Multi-agent participation load management cost benefit quantitative evaluation method and system

Abstract

The invention discloses a multi-agent participation load management cost benefit quantitative evaluation method and system, and relates to the field of power system load management, and the method comprises the steps: carrying out the distributed photovoltaic cost benefit analysis according to the load management economic index characteristics, and constructing a cost benefit comprehensive evaluation model; based on the cost-benefit comprehensive evaluation model, performing quantitative evaluation on the load management cost-benefit under different subsidy standards and actual response capacities; and carrying out accuracy analysis on the evaluation result. According to the method, the operation cost of the power grid and the power consumption cost of the user are comprehensively considered, and the balance between power supply and user requirements is realized by establishing the cost-benefit evaluation model. Advanced mathematical analysis methods such as regression analysis and Bayesian classification are adopted, a load baseline calculation step and a clustering technology-based power consumer load classification method are introduced, and the performance of the model is further optimized.

Description

Technical Field

[0001] The present invention relates to the field of load management of electric power systems, and in particular to a method and system for quantitatively evaluating the cost-effectiveness of load management taking into account the participation of multiple subjects. Background Art

[0002] The existing cost-benefit evaluation model for power system load management has the following main defects and shortcomings:

[0003] 1. The model is too simplified: Existing models are usually based on a single load management strategy and cannot adapt to the complex situation in actual power grid operation, especially when multiple entities are involved. A single model often cannot fully consider the power grid operation cost and user electricity cost, and cannot optimize the balance between power grid supply and user demand.

[0004] 2. Insufficient accuracy: Most of the existing models do not take into account the complexity and randomness of the actual power system, such as load volatility and demand uncertainty. Therefore, in actual application, there may be problems such as large prediction errors and low model accuracy.

[0005] 3. Poor flexibility: Existing models usually fix the frequency and amplitude of load scheduling and cannot flexibly respond to changes in factors such as electricity market prices and load conditions. Therefore, they may not be able to make the best decision in some special circumstances.

[0006] 4. Simple considerations: Existing models often only consider the optimization of electricity costs when considering grid operation strategies, while ignoring many factors such as user demand response, grid system stability and security, resulting in poor application scope and practicality of the models.

[0007] The main technical problem to be solved by the present invention is how to accurately and effectively establish a quantitative evaluation model for the cost-effectiveness of load management by combining the grid operation cost and the user's electricity cost with the participation of multiple subjects.

[0008] First, the model needs to be able to adapt to different power market environments and different load management strategies, including load operation scale, demand response subsidy standards, and the number and magnitude of load dispatch. This requires the model to be flexible and adaptable enough to provide more accurate cost-benefit evaluation results in different situations and provide effective decision support for power grid operators and users.

[0009] Secondly, this model also needs to consider and deal with the constraints of load scheduling, which is a complex optimization problem that involves the stability and security of the power grid, and also affects the user's electricity costs and electricity demand.

[0010] Thirdly, the present invention also needs to solve how to introduce an incentive mechanism for demand response, which is not only an important part of the operation of the power market, but also a key factor affecting user participation and load management effects. The incentive mechanism for demand response needs to comprehensively consider factors such as the user's electricity demand, market prices, and the user's sensitivity to price fluctuations.

[0011] In summary, the technical solutions proposed in the present invention need to solve many complex problems, aiming to achieve effective operation and fair trading in the electricity market, improve the operating efficiency of the power system, and ensure the stability of the power system and the rights and interests of users. Summary of the invention

[0012] In view of the above-mentioned problems, the present invention is proposed.

[0013] Therefore, the problem to be solved by the present invention is: how to accurately and effectively establish a quantitative evaluation model for the cost-effectiveness of load management by combining the grid operation cost and the user's electricity cost with the participation of multiple subjects.

[0014] To solve the above technical problems, the present invention provides the following technical solutions: a quantitative evaluation method for the cost-benefit of load management taking into account the participation of multiple subjects, which includes: conducting a distributed photovoltaic cost-benefit analysis and constructing a comprehensive cost-benefit evaluation model based on the characteristics of load management economic indicators; based on the comprehensive cost-benefit evaluation model, quantitatively evaluating the cost-benefit of load management under different subsidy standards and actual response capacities; and analyzing the accuracy of the evaluation results.

[0015] As a preferred solution of the method for quantitatively evaluating the cost-benefit of load management taking into account multiple subjects involved in the present invention, the distributed photovoltaic cost-benefit analysis includes the construction of a distributed photovoltaic power generation cost model and a benefit model; the distributed photovoltaic power generation cost model is composed of the total cost of distributed photovoltaic power generation, the initial investment cost during the construction period, and the system operation and maintenance cost during the operation period, and is expressed as:

[0016] C=ω*(C 0 +C 1 )

[0017] Where C represents the total cost of distributed photovoltaic power generation, ω represents the installed capacity of distributed photovoltaic power generation projects, and C 0 Expressed as unit investment cost, C 1 Expressed as unit operation and maintenance cost.

[0018] As a preferred solution of the method for quantitatively evaluating the cost-effectiveness of load management taking into account multiple subjects involved in the present invention, the construction of the revenue model includes: the operation modes of distributed photovoltaics are full grid-connection and self-use, and the surplus grid-connection; when the operation mode is full grid-connection, the total revenue B totalEqual to the electricity sales revenue of the load management project B feed-in , the revenue model is expressed as,

[0019] B total =B feed-in

[0020] When the operation mode is self-use and the surplus is connected to the Internet, the revenue model is expressed as:

[0021] B total =B feed-in +B subsidy +B save

[0022] Among them, B subsidy Expressed as subsidy income, B save Expressed as saved electricity bill income.

[0023] As a preferred solution of the method for quantitatively evaluating the cost-effectiveness of load management taking into account the participation of multiple subjects in the present invention, the construction of the comprehensive cost-effectiveness evaluation model includes defining an objective function to minimize the sum of the grid operation cost and the user's electricity cost. The objective function is expressed as:

[0024] minC=C grid +C user

[0025] Among them, C grid Expressed as the grid operation cost, C user It is expressed as the user's electricity cost; the grid operation cost is calculated based on load scheduling and energy purchase cost, and the formula is expressed as follows:

[0026]

[0027] Among them, C t Expressed as the unit electricity cost in period t, L t It is expressed as the total load demand in time period t; the user's electricity cost is calculated based on electricity consumption, demand response subsidy and load dispatch strategy, and the formula is expressed as follows:

[0028]

[0029] Among them, P i Expressed as the unit electricity cost of user i, l i,t It is represented by the electricity consumption of user i in period t, s i Demand response subsidy standard, DR i,t It is expressed as the demand response of user i in time period t. In order to ensure the stability and safety of the power grid, constraints are set for load scheduling. The constraint formula is expressed as follows:

[0030]

[0031] Among them, L min,t and L max,t They are respectively represented as the minimum and maximum total load demands in period t.

[0032] As a preferred solution of the method for quantitatively evaluating the cost-effectiveness of load management taking into account the participation of multiple subjects in the present invention, the quantitative evaluation of the cost-effectiveness of load management under different subsidy standards and actual response capacities includes classifying the loads of power users based on clustering technology, using the shortest distance method to perform system clustering, and obtaining the cluster center as the original cluster center of the next step of fuzzy C-means clustering, and the shortest distance D between each class KL The calculation formula is,

[0033] D KL =min{d ij :p i G K ,p j G L}

[0034] Among them, G K , G L Represented as any two clusters, p i 、p j It is represented as the element in the corresponding cluster, which is the 96-point load data sequence of the user every day, d ij It is expressed as the distance between cluster elements;

[0035] For the distance between any two cluster elements, the calculation formula is expressed as:

[0036]

[0037] Among them, p i,t 、p j,t They are respectively represented as the elements in the i-th cluster and the j-th cluster at time t; after obtaining the cluster center through the system clustering method, the membership matrix u is determined according to the cluster center ij It is expressed as,

[0038]

[0039] dc ij =||c i -p j ||

[0040] Among them, m represents the number of clusters, d cij It is expressed as the distance between the i-th cluster center and the j-th cluster center, c i Represented as the center point of the i-th cluster, according to the shortest distance D KLDetermine, q is represented by a weighted index, 1∈[1,+∞); calculate the value function J as,

[0041]

[0042] If the current J value is less than the preset threshold, or the change in the value function value of the current J value relative to the previous iteration is less than the preset threshold, the iteration is stopped; if the system still needs to iterate, the new cluster center is calculated as follows:

[0043]

[0044] Re-iterate based on the new cluster centers.

[0045] As a preferred solution of the method for quantitatively evaluating the cost-effectiveness of load management taking into account the participation of multiple subjects in the present invention, wherein: the quantitative evaluation of the cost-effectiveness of load management under different subsidy standards and actual response capacities also includes calculating the baseline load, calculating the prior probability P(c i , i = 1, 2...m, then assume that P(c 1 )=P(c 2 )=……=P(c m ), otherwise P(c i )=|c i,D ∣ / ∣D∣;

[0046] Where D represents the set of all training data, |c i,D | represents the number of training tuples of the class in D, and |D| represents the total number of training tuples in D. Assuming that the categories are independent of each other, given the class label of the tuple, and the features are independent of each other, that is, there is no dependency between the features, then,

[0047]

[0048] Among them, X represents a specific data tuple, x k It is represented as the value of the kth feature of the tuple X; for each feature, it is examined whether the feature is a discrete value or a continuous value. If x k is a discrete value, then,

[0049]

[0050] If x k If is a continuous value, it is usually assumed that the feature value follows a Gaussian distribution, expressed as,

[0051]

[0052] in, It is represented as the mean value of each attribute of the class training tuple, Expressed as the standard deviation of the value of the kth attribute of the class training tuple; P(c i |X), then the baseline load on the forecast day is,

[0053]

[0054] Among them, c s (t) Predict the power load at time t on the day.

[0055] As a preferred solution of the method for quantitatively evaluating the cost-effectiveness of load management taking into account the participation of multiple subjects described in the present invention, the accuracy analysis of the evaluation results includes evaluating and correcting the model prediction values ​​using the mean absolute error and the mean square error.

[0056] Another object of the present invention is to provide a cost-effectiveness quantitative evaluation system for load management taking into account the participation of multiple subjects. This system can quantitatively evaluate the cost-effectiveness of load management by combining the grid operation cost and the user's electricity cost.

[0057] In order to solve the above technical problems, the present invention provides the following technical solutions: a system for quantitative evaluation of the cost-effectiveness of load management taking into account the participation of multiple subjects, comprising: an evaluation model construction module, an evaluation module and an analysis module; the evaluation model construction module performs distributed photovoltaic cost-benefit analysis according to the characteristics of load management economic indicators and constructs a comprehensive cost-effectiveness evaluation model; the evaluation module performs quantitative evaluation of the cost-effectiveness of load management under different subsidy standards and actual response capacities based on the comprehensive cost-effectiveness evaluation model;

[0058] The analysis module performs accuracy analysis on the evaluation results.

[0059] A computer device includes a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of the method for quantitatively evaluating the cost-effectiveness of load management taking into account the participation of multiple subjects are implemented.

[0060] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the method for quantitatively evaluating the cost-effectiveness of load management taking into account the participation of multiple agents.

[0061] The beneficial effects of the present invention are as follows: the present invention comprehensively considers the grid operation cost and the user's electricity cost, and achieves a balance between electricity supply and user demand by establishing a cost-effectiveness evaluation model.

[0062] The model of the present invention can adapt to different load management schemes, including load operation scale, demand response subsidy standard, and the number and amplitude of load dispatching, and has strong adaptability and flexibility.

[0063] The present invention adopts advanced mathematical analysis methods such as regression analysis and Bayesian classification, introduces a load baseline calculation step, and a power user load classification method based on clustering technology, so as to further optimize the performance of the model.

[0064] The model of the present invention provides a practical decision support tool for power grid operators and users, which can help them make more scientific and reasonable decisions. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. Among them:

[0066] Figure 1 This is a flow chart of the method for quantitatively evaluating the cost-effectiveness of load management taking into account the participation of multiple agents in Example 1.

[0067] Figure 2 This is a diagram of the baseline calculation and analysis steps of the method for quantitatively evaluating the cost-effectiveness of load management taking into account the participation of multiple agents in Example 1.

[0068] Figure 3 FIG. 1 is a diagram of the MAE simulation experiment curve (1) of the method for quantitatively evaluating the cost-effectiveness of load management taking into account multiple agents involved in Example 2.

[0069] Figure 4 FIG. 2 is a MAE simulation experiment curve (2) of the method for quantitatively evaluating the cost-effectiveness of load management taking into account multiple agents involved in Example 2.

[0070] Figure 5 FIG. 1 is a diagram showing the simulation results of the MAE cost-benefit evaluation of the cost-benefit quantitative evaluation method for load management taking into account multiple agents in Example 2 (1).

[0071] Figure 6 FIG. 2 is a simulation result of the MAE cost-benefit evaluation of the cost-benefit quantitative evaluation method for load management taking into account multiple agents involved in Example 2 (2). DETAILED DESCRIPTION

[0072] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0073] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0074] Example 1, reference Figure 1 and Figure 2 , which is the first embodiment of the present invention, and provides a method for quantitatively evaluating the cost-effectiveness of load management taking into account the participation of multiple subjects, including: Figure 1 As shown:

[0075] Step 1: Conduct a distributed photovoltaic cost-benefit analysis based on the characteristics of load management economic indicators and build a comprehensive cost-benefit evaluation model;

[0076] Distributed photovoltaic cost-benefit analysis includes the construction of distributed photovoltaic power generation cost model and benefit model; the distributed photovoltaic power generation cost model consists of the total cost of distributed photovoltaic power generation, the initial investment cost during the construction period, and the system operation and maintenance cost during the operation period, which is expressed as,

[0077] C=ω*(C 0 +C 1 )

[0078] Where C represents the total cost of distributed photovoltaic power generation, ω represents the installed capacity of distributed photovoltaic power generation projects, and C 0 Expressed as unit investment cost, C 1 Expressed as unit operation and maintenance cost.

[0079] Construction of revenue model: The operating modes of distributed photovoltaics mainly include "full access to the grid" and "self-generation for own use, surplus access to the grid". The composition of load management revenue is different in different operating modes. The load management revenue model is described as follows.

[0080] When the operation mode is full access to the Internet, the total revenue is B total Equal to the electricity sales revenue of the load management project B feed-in , the revenue model is expressed as,

[0081] B total =B feed-in

[0082] In this model, all power generated by the load management project is purchased by the grid company according to the benchmark photovoltaic grid-connected electricity price in the local resource area, and the total revenue is B total Equal to the electricity sales revenue of the load management project B feed-in , which is the annual power generation multiplied by the benchmark photovoltaic grid-connected electricity price.

[0083] When the operation mode is self-use and the surplus is connected to the Internet, the revenue model is expressed as:

[0084] B total =B feed-in +B subsidy +B save

[0085] Among them, B subsidy Expressed as subsidy income, B save Expressed as saved electricity bill income.

[0086] In this mode, the electricity revenue saved by users for self-generation and self-use is calculated according to the sales electricity price, that is, the electricity revenue saved is B savc It is equal to the self-consumption electricity multiplied by the sales electricity price. The remaining electricity is purchased by the power grid company according to the local coal-fired benchmark on-grid electricity price, and the sales income is B feed-in It is equal to the grid-connected electricity multiplied by the benchmark grid-connected electricity price for coal-fired power. The government implements a full electricity subsidy policy for users who adopt the "self-generation and self-use, surplus grid-connected" operation mode, that is, to subsidize all the power generated by the load management project, and the subsidy income B subsidy It is equal to the annual power generation multiplied by the load management quota subsidy standard.

[0087] This study adopted a comprehensive mathematical modeling approach when constructing an electricity cost-effectiveness optimization model for both the grid and user levels. First, an objective function was defined to minimize the sum of grid operating costs and user electricity costs in order to optimize the overall cost-effectiveness. The grid operating cost takes into account the load scheduling and energy purchase costs, while the user electricity cost is based on each user's electricity consumption, demand response subsidies, and load scheduling strategies. In addition, the model ensures the stability and security of the grid by introducing load scheduling constraints, while also taking into account the incentive mechanism of demand response. Through this model, the grid supply and user demand can be effectively balanced to achieve the optimization of electricity costs. The key to this method is that it can flexibly adapt to different load management schemes, including load operation scale, demand response subsidy standards, and the number and amplitude of load scheduling, thereby providing a practical decision support tool for grid operators and users.

[0088] Define an objective function to minimize the sum of grid operation cost and user electricity cost. The objective function is expressed as:

[0089] minC=C grid +C user

[0090] Among them, C grid Expressed as the grid operation cost, C user Expressed as the user's electricity cost;

[0091] The grid operation cost is calculated based on load dispatch and energy purchase cost, and the formula is:

[0092]

[0093] Among them, C t Expressed as the unit electricity cost in period t, L t It is expressed as the total load demand in time period t;

[0094] The user's electricity cost is calculated based on electricity consumption, demand response subsidies, and load dispatch strategy. The formula is:

[0095]

[0096] Among them, P i It is expressed as the unit electricity cost of user i, l i,t It is represented by the electricity consumption of user i in period t, s i Demand response subsidy standard, DR i,t It is expressed as the demand response amount of user i in time period t;

[0097] In order to ensure the stability and safety of the power grid, constraints are set for load scheduling. The constraint formula is expressed as follows:

[0098]

[0099] Among them, L min,t and L max,t They are respectively represented as the minimum and maximum total load demands in period t.

[0100] Step 2: Based on the comprehensive cost-effectiveness evaluation model, quantitatively evaluate the cost-effectiveness of load management under different subsidy standards and actual response capacities.

[0101] The core method of the original cost-effectiveness evaluation method is the industrial baseline load calculation method. The industrial load baseline analysis should adopt regression analysis, Bayesian classification and other methods. The calculation model should be analyzed in combination with important influencing factors such as the user's equipment characteristics, industry characteristics, production nature, economy and time. Regression analysis: a statistical analysis method to determine the quantitative relationship between two or more variables that are interdependent. According to the number of variables involved, it can be divided into univariate regression and multivariate regression analysis; according to the number of dependent variables, it can be divided into simple regression analysis and multiple regression analysis; according to the type of relationship between independent variables and dependent variables, it can be divided into linear regression analysis and nonlinear regression analysis. Correlation analysis studies whether phenomena are related, the direction and closeness of the correlation, and generally does not distinguish between independent variables or dependent variables. Regression analysis analyzes the specific form of correlation between phenomena, determines its causal relationship, and uses mathematical models to express its specific relationship. Bayesian classification: given a certain conditional probability, how to obtain the probability of two events after exchange, that is, how to obtain P(B|A) when P(A|B) is known. P(A|B) represents the probability of event A occurring under the premise that event B has already occurred, which is called the conditional probability of event A under the occurrence of event B. The steps for calculating the load baseline are as follows: Figure 2 shown.

[0102] Based on clustering technology, the power user load is classified, and the shortest distance method is used for system clustering. The cluster center is used as the original cluster center for the next step of fuzzy C-means clustering. The shortest distance D between each class is KL The calculation formula is,

[0103] D KL =min{d ij :p i G K ,p j G L}

[0104] Among them, G K , G L Represented as any two clusters, p i 、p j It is represented as the element in the corresponding cluster, which is the 96-point load data sequence of the user every day, d ij It is expressed as the distance between cluster elements;

[0105] For the distance between any two cluster elements, the calculation formula is expressed as:

[0106]

[0107] Among them, p i,t 、p j,t They are represented as the elements in the i-th cluster and the j-th cluster at time t respectively;

[0108] After obtaining the cluster center through the system clustering method, the membership matrix u is determined according to the cluster center ij It is expressed as,

[0109]

[0110] d cij =||c i -p j ||

[0111] Among them, m represents the number of clusters, d cij It is expressed as the distance between the i-th cluster center and the j-th cluster center, c i Represented as the center point of the i-th cluster, according to the shortest distance D KL Determine, q is expressed as a weighted index, q∈[1,+∞);

[0112] The computational value function J is expressed as,

[0113]

[0114] If the current J value is less than the preset threshold, or the change in the value function value of the current J value relative to the previous iteration is less than the preset threshold, the iteration is stopped;

[0115] If the system still needs iteration, the new cluster center is calculated as:

[0116]

[0117] Re-iterate according to the new cluster center. Through the above three steps of processing load big data, the power consumption behavior patterns of power users can be divided into m categories, and the result is the basis for calculating the predicted daily baseline load.

[0118] The baseline load calculation method is as follows: calculate the prior probability P(c i ), i = 1, 2...m, then assume that P(c 1 )=P(c 2 )=……=P(c m ), otherwise P(c i )=|c i,D ∣ / ∣D∣;

[0119] Where D represents the set of all training data, |c i,D | represents the number of training tuples of the class in D, and |D| represents the total number of training tuples in D;

[0120] Assuming that the categories are independent of each other, given the class label of the tuple, and the features are independent of each other, that is, there is no dependency between the features, then,

[0121]

[0122] Among them, X represents a specific data tuple, x k Represented as the value of the kth feature of tuple X;

[0123] For each feature quantity, examine whether the feature quantity is discrete or continuous. If x k is a discrete value, then,

[0124]

[0125] If x k If is a continuous value, it is usually assumed that the feature value follows a Gaussian distribution, expressed as,

[0126]

[0127] in, It is represented as the mean value of each attribute of the class training tuple, It is expressed as the standard deviation of the value of the kth attribute of the class training tuple;

[0128] According to Bayes' theorem, P(c i |X), then the baseline load on the forecast day is,

[0129]

[0130] Among them, c s (t) Forecast the power load at time t on the day

[0131] Step 3: Analyze the accuracy of the evaluation results. Use the mean absolute error and mean square error to evaluate and correct the model prediction value.

[0132] Example 2, reference Figure 3-Figure 6 , which is the second embodiment of the present invention, and is different from the first embodiment in that: the quantitative evaluation method for cost-effectiveness of load management taking into account the participation of multiple subjects also includes, in order to verify and illustrate the technical effects adopted in this method, this embodiment compares the test results with scientific argumentation to verify the real effect of this method.

[0133] The mean absolute error (MeanAbsoluteError) is the sample standard deviation of the difference between the predicted value and the observed value (called residual). The root mean square error is used to illustrate the degree of dispersion of the sample. It refers to the average value of the distance between the model predicted value f(x) and the true value y of the sample. Its formula is as follows:

[0134]

[0135] To simplify the discussion, we ignore the subscript i, m = 1, and use yf(x) as the horizontal coordinate and MAE as the vertical coordinate to draw the graph of its loss function as shown in Figure 3 As shown:

[0136] Intuitively, the MAE curve is V-shaped, continuous but not differentiable at yf(x) = 0, and it is difficult for a computer to solve the derivative. Moreover, the gradient of MAE is equal in most cases, which means that even for small loss values, its gradient is large. This is not conducive to function convergence and model learning.

[0137] It is worth mentioning that one advantage of MAE over MSE is that it is less sensitive to outliers and more tolerant. Because MAE calculates the absolute value of the error yf(x), whether yf(x)>1 or yf(x)<1, there is no square term, the penalty intensity is the same, and the weight is the same.

[0138] The mean square error (MSE) is an important indicator that reflects the degree of match between the predicted value and the true value of the regression model. It mainly includes the average value of the square of the distance between the model predicted value f(x) and the true value y of the sample. Its formula is as follows:

[0139]

[0140] Among them, yi and f(xi) represent the true value and predicted value of the i-th sample respectively, and M is the number of samples.

[0141] To simplify the discussion, we ignore the subscript i, m = 1, and use yf(x) as the horizontal coordinate and MSE as the vertical coordinate to draw the graph of its loss function as shown in Figure 4 As shown:

[0142] The MSE curve is characterized by smoothness, continuity, and differentiability, which makes it easy to use the gradient descent algorithm. It is a commonly used loss function. Moreover, as the error decreases, the gradient of MSE also decreases, which is conducive to the convergence of the function. Even if the learning factor is fixed, the function can quickly reach the minimum value. The square error has a characteristic that when the difference between yi and f(xi) is greater than 1, the error will increase; when the difference between yi and f(xi) is less than 1, the error will decrease. This is determined by the characteristics of the square. In other words, MSE will give a greater penalty to cases with larger errors (>1) and a smaller penalty to cases with smaller errors (<1). From a training perspective, the model will be more inclined to punish points with larger penalties and give them greater weights.

[0143] like Figure 5As shown in the figure, if there are outliers in the sample, MSE will give higher weight to the outliers, but at the expense of the prediction effect of other normal data points, which will eventually reduce the overall performance of the model. We propose to adopt the peak-valley management benefit evaluation method and introduce the concept of different scenarios. The difference from the original cost-effectiveness evaluation method is that each day is refined into peak and valley hours, thereby reducing MAE and MSE. Compared with the original cost-effectiveness evaluation method, it improves MAE and MSE by 22% and 21% respectively.

[0144] The horizontal axis is the difference between the model prediction value and the true value of the sample every year since June 2023, and the vertical axis is the MAE value. The blue line is the original cost-benefit evaluation, and the red line is after adopting our cost-benefit evaluation method. After adopting our cost-benefit analysis method, the MAE value is reduced by an average of 22%.

[0145] like Figure 6 As shown, the horizontal axis is the square of the difference between the model prediction value and the true value of the sample each year from June 2022 to June 2023, and the vertical axis is the MSE value. The blue line is the original cost-effectiveness evaluation, and the red line is after adopting our cost-effectiveness evaluation method. After adopting our cost-effectiveness analysis method, the MSE value is reduced by an average of 21%.

[0146] The present invention utilizes the peak-valley management benefit evaluation method and introduces the concept of different scenarios. Based on the original cost-benefit evaluation method, it successfully reduces the values ​​of Mean Absolute Error (MAE) and Mean Squared Error (MSE) by 22% and 21% respectively, significantly improving the accuracy and reliability of the model.

[0147] Embodiment 3 is the third embodiment of the present invention, which is different from the previous two embodiments in that: a system for quantitative evaluation of the cost-effectiveness of load management taking into account the participation of multiple subjects, including an evaluation model construction module, an evaluation module and an analysis module; the evaluation model construction module conducts distributed photovoltaic cost-benefit analysis according to the characteristics of load management economic indicators, and constructs a comprehensive cost-effectiveness evaluation model; the evaluation module conducts quantitative evaluation of the cost-effectiveness of load management under different subsidy standards and actual response capacities based on the comprehensive cost-effectiveness evaluation model; the analysis module performs accuracy analysis on the evaluation results.

[0148] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program codes.

[0149] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0150] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0151] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, a plurality of steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0152] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A quantitative evaluation method for the cost-effectiveness of load management taking into account the participation of multiple entities. Features: include, According to the characteristics of load management economic indicators, conduct distributed photovoltaic cost-benefit analysis and build a comprehensive cost-benefit evaluation model; Based on the comprehensive cost-effectiveness evaluation model, the cost-effectiveness of load management under different subsidy standards and actual response capacities is quantitatively evaluated; Conduct accuracy analysis on the evaluation results.

2. The method for quantitatively evaluating the cost-effectiveness of load management taking into account the participation of multiple agents as claimed in claim 1, Features: The distributed photovoltaic cost-benefit analysis includes the construction of a distributed photovoltaic power generation cost model and a benefit model; The distributed photovoltaic power generation cost model consists of the total cost of distributed photovoltaic power generation, the initial investment cost during the construction period, and the system operation and maintenance cost during the operation period, which is expressed as: C=ω*(C 0 +C 1 ) Where C represents the total cost of distributed photovoltaic power generation, ω represents the installed capacity of distributed photovoltaic power generation projects, and C 0 Expressed as unit investment cost, C 1 Expressed as unit operation and maintenance cost.

3. The method for quantitatively evaluating the cost-effectiveness of load management taking into account the participation of multiple agents as claimed in claim 2, Features: The construction of the revenue model includes the following: the operation modes of distributed photovoltaic power generation include full grid access and self-use, and the surplus grid access; When the operation mode is full access to the Internet, the total revenue is B total Equal to the electricity sales revenue of the load management project B feed-in , the revenue model is expressed as, B total =B feed-in When the operation mode is self-use and the surplus is connected to the Internet, the revenue model is expressed as: B total =B feed-in +B subsidy +B save Among them, B subsidy Expressed as subsidy income, B save Expressed as saved electricity bill income.

4. The method for quantitatively evaluating the cost-effectiveness of load management taking into account the participation of multiple agents as claimed in claim 3, Features: The construction of the comprehensive cost-effectiveness evaluation model includes defining an objective function to minimize the sum of the grid operation cost and the user's electricity cost. The objective function is expressed as: my C=C grid +C user Among them, C grid Expressed as the grid operation cost, C user Expressed as the user's electricity cost; The grid operation cost is calculated based on load dispatch and energy purchase cost, and the formula is: Among them, C t Expressed as the unit electricity cost in period t, L t It is expressed as the total load demand in time period t; The user's electricity cost is calculated based on electricity consumption, demand response subsidies, and load dispatch strategy. The formula is: Among them, P i Expressed as the unit electricity cost of user i, l i,t It is represented by the electricity consumption of user i in period t, s i Demand response subsidy standard, DR i,t It is expressed as the demand response amount of user i in time period t; In order to ensure the stability and safety of the power grid, constraints are set for load scheduling. The constraint formula is expressed as follows: Among them, L min,t and L max,t They are respectively represented as the minimum and maximum total load demands in period t.

5. The method for quantitatively evaluating the cost-effectiveness of load management taking into account the participation of multiple agents as claimed in claim 4, Features: The quantitative evaluation of the cost-effectiveness of load management under different subsidy standards and actual response capacities includes classifying the loads of power users based on clustering technology, using the shortest distance method to perform system clustering, and obtaining the cluster center as the original cluster center of the next fuzzy C-means clustering. The shortest distance D between each class is KL The calculation formula is, D KL =min{d ij :p i G K ,p j G L } Among them, G K , G L Represented as any two clusters, p i 、p j It is represented as the element in the corresponding cluster, which is the 96-point load data sequence of the user every day, d ij It is expressed as the distance between cluster elements; For the distance between any two cluster elements, the calculation formula is expressed as: Among them, p i,t 、p j,t They are represented as the elements in the i-th cluster and the j-th cluster at time t respectively; After obtaining the cluster center through the system clustering method, the membership matrix u is determined according to the cluster center ij It is expressed as, dc ij =||c i -p j || Among them, m represents the number of clusters, d cij It is expressed as the distance between the i-th cluster center and the j-th cluster center, c i Represented as the center point of the i-th cluster, according to the shortest distance D KL Determine, q is expressed as a weighted index, q∈[1,+∞); The computational value function J is expressed as, If the current J value is less than the preset threshold, or the change in the value function value of the current J value relative to the previous iteration is less than the preset threshold, the iteration is stopped; If the system still needs iteration, the new cluster center is calculated as: Re-iterate based on the new cluster centers.

6. The method for quantitatively evaluating the cost-effectiveness of load management taking into account the participation of multiple agents as claimed in claim 5, Features: The quantitative evaluation of the cost-effectiveness of load management under different subsidy standards and actual response capacity also includes calculating the baseline load, including: Calculate the prior probability P(c i ), i = 1, 2...m, then assume that P(c 1 )=P(c 2 )=……=P(c m ), otherwise P(c i )=|c i,D ∣ / ∣D∣; Where D represents the set of all training data, |c i,D | represents the number of training tuples of the class in D, and |D| represents the total number of training tuples in D; Assuming that the categories are independent of each other, given the class label of the tuple, and the features are independent of each other, that is, there is no dependency between the features, then, Among them, X represents a specific data tuple, x k Represented as the value of the kth feature of tuple X; For each feature quantity, examine whether the feature quantity is discrete or continuous. If x k is a discrete value, then, If x k If is a continuous value, it is usually assumed that the feature value follows a Gaussian distribution, expressed as, in, Represented as the mean of the values ​​of each attribute of the class training tuple, σ Ci It is expressed as the standard deviation of the value of the kth attribute of the class training tuple; According to Bayes' theorem, P(c i |X), then the baseline load on the forecast day is, Among them, c s (t) Predict the power load at time t on the day.

7. The method for quantitatively evaluating the cost-effectiveness of load management taking into account the participation of multiple agents as claimed in claim 6, Features: The accuracy analysis of the evaluation results includes evaluating and correcting the model prediction values ​​using mean absolute error and mean square error.

8. A system using the method for quantitatively evaluating the cost-effectiveness of load management taking into account the participation of multiple agents as described in any one of claims 1 to 7, Features: It includes evaluation model building module, evaluation module and analysis module; The evaluation model building module performs distributed photovoltaic cost-benefit analysis according to the characteristics of load management economic indicators and builds a comprehensive cost-benefit evaluation model; The evaluation module quantitatively evaluates the cost-effectiveness of load management under different subsidy standards and actual response capacities based on a comprehensive cost-effectiveness evaluation model; The analysis module performs accuracy analysis on the evaluation results.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program. Features: When the processor executes the computer program, the steps of the method for quantitatively evaluating the cost-effectiveness of load management taking into account the participation of multiple agents are implemented as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, Features: When the computer program is executed by a processor, the steps of the method for quantitatively evaluating the cost-effectiveness of load management taking into account the participation of multiple agents are implemented as described in any one of claims 1 to 7.

Images