dynamic adjustment mechanism module, an intelligent grading and result output module, and a dynamic adjustment mechanism module;

[0078] The data input module is used to collect various indicators related to the development capacity of pumped storage, form original indicator data, and send it to the data preprocessing module;

[0079] The data preprocessing module includes a data cleaning unit, a data standardization unit, and a data fusion unit. The data cleaning unit is responsible for removing invalid or incorrect original indicator data; the data standardization unit converts the original indicator data from different sources into a unified format and unit; the data fusion unit is responsible for integrating various original indicator data types into a consistent data set;

[0080] The feature extraction and weight assignment module is used to construct a judgment matrix based on the data set to obtain the weights of various indicators;

[0081] The comprehensive evaluation model construction module constructs a Beta distribution density function based on the AHP evaluation model, the FCE evaluation model, and the TOPSIS evaluation model to obtain an evaluation report;

[0082] The intelligent grading and result output module grades based on the evaluation report and generates charts from the evaluation report;

[0083] The dynamic adjustment mechanism module is used to update the original indicator data.

[0084] Beneficial effects:

[0085] The present invention can conduct theoretical analysis on the scoring tendencies of different evaluation methods (AHP, FCE, and TOPSIS), and calibrate by setting reasonable Beta distribution parameters to eliminate the scoring biases of each method in specific applications, which helps to achieve a more comprehensive and balanced evaluation, so as to obtain more valuable evaluation results. Description of the drawings

[0086] Figure 1 is the overall flow chart of the multi-dimensional evaluation system for the development capacity of regional pumped storage based on data feature modeling and Beta distribution according to the present invention;

[0087] Figure 2 is Figure 1 the schematic structural diagram of the analytic hierarchy process (AHP) in

[0088] Figure 3 is Figure 1 the schematic structural diagram of the fuzzy comprehensive evaluation (FCE) in

[0089] Figure 4 For Figure 1 the structural schematic diagram of the superiority solution distance method (TOPSIS);

[0090] Figure 5 This is the flow chart of the multi-dimensional evaluation system method for the regional pumped storage development ability based on data feature modeling and Beta distribution in the present invention. Specific implementation manners

[0091] It should be noted that

[0092] The technical solution of the present invention will be described in detail below through the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present invention and the specific features in the embodiments are detailed descriptions of the technical solution of the present invention, rather than limitations on the technical solution of the present invention. Without conflict, the embodiments of the present invention and the technical features in the embodiments can be combined with each other.

[0093] The term "and / or" is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist simultaneously, and B exists alone. In addition, the character " / " generally represents an "or" relationship between the associated objects before and after.

[0094] Embodiment 1

[0095] To make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0096] As shown in Figures 1-5 , this embodiment provides a multi-dimensional evaluation method for the regional pumped storage development ability based on data feature modeling and Beta distribution, including:

[0097] Construct the score input layer of the scoring method, collect various indicators related to the regional pumped storage development ability, and use the AHP method, FCE method and TOPSIS method to score each indicator to obtain the AHP score, FCE score and TOPSIS score;

[0098] Such as Figure 2The evaluation process of the Analytic Hierarchy Process (AHP) shown aims to conduct in-depth analysis of multi-faceted data and finally obtain a comprehensive evaluation result. The entire process is divided into multiple key levels. At the data preparation level, various raw data are collected, including key data such as development needs, resource endowments, economic carrying capacity, and construction constraints. These data form the basis of the analysis, covering all important aspects of the evaluation object. Then comes data preprocessing, which includes Z-score standardization, extreme value processing, and interval transformation. Z-score standardization is used to eliminate the dimensional differences of data to make it comparable; extreme value processing is used to solve the problem of data outliers; interval transformation ensures that all data fall within a unified interval, facilitating subsequent calculations. At the feature calculation level, by calculating eigenvalue such as correlation coefficient, coefficient of variation, and information entropy, the internal characteristics of the data are deeply analyzed. The correlation coefficient is used to judge the correlation between indicators, the coefficient of variation reflects the degree of dispersion of the data, and information entropy measures the degree of chaos of the data. These indicators together help to understand the complexity of the data. The judgment matrix construction layer is a key step in constructing the A1-A4 sub-criterion judgment matrix and the criterion layer judgment matrix. These matrices are used to express the relative importance relationships between sub-criteria and criteria and are the basis for weight calculation. At the weight calculation layer, by calculating eigenvalues and eigenvectors, consistency tests are carried out to ensure the rationality of the judgment matrix, and normalization processing of weights is carried out. In this way, the relative importance of each indicator can be accurately determined, making the evaluation result more objective and reasonable. Finally, at the score calculation layer, score calculations are carried out for different dimensions to obtain the scores of development demand degree, resource endowment degree, economic carrying capacity, and construction constraint degree. By synthesizing the scores of each dimension, the final AHP score is calculated, providing a clear evaluation basis for decision-making.

[0099] Among them, the AHP score is obtained through the following steps:

[0100] Construct the theoretical distribution construction layer, set specific distribution parameters for the AHP method, FCE method, and TOPSIS method, construct the Beta distribution density function, and generate quantiles;

[0101] Construct the absolute level evaluation layer. Based on the AHP score, FCE score, and TOPSIS score, construct a score position calculation model to obtain the score position and the quantified rating score, and based on the score position, construct a development potential calculation model to obtain the development potential score;

[0102] Construct the comprehensive ability evaluation layer, and calculate the mean value of the development potential scores of each indicator under the AHP method, FCE method, and TOPSIS method;

[0103] Construct a comprehensive evaluation layer, and based on the quantitative rating scores and development potential scores of each index under the AHP method, FCE method, and TOPSIS method, construct a basic index calculation model to obtain the basic ability score, development potential score, and balance score, and perform weighted calculation on them to obtain the comprehensive score;

[0104] Construct a feature analysis layer. Based on the quantitative rating scores of each index under the AHP method, FCE method, and TOPSIS method, construct a score feature analysis model to obtain the score dispersion degree and coefficient of variation. Based on the development potential scores of each index under the AHP method, FCE method, and TOPSIS method, construct a feature analysis model to obtain the potential mean value, potential dispersion degree, and dominant development direction.

[0105] Preferably, the indexes included in the score input layer of the scoring method include the development demand degree (A1), resource endowment degree (A2), economic carrying capacity (A3), and construction constraint degree (A4).

[0106] Among them, the development demand degree (A1), resource endowment degree (A2), economic carrying capacity (A3), and construction constraint degree (A4) are obtained through the following steps:

[0107] I. Development demand degree (A1):

[0108] 1. Regional renewable energy characteristic quantity:

[0109] Renewable energy installed capacity ratio R i,t , where i represents the regional identifier and t represents time; renewable energy development coefficient α i , which is used to characterize the development potential of regional renewable energy.

[0110] 2. Power load characteristic quantity

[0111] Peak-valley difference rate P i,t , which is used to characterize the regional electricity load characteristics;

[0112] Regional load factor β i , which is used to characterize the concentration degree of regional industrial load.

[0113] 3. Grid accommodation characteristic quantity

[0114] Abandoned wind and light rate A i,t , which characterizes the renewable energy accommodation capacity;

[0115] Grid adaptability coefficient γ i , which characterizes the regional grid acceptance capacity.

[0116] For the evaluation region i, in the time series T, form a data matrix D i :

[0117]

[0118] Among them, D i is the development demand degree.

[0119] II. Resource endowment degree (A2):

[0120] 1. Geographic condition evaluation index set G, including:

[0121] Terrain drop h i , in meters;

[0122] Available site area a i , in square kilometers;

[0123] Water resource endowment coefficient w i ';

[0124] Geological condition coefficient g i ∈[0,1].

[0125] 2. Power grid access condition index set E, including:

[0126] Voltage level v i , value set V = {110, 220, 330, 500} kV;

[0127] Power grid stability coefficient s i ∈[0,1];

[0128] Substation distance d s,i , in kilometers;

[0129] Load center distance d l,i , in kilometers.

[0130] 3. Traffic condition index set T, including:

[0131] Highway accessibility distance r d,i , in kilometers;

[0132] Railway accessibility distance r t,i , in kilometers;

[0133] Road grade l i , value set L = {1, 2, 3, 4};

[0134] Construction convenience coefficient c i ∈[0,1].

[0135] For the evaluation area i, construct the evaluation matrix M i :

[0136]

[0137] After standardizing each index, a standardized score matrix S is obtained. i :

[0138]

[0139] Among them, each standardized score satisfies: S x,i ∈[0, 100].

[0140] By introducing the weight coefficient matrix W, the final resource endowment degree can be obtained:

[0141]

[0142] Among them:

[0143] III. Economic carrying capacity (A3):

[0144] 1. Economic development index set E, including:

[0145] Gross domestic product per capita G i (t), with the unit of 10,000 yuan per person;

[0146] Economic growth rate α i (t);

[0147] Regional population base P i , with the unit of 10,000 people;

[0148] Fiscal revenue scale F i (t), with the unit of 100 million yuan;

[0149] Fiscal growth rate β i (t).

[0150] 2. Energy economy index set U, including:

[0151] Unit GDP energy consumption coefficient e i (t), satisfying the time decay function e i (t) = e i (0)(1 - η) t . Among them, η is the annual energy efficiency improvement rate.

[0152] Benchmark electricity price p b (t), satisfying: p b (t) = p b (0)(1 + γ) t . Among them, γ is the annual electricity price growth rate.

[0153] 3. Resident payment ability index set A, including:

[0154] Per capita disposable income I i (t), with the unit of 10,000 yuan per person;

[0155] Income growth rate δ i (t);

[0156] Electricity price affordability coefficient C i (t), and the calculation formula is:

[0157]

[0158] where Q r is the annual benchmark electricity consumption of residents.

[0159] For the evaluation area i at the time series t, construct the economic condition evaluation matrix M i (t):

[0160]

[0161] Introduce the economic cycle influence factor ω(t):

[0162]

[0163] where T is the economic cycle length and ∈ is the fluctuation amplitude coefficient.

[0164] Finally, obtain the standardized economic carrying capacity S i (t):

[0165] S i (t) = f(M i (t), ω(t)) ∈ [0, 100]

[0166] IV. Construction constraint degree (A4):

[0167] 1. Spatial constraint index set S, including:

[0168] Total area A of the region i , with the unit of square kilometers;

[0169] Proportion R of the ecological protection area p (t), which satisfies the dynamic evolution equation:

[0170] R p (t) = R p (0) + λt

[0171] where λ is the annual growth rate of the protection intensity, and R p (t) ≤ R max .

[0172] 2. Population density constraint index set P, including:

[0173] Regional population quantity N i (t), which satisfies:

[0174] N i N(t) = N i (0)(1 + r i ) t

[0175] where r i is the natural population growth rate.

[0176] The urbanization rate U i (t) satisfies:

[0177] U i U(t) = U i (0)(1 + μ) t

[0178] where μ is the annual urbanization growth rate.

[0179] The population density D i (t), the calculation formula:

[0180]

[0181] 3. The load constraint index set L includes:

[0182] The load center capacity C i (t), in units of 10,000 kW;

[0183] The industrial density coefficient η i ;

[0184] The load growth rate α i ;

[0185] The load evolution equation:

[0186] C i C(t) = C i (0)(1 + η i α i ) t

[0187] For the evaluation area i in the time series t, construct the constraint condition matrix K i (t):

[0188]

[0189] Based on this constraint matrix, the regional construction constraint degree Ω i (t) can be obtained:

[0190] Ω i Ω(t) = {x|g(K i (t), x) ≤ 0}

[0191] Among them, g(K i (t), x) is a set of constraint condition functions.

[0192] Please refer to Figure 2 , the Analytic Hierarchy Process (AHP) evaluation algorithm is a structured decision-making support method aimed at solving complex multi-level decision-making problems. First, AHP systematizes the problem by establishing a four-layer progressive evaluation structure. The goal layer (L0) is defined as the comprehensive score of regional development ability, and the criterion layer (L1) consists of development demand degree (A1), resource endowment degree (A2), economic carrying capacity (A3), and construction constraint degree (A4). On this basis, the sub-criterion layer (L2) refines the specific evaluation indicators under each criterion, and the scheme layer (L3) lists the regions to be evaluated. The design of this hierarchical structure helps to clarify the goals and directions of the evaluation.

[0193] In the objective construction of the judgment matrix, first establish the judgment matrix of the criterion layer:

[0194]

[0195] In the formula, I i is the information entropy of index i, ρ i is the correlation coefficient with the goal, and CV i is the coefficient of variation.

[0196] This judgment matrix is based on the contribution degree of historical data and determines the elements of the judgment matrix by calculating the information entropy, correlation coefficient with the goal, and coefficient of variation of each index. It ensures that the judgment matrix reflects the relative importance of the indexes in decision-making.

[0197] Subsequently, for the construction of the judgment matrix of the sub-criterion layer:

[0198]

[0199] By constructing the judgment matrix of the sub-criterion layer for the sub-indexes under each criterion, calculate the relative importance:

[0200]

[0201] Where: e i is the information entropy, h i is the redundancy, and s i is the sensitivity coefficient.

[0202] By calculating the relative importance, combine the information entropy, redundancy, and sensitivity coefficient together to quantify the relative weight of each sub-index. It ensures the objectivity and accuracy of the evaluation.

[0203] In the stage of optimizing the consistency of the judgment matrix, first construct the optimization model:

[0204]

[0205] The constraints are as follows:

[0206]

[0207] Through the iterative optimization algorithm, first initialize the judgment matrix A; then calculate the consistency ratio CR; if CR ≥ 0.1, then make corrections:

[0208]

[0209] Finally, repeat the steps until the consistency requirement is met.

[0210] In terms of weight calculation, the eigenvalue method is used to calculate the eigenvalues:

[0211] (A - λI)W = 0

[0212] And normalize the eigenvalues to obtain the weights of each index.

[0213]

[0214] In the multi-level comprehensive scoring, first calculate the scores of each evaluation object through local weight combination. The specific formula combines the score of the kth evaluation object, the weight of the ith index, and the standardized index value. Finally, the global score is calculated by summarizing the weights and standardized index values of each level, taking into account the number of levels L and the number of indexes in each level, realizing a comprehensive evaluation of all evaluation objects.

[0215] The local weight combination formula is:

[0216]

[0217] Where S k is the score of the kth evaluation object, w i is the weight of the ith index, S ki is the standardized index value.

[0218] AHP scoring calculation:

[0219]

[0220] Where L is the number of levels, n l is the number of indexes in the lth layer, w i l is the weight of the ith index in the lth layer, S ki l is the corresponding standardized index value.

[0221] The Analytic Hierarchy Process provides a scientific and feasible solution for complex decision-making problems through a systematic hierarchical structure, objective judgment matrix construction, and strict consistency optimization.

[0222] Please refer to Figure 3 , the Fuzzy Comprehensive Evaluation Algorithm (FCE) is an effective method for quantitatively and qualitatively evaluating complex systems. First, a five-level quantitative evaluation scale is constructed: V = {v 1 , v 2 , v 3 , v 4 , v 5}, specifically including {poor (20), relatively poor (40), average (60), relatively good (80), and good (100)}, and the corresponding quantitative mapping function is defined:

[0223]

[0224] To achieve a clear division of the evaluation levels. Then, a dynamic membership function is designed, and the membership degree is calculated using the improved Gaussian function form. The central value is determined through the mean of historical data, and a basic adjustment coefficient and a variation adjustment coefficient are introduced. The calculation of the dynamic fuzziness parameter combines the index variation coefficient and the sample mean. This process ensures the flexibility and accuracy of the evaluation.

[0225] Among them, the membership degree calculation formula based on the improved Gaussian function is:

[0226]

[0227] Determination of the central value:

[0228] c k = 20(k + 1), k = 0, 1, 2, 3, 4

[0229] Dynamic fuzziness parameter:

[0230]

[0231] In the formula, β is the basic adjustment coefficient, and its value range is [0.1, 0.3], γ is the variation adjustment coefficient, and its value range is [0.2, 0.5], CV x is the index variation coefficient, is the sample mean.

[0232] In the construction of the fuzzy relation matrix, for each level of each index, the elements of the fuzzy relation matrix are calculated, combined with the tail attenuation coefficient, to reflect the relative influence of different levels. For the jth level of the ith index:

[0233]

[0234] where δ is the tail attenuation coefficient with a value of 0.1

[0235] Subsequently, a multi-level fuzzy comprehensive operation is carried out. First, fuzzy comprehensiveness is performed at the sub-criterion level by combining the weights of each sub-criterion with its fuzzy relationship to obtain the comprehensive result of the sub-criterion. Then, synthesis is carried out at the criterion level to further integrate the results of each sub-criterion through the weights of each criterion, thus forming a comprehensive evaluation system.

[0236] Fuzzy comprehensiveness at the sub-criterion level:

[0237]

[0238] Fuzzy synthesis at the criterion level:

[0239] B = W[B 1 ; B 2 ; B 3 ; B 4

[0240] Finally, based on the defuzzification process of the centroid method, the FCE score is obtained:

[0241]

[0242] This score is obtained by weighted calculation of the comprehensive result of each index and its corresponding central value, providing a scientific basis for decision-making. Generally speaking, the fuzzy comprehensive evaluation algorithm realizes the effective evaluation of multi-dimensional complex problems through flexible parameter settings and hierarchical calculation methods.

[0243] Please refer to Figure 4 , the TOPSIS evaluation algorithm is a method widely used in multi-attribute decision-making analysis. Based on the principles of distance measure and relative closeness, it supports scientific decision-making by constructing an optimized decision matrix, ideal solution, and evaluation mechanism.

[0244] First, the algorithm processes the decision matrix through vector normalization:

[0245]

[0246] to ensure the comparability of indicators with different dimensions. In terms of the optimization and integration of weights, the analytic hierarchy process (AHP), entropy weight method, and coefficient of variation weights are comprehensively considered to form a relatively objective weight structure, making the influence of different indicators on the decision result more accurate. The weight optimization and integration formula is:

[0247]

[0248] where w j AHP is the AHP weight, w j E ​is the entropy weight, w j V is the coefficient of variation weight, where α + β + γ = 1.

[0249] To construct the dynamic ideal solution, the algorithm defines the positive ideal solution:

[0250]

[0251] Negative ideal solution:

[0252]

[0253] In the formula, σ j is the standard deviation of index j, and δ is the expansion coefficient, with a value range of [0.1, 0.3].

[0254] The advantages and disadvantages of each evaluation object are measured by maximizing the positive ideal solution and minimizing the negative ideal solution. When defining the ideal solution, the standard deviation of each index and an expansion coefficient with a value range from 0.1 to 0.3 are considered to ensure that the ideal solution has dynamic adaptability and can better reflect the actual situation.

[0255] When calculating the distance between each evaluation object and the ideal solution, the TOPSIS algorithm uses the Minkowski - Mahalanobis mixed distance:

[0256]

[0257] In the formula, p is the distance parameter, with a value range of [1, 2], Σ is the covariance matrix, and V i is the index vector of evaluation object i.

[0258] This distance measure combines the flexibility of the Minkowski distance and the consideration of variable correlation in the Mahalanobis distance, making the distance measure more generalized and accurate. Here, the value range of the distance parameter p is between 1 and 2. In addition, the covariance matrix needs to be combined to fully consider the mutual relationship between each index, and finally form the generalized distance between the index vector of evaluation object i and the ideal solution.

[0259] Next, the TOPSIS algorithm introduces the relative closeness of the adjustment coefficient:

[0260]

[0261] The maximum negative ideal solution distance is corrected by the balance coefficient λ (with a value range of 0.4 to 0.6), making the closeness calculation more robust and adaptable.

[0262] In terms of the integration mechanism of the evaluation results, TOPSIS first performs pre - processing standardization:

[0263]

[0264] Introduce the IQR k (Interquartile Range), making the distribution of data more reasonable.

[0265] Then, in the generation of combined weights:

[0266]

[0267] Consider the r k (Average correlation coefficient) and CV k (Coefficient of variation) to ensure that the weight allocation of different evaluation methods is more balanced and reasonable.

[0268] Finally, TOPSIS comprehensively corrects the scores through CR k (Consistency ratio) and η (correction coefficient) (with a value between 0.1 and 0.2), generating the TOPSIS scores of each evaluation object:

[0269]

[0270] This score comprehensively considers the ideal solution distance, weights, and their consistency of various indicators, providing scientific decision-making support for decision-makers.

[0271] The TOPSIS evaluation algorithm ensures efficient evaluation under multi-dimensional indicators through the flexible application of dynamic ideal solutions, generalized distance measures, and combined weights, providing a solid methodological foundation for complex decision-making problems.

[0272] In this embodiment, constructing the Beta distribution density function in the theoretical distribution construction layer includes:

[0273] Construct the Beta distribution, and its formula is:

[0274]

[0275] where x is one of the AHP score, FCE score, and TOPSIS score, x ∈ [0, 1], and α, β are distribution parameters,

[0276]

[0277] Cumulative Beta distribution, and its formula is:

[0278]

[0279] Calculate the quantile, and its formula is:

[0280] Q(p) = F -1(p; α, β) = {x | F(x; α, β) = p}

[0281] Preferably, constructing a scoring position calculation model in the absolute level evaluation layer includes:

[0282] Probability density discretization, and its formula is:

[0283]

[0284] Calculating the density value sequence, and its formula is:

[0285] Y = {y i = f(x i ; α, β) | i = 0, 1,..., N - 1}

[0286] Calculating the scoring position, and its formula is:

[0287]

[0288] Calculating the quantified rating score, and its formula is:

[0289]

[0290] Preferably, constructing a development potential calculation model includes:

[0291] Calibrating the current position, and its formula is:

[0292] p c = P(x)

[0293] where p c is the current position;

[0294] Determining the target threshold, and its formula is:

[0295] T(p c ) = min{p ∈ {0.9, 0.75, 0.5, 0.25} | p > p c}

[0296] where T(p c ) is the target threshold;

[0297] Calculating the development distance, and its formula is:

[0298]

[0299] where Δ(x) is the development distance;

[0300] Calculating the development potential score, and its formula is:

[0301]

[0302] Wherein, D(x) is the development potential score.

[0303] Preferably, constructing the basic index calculation model in the comprehensive evaluation layer includes:

[0304] Calculating the basic ability score:

[0305]

[0306] Wherein, C b is the basic ability score, M = 3, s i = L(x i ) score is the quantitative rating score of the i-th method, and score is one of the AHP method, FCE method and TOPSIS method;

[0307] Calculating the development potential score:

[0308]

[0309] Wherein, C p is the development potential score, M = 3, D(x i ) is the development potential score at x i .

[0310] Calculating the balance score:

[0311]

[0312] Wherein, C e is the balance score,

[0313] Preferably, the formula for the comprehensive score in the comprehensive evaluation layer is:

[0314] S = w 1 C b + w 2 C p + w 3 C e

[0315] Wherein, w 1 = 0.5, w 2 = 0.3, w 3 = 0.2, C b is the basic ability score, C p is the development potential score, C e is the balance score.

[0316] Preferably, constructing the score feature analysis model in the feature analysis layer includes:

[0317] Calculating the score divergence, and its formula is:

[0318]

[0319] Wherein, σ s is the obtained divergence;

[0320] Calculate the coefficient of variation, and its formula is:

[0321]

[0322] Wherein, CV s is the coefficient of variation.

[0323] Preferably, constructing the feature analysis model in the feature analysis layer includes:

[0324] Calculate the potential mean value, and its formula is:

[0325]

[0326] Wherein, is the potential mean value;

[0327] Calculate the potential divergence, and its formula is:

[0328]

[0329] Wherein, σ D is the potential divergence;

[0330] Calculate the dominant development direction, and its formula is:

[0331] d * = argmax i {D(x i )}

[0332] Wherein, d * is the dominant direction.

[0333] In the above method, by constructing the Beta distribution density function and calculating the theoretical quantile, the intelligent grading of scores is realized, ensuring the objectivity and credibility of the evaluation results. At the same time, this method comprehensively considers the basic ability and development potential, allows the quantitative analysis of development potential, can comprehensively reflect the comprehensive development status within the region, and through detailed data feature analysis, helps to deeply understand the regional development differences, provides important support for policy formulation and resource allocation, and finally the provided data has high credibility and reference value.

[0334] Example 2

[0335] This embodiment provides a multi-dimensional evaluation system for the development capacity of regional pumped storage based on data feature modeling and Beta distribution, including: a data input module, a data preprocessing module, a feature extraction and weight assignment module, a comprehensive evaluation model construction module, an intelligent grading and result output module, a dynamic adjustment mechanism module, an intelligent grading and result output module, and a dynamic adjustment mechanism module;

[0336] The data input module is used to collect various indicators related to the development capacity of pumped storage, form original indicator data, and send it to the data preprocessing module;

[0337] The data preprocessing module includes a data cleaning unit, a data standardization unit, and a data fusion unit. The data cleaning unit is responsible for removing invalid or incorrect original indicator data; the data standardization unit converts the original indicator data from different sources into a unified format and unit; the data fusion unit is responsible for integrating various types of original indicator data into a consistent data set;

[0338] The feature extraction and weight assignment module is used to construct a judgment matrix based on the data set to obtain the weights of various indicators; different from traditional methods, the present invention does not rely on expert scoring to determine the weights, but automatically calculates the importance weights of various indicators through data analysis.

[0339] The comprehensive evaluation model construction module constructs a Beta distribution density function based on the AHP evaluation model, the FCE evaluation model, and the TOPSIS evaluation model to obtain an evaluation report;

[0340] The intelligent grading and result output module grades based on the evaluation report and generates charts from the evaluation report;

[0341] The dynamic adjustment mechanism module is used to update the original indicator data. This mechanism allows self-update according to the latest original indicator data and information, ensuring that the evaluation results always conform to the actual situation. This mechanism not only improves the timeliness of the evaluation but also enhances the adaptability and flexibility of the evaluation method.

[0342] Through the data input, preprocessing, and fusion modules, the system achieves effective organization and standardization of relevant metrics, ensuring data accuracy and consistency. Secondly, the feature extraction and weight assignment module objectively determines the importance of each metric by automatically analyzing the data rather than relying on expert scoring, improving the scientific nature of weight assignment. The comprehensive evaluation module combines the AHP, FCE, and TOPSIS models, makes full use of the Beta distribution to generate accurate evaluation reports, and, paired with intelligent grading and result output, intuitively displays the evaluation results. In addition, the dynamic adjustment mechanism ensures that the system can be updated in real time, always reflecting the latest situation, which not only enhances the timeliness and adaptability of the evaluation but also provides more reliable decision-making support for policymakers, contributing to the sustainable development of regional pumped storage energy.

[0343] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0344] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to the embodiments of the present invention. It should be understood that each flow and / or block in the flowchart and / or block diagram, as well as the combination of flows and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing devices generate means for implementing the functions specified in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0345] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including instruction means that implement the functions specified in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0346] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus, so that a series of operation steps are performed on the computer or other programmable apparatus to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable apparatus provide steps for implementing the functions specified in one process or a plurality of processes and / or boxes Figure 1 one process or a plurality of processes and / or boxes Figure 1 steps for implementing the functions specified in one box or a plurality of boxes.

[0347] The embodiments of the present invention have been described above in conjunction with the accompanying drawings. However, the present invention is not limited to the above specific embodiments. The above specific embodiments are merely illustrative and not restrictive. Under the inspiration of the present invention, those of ordinary skill in the art can also make many forms without departing from the spirit of the present invention and the scope protected by the claims. These all fall within the protection scope of the present invention.

Claims

1. A method for evaluating regional pumped storage development capacity based on data feature modeling and Beta distribution, characterized in that: include: Construct the scoring input layer of the scoring method, collect various indicators related to the regional pumped storage development capacity, and use the AHP method, FCE method and TOPSIS method to score each indicator to obtain the AHP score, FCE score and TOPSIS score; Construct a theoretical distribution construction layer, set specific distribution parameters for the AHP method, FCE method, and TOPSIS method, construct the Beta distribution density function, and generate quantiles; Construct an absolute level evaluation layer, and build a score position calculation model based on AHP score, FCE score and TOPSIS score to obtain the score position and quantitative rating score. Based on the score position, build a development potential calculation model to obtain the development potential score. Construct a comprehensive evaluation layer, and build a basic indicator calculation model based on the quantitative rating scores and development potential scores of various indicators under the AHP method, FCE method and TOPSIS method, to obtain the basic ability score, development potential score and balance score, and perform weighted calculations on them to obtain a comprehensive score; Construct a feature analysis layer, and construct a score feature analysis model based on the quantitative rating scores of each indicator under the AHP method, FCE method and TOPSIS method to obtain the score dispersion and coefficient of variation. Construct a feature analysis model based on the development potential scores of each indicator under the AHP method, FCE method and TOPSIS method to obtain the potential mean, potential dispersion and dominant development direction.

2. The method for evaluating regional pumped storage development capacity based on data feature modeling and Beta distribution according to claim 1 is characterized in that: The indicators in the scoring input layer of the scoring method include development demand (A1), resource endowment (A2), economic carrying capacity (A3) and construction constraint (A4).

3. The method for evaluating regional pumped storage development capacity based on data feature modeling and Beta distribution according to claim 1 is characterized in that: Building the theoretical distribution building layer includes: Step 1: Construct Beta distribution: Where x is one of the AHP score, FCE score and TOPSIS score, x∈[0,1], B(α,β) represents the Beta function, and α and β are distribution parameters; The specific distribution parameters set for the AHP method, the FCE method and the TOPSIS method include: AHP method: AHP (x) = f(x; 4, 2) FCE method: f FCE (x) = f(x; 3, 3) TOPSIS method: f TOPSIS (x) = f(x; 2, 4); Step 2: Quantile calculation Construct a cumulative distribution function, the formula of which is: Where F represents the cumulative distribution function; Calculate the quantile, the calculation formula of the quantile is: Q(p)=F -1 (p; α, β) = {x|F(x; α, β) = p} (3) where Q represents the quantile.

4. The method for evaluating regional pumped storage development capacity based on data feature modeling and Beta distribution according to claim 3 is characterized in that: Constructing the absolute level assessment layer includes: Step 1: For a given score x, determine the score position The probability density is discretized, dividing the [0,1] interval into N points. Based on the equally divided [0,1] interval, calculate the density value sequence: Y={y i =f(x i ;a,b)|i=0,1,...,N-1} (5) According to the density value sequence, the score position of the score x is determined: According to the score position, a quantitative rating score is determined by:

5. The method for evaluating regional pumped storage development capacity based on data feature modeling and Beta distribution according to claim 4 is characterized in that: The construction of the development potential calculation model includes: According to the score position, the current position is calibrated: p c =P(x) (8) In the formula, p c is the current location; Based on the current position, determine the target threshold: T(p c )=min{p∈{0.9,0.75,0.5,0.25}|p>p c } (9) Where T is the target threshold; According to the target threshold and the current position, the development distance is determined, and the calculation formula of the development distance is: In the formula, Δ is the development distance; According to the development distance, a development potential score is determined, and the calculation formula of the development potential score is: Where D is the development potential score.

6. The method for evaluating regional pumped storage development capacity based on data feature modeling and Beta distribution according to claim 5 is characterized in that: The basic indicator calculation model constructed in the comprehensive evaluation layer includes: Step 1: Build a basic indicator system Calculate the average of the quantitative rating scores of each scoring method as the basic ability score: Among them, C b For basic ability scoring, M = 3, s i =L(x i ) score is the quantitative rating score of the i-th method, and score is one of the AHP method, FCE method, and TOPSIS method; Calculate the average of the development potential scores of each scoring method to calculate the development potential score: In the formula, C p For development potential score, M = 3, D(x i ) is x i The development potential score of the place. Calculating the Balance Rating: Among them, C e For balance score, 7. The method for evaluating regional pumped storage development capacity based on data feature modeling and Beta distribution according to claim 6 is characterized in that: The formula for constructing the comprehensive score in the comprehensive evaluation layer is: S=w1C b +w2C p +w3C e (15) Among them, S represents the comprehensive score, w1=0.5, w2=0.3, w3=0.2, C b For basic ability score, C p Score the development potential, C e Score for balance.

8. The method for evaluating regional pumped storage development capacity based on data feature modeling and Beta distribution according to claim 6 is characterized in that: Constructing the score feature analysis model in the feature analysis layer includes: Calculate the score dispersion: In the formula, σ s is the score dispersion; Calculate the coefficient of variation using the formula: In the formula, CV s is the coefficient of variation.

9. The method for evaluating regional pumped storage development capacity based on data feature modeling and Beta distribution according to claim 6 is characterized in that: Constructing the feature analysis model in the feature analysis layer includes: The potential mean is calculated as follows: In the formula, is the potential mean; Calculate the potential dispersion, the formula is: In the formula, σ D is the potential dispersion; Calculate the dominant development direction, the formula is: d * =argnax i {D(x i )} (20) Where, d * As the leading direction.

10. A regional pumped storage development capacity evaluation system based on data feature modeling and Beta distribution, characterized in that: Used to execute the regional pumped storage development capacity assessment method based on data feature modeling and Beta distribution as described in claims 1-9.

Images