Method and system for determining a new energy consumption model based on lasso regression

Abstract

The present application relates to a kind of based on LASSO regression determination new energy consumption model method and system, the method includes: constructing new energy consumption model;LASSO regression is calculated using to the new energy consumption model, and it is judged whether there is mathematical relationship between each factor: if yes, then the multi-dimensional model is represented by factor function relationship formula;If no, then the multi-dimensional model is represented by Peaks function.The multi-dimensional new energy response surface model based on LASSO regression of the present application can be graphically depicted that new energy utilization rate changes with factor, and it is of great significance to study the influence of multiple uncertain factors on new energy utilization rate.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of new energy evaluation, in particular to a method and system for determining a new energy consumption model based on LASSO regression. BACKGROUND

[0002] Energy is the basis for sustainable economic and social development, and is an indispensable power guarantee for human production and life. With the increasingly prominent problems of energy security, ecological environment and climate change, accelerating the development of new energy has become a universal consensus and concerted action of the international community to promote energy transformation and development and respond to global climate change.

[0003] China actively promotes the development of new energy. By the end of 2015, China's cumulative installed capacity of new energy reached 171.48GW, ranking among the world's top. Due to the randomness and volatility of wind and light new energy output, and because of the endowment of the reverse distribution of new energy resources in China and the complex power system and market mechanism, the new energy consumption problem faces greater challenges. The problem of abandoned wind and light in China's three northern regions is more prominent. In order to solve this situation, a new energy consumption model that conforms to reality must be constructed. The traditional method can only determine the corresponding new energy consumption capacity according to the change of a single factor. The new energy output model has not yet been clearly understood under the condition of simultaneous change of multiple factors. Multidimensional space refers to a space composed of four or more dimensions. The definition of "dimension" is a measure. In three-dimensional space coordinates, plus time, time and space are related to each other, forming a four-dimensional space-time.

[0004] Currently, the methods for determining the new energy consumption capacity mainly include determining the new energy consumption model, using the power grid planning model, and the time sequence production simulation method. These methods mostly rely on mathematical methods, and it is difficult to obtain a clear and optimal solution for the final optimization result. Even if the optimal solution is obtained, it is also difficult to obtain the optimization interval. SUMMARY

[0005] The purpose of the present application is to provide a method and system for determining a new energy consumption model based on LASSO regression.

[0006] To achieve the above purpose, the present application provides the following solutions:

[0007] A method for determining a new energy consumption model based on LASSO regression, comprising:

[0008] constructing a new energy consumption model;

[0009] using LASSO regression to perform dimensionality reduction calculation on the new energy consumption model, and determining whether there is a mathematical relationship between the factors:

[0010] if yes, then representing the multidimensional model through a factor function relationship.

[0011] If not, the Peaks function is used to represent the multi-dimensional model.

[0012] Preferably, the new energy consumption model is constructed, comprising:

[0013] The minimum new energy curtailment is taken as the objective function.

[0014] A constraint condition set is constructed; the constraint condition set includes new energy power curtailment constraint, main transformer capacity constraint, active power balance constraint, conventional unit output constraint, conventional unit ramping constraint, system positive reserve constraint, system negative reserve constraint and line flow constraint.

[0015] The new energy consumption model is determined according to the objective function and the constraint condition set.

[0016] Preferably, the LASSO regression is used to perform dimensionality reduction calculation on the new energy consumption model, comprising:

[0017] The new energy consumption model is simulated to obtain basic data.

[0018] The basic data is subjected to regularization processing to obtain processed data.

[0019] According to the processed data, the influence characteristics of multiple uncertainty factors on new energy consumption capacity are determined.

[0020] Based on the influence characteristics, the LASSO regression is used to obtain a new energy consumption capacity model suitable for multi-dimensional space theory.

[0021] Preferably, the multi-dimensional model is represented by a factor function relationship, comprising:

[0022] The least square method is used for polynomial fitting, and the cftool command of MATLAB is used to make the residual error 0 and the sum of squared errors minimum, so as to obtain the optimal matching function.

[0023] According to the optimal matching function, a first three-dimensional model between variables is constructed; the variables include installed capacity, power load and resource hours; and the first three-dimensional model is used to represent the multi-dimensional factor new energy consumption capacity model through color dimension.

[0024] Preferably, the Peaks function is used to represent the multi-dimensional model, comprising:

[0025] The new energy consumption model is plotted on MATLAB to obtain a density map.

[0026] The normality of the density map is checked by using a Kolmogorov-Smirnov test, and a new energy consumption model conforming to a normal distribution is obtained;

[0027] The new energy consumption model conforming to the normal distribution is fitted by using a Peaks function as a representative normal function.

[0028] A system for determining a new energy consumption model based on LASSO regression includes:

[0029] A construction module is configured to construct a new energy consumption model.

[0030] A judgment module is configured to perform dimension reduction calculation on the new energy consumption model by using LASSO regression, and determine whether there is a mathematical relationship between factors:

[0031] A first representation module is configured to represent a multi-dimensional model by a factor function relationship when the result of the judgment module is yes.

[0032] A second representation module is configured to represent a multi-dimensional model by using a Peaks function when the result of the judgment module is no.

[0033] Preferably, the construction module specifically includes:

[0034] A target function unit is configured to take the minimum amount of new energy curtailment as a target function.

[0035] A constraint condition unit is configured to construct a constraint condition set; the constraint condition set includes a new energy curtailment constraint, a main transformer capacity constraint, an active power balance constraint, a conventional unit output constraint, a conventional unit ramping constraint, a system positive reserve constraint, a system negative reserve constraint, and a line power flow constraint.

[0036] A model determination unit is configured to determine the new energy consumption model according to the target function and the constraint condition set.

[0037] Preferably, the judgment module specifically includes:

[0038] A simulation unit is configured to simulate the new energy consumption model to obtain basic data.

[0039] A preprocessing unit is configured to perform regularization processing on the basic data to obtain processed data.

[0040] A feature determination unit is configured to determine the influence characteristics of multiple uncertainty factors on new energy consumption capacity according to the processed data.

[0041] A model regression unit is configured to obtain a new energy consumption capacity model suitable for a multi-dimensional space theory by using LASSO regression based on the influence characteristics.

[0042] Preferably, the first characterization module specifically comprises:

[0043] a first fitting unit, configured to perform polynomial fitting by using a least square method, and make residual error 0 and the sum of square errors minimum by using a cftool command of MATLAB, to obtain an optimal matching function;

[0044] a first three-dimensional model between variables is constructed according to the optimal matching function; the variables include installed capacity, power load and resource hours; and the first three-dimensional model is used to represent a multi-dimensional factor new energy consumption capacity model by color dimension.

[0045] Preferably, the second characterization module specifically comprises:

[0046] a plotting unit, configured to plot a density map on MATLAB by using the new energy consumption model;

[0047] a verification unit, configured to verify normality of the density map by using a Kolmogorov-Smirnov test, to obtain a new energy consumption model conforming to normal distribution;

[0048] a second fitting unit, configured to fit the new energy consumption model conforming to normal distribution by using a Peaks function as a representative of a normal function.

[0049] According to the specific embodiments provided by the present application, the following technical effects are disclosed:

[0050] The present application provides a method and system for determining a new energy consumption model based on LASSO regression, the method comprising: constructing a new energy consumption model; performing dimension reduction calculation on the new energy consumption model by using LASSO regression, and determining whether there is a mathematical relationship between factors; if yes, representing a multi-dimensional model by a factor function relationship; and if no, representing a multi-dimensional model by a Peaks function. The multi-dimensional new energy response surface model based on LASSO regression can visually depict the situation of new energy utilization rate changing with factors, and has important significance for studying the influence of multiple uncertain factors on new energy utilization rate. BRIEF DESCRIPTION OF DRAWINGS

[0051] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed in the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0052] Figure 1A flowchart of the method for determining a new energy consumption model based on LASSO regression in the embodiment provided by the present application;

[0053] Figure 2 A schematic diagram of the overall scheme structure in the embodiment provided by the present application;

[0054] Figure 3 A schematic diagram of the LASSO regression model in the embodiment provided by the present application;

[0055] Figure 4 A schematic diagram of the peaks function multidimensional model in the embodiment provided by the present application;

[0056] Figure 5 A schematic diagram of the Peaks function-based multidimensional model example verification in the embodiment provided by the present application;

[0057] Figure 6 A schematic diagram of the factor function multidimensional model in the embodiment provided by the present application;

[0058] Figure 7 A schematic diagram of the factor function-based multidimensional model example verification in the embodiment provided by the present application. DETAILED DESCRIPTION

[0059] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0060] There are currently three common methods for determining new energy consumption capacity:

[0061] 1. Low-dimensional function mapping method

[0062] Mapping a two-dimensional model to a high-dimensional model is a common method for constructing a three-dimensional model. The kernel function under the support vector machine is used to map the two-dimensional model to a three-dimensional space. The support vector machine is used to solve the problem that data is not easy to linearly divide in low dimensions. A certain nonlinear transformation is used to map the input space to a high-dimensional feature space. The kernel function is introduced to reduce the data complexity and prevent dimension disaster.

[0063] 2. Power grid planning model method

[0064] Divide a year into 12 units, and each unit has N i day load curves. The simulated time-series output is established in the output curve library according to the unit. There are N jThe daily power output curve, the daily load curve in each unit combined with the daily power output curve in the daily power output curve library of the corresponding unit, and these power-load combinations constitute the annual consumption capacity determination set Ω; in the annual consumption capacity determination set Ω, the nth load-power combination is composed of the jth load curve of the ith unit and the kth daily power output curve in the power output curve library of the ith unit; according to the annual wind power consumption capacity determination set Ω, the determination of the wind power consumption capacity under the transmission capacity constraint is performed. When the curtailed power of the nth load-power combination , it indicates that the nth load-power combination passes the network transmission capacity constraint; when the curtailed wind power of the nth load-power combination , it indicates that the nth load-power combination does not pass the network transmission capacity constraint. It is determined whether the calculation of all load-power combinations in the annual consumption capacity determination set Ω is completed, and if so, the wind power consumption ratio λ grid under the network transmission capacity constraint is obtained. grid The determination of the wind power consumption capacity under the network transmission capacity constraint is as follows:

[0065]

[0066] 3. Time series production simulation method

[0067] The new energy time series production simulation model is based on the time series production simulation method, and under the constraints of supply and demand balance at each time section and other power system operation boundary conditions, the maximum new energy consumption capacity is taken as the objective function to simulate the operation status of each generator unit and obtain optimal indicators. For this mixed integer programming problem, the GAMS solver is used for solving.

[0068] The deficiencies of the above three methods are as follows:

[0069] 1. Low-dimensional function mapping method

[0070] (1) For complex two-dimensional functions, it is very difficult to map from low dimension to high dimension, and it is necessary to ensure that the synthesized multi-dimensional function has a real physical meaning.

[0071] (2) The calculation speed is slow, overfitting phenomenon is easy to occur, and the interpretability is poor.

[0072] 2. Power grid planning simulation method

[0073] (1) It is more dependent on basic data, and the current power grid updates at a faster speed, so many new data cannot be timely counted.

[0074] (2) Due to the involvement of many other factors in power grid construction, it is difficult for the current algorithm to solve some uncertain factors.

[0075] 3. Time series production simulation method

[0076] (1) The future mutation situation must be considered in the process of production simulation, which may not be the same as the historical regular trend.

[0077] (2) The analysis speed is slow, and some key factors may be forgotten in the analysis process, and not all variables are sensitive to the analysis path.

[0078] Figure 1 The flowchart of the method for determining the new energy consumption model based on LASSO regression in the embodiment provided by the present application is shown in Figure 1 The present application aims to provide a method for determining a new energy consumption model based on LASSO regression, which comprises:

[0079] Step 100: Construct a new energy consumption model;

[0080] Step 200: Perform dimensionality reduction calculation on the new energy consumption model by using LASSO regression, and determine whether there is a mathematical relationship between factors:

[0081] Step 300: If yes, represent the multi-dimensional model through a factor function relationship;

[0082] Step 301: If no, represent the multi-dimensional model by using Peaks function.

[0083] Figure 2 The overall scheme structure diagram in the embodiment provided by the present application is shown in Figure 2 When using the method, a new energy consumption model is first constructed, with the minimum new energy curtailment as the objective function, and with new energy curtailment, limited main transformer, active power balance, conventional unit output, conventional unit climbing, system positive reserve, system negative reserve, and line power flow as the constraint conditions. Then, dimensionality reduction calculation is performed on the new energy consumption model by using LASSO regression, and whether there is a mathematical relationship between factors is considered. If there is a mathematical relationship between factors, polynomial fitting is considered by using the least square method, and the residual error is made to be 0 and the sum of square errors is made to be minimum by using the cftool command of MATLAB, so as to construct a three-dimensional model between variables. If there is no mathematical relationship between variables, normality test is performed by using MATLAB, and Peaks function is used to represent the new energy consumption capacity by randomly selecting new energy factors.

[0084] Preferably, the step 100 comprises:

[0085] The minimum new energy curtailment is taken as the objective function;

[0086] A constraint condition set is constructed; the constraint condition set comprises a new energy power cut constraint, a main transformer capacity limited constraint, an active power balance constraint, a conventional unit output constraint, a conventional unit ramping constraint, a system positive reserve constraint, a system negative reserve constraint, and a line power flow constraint;

[0087] The new energy consumption model is determined according to the objective function and the constraint condition set.

[0088] Further, the steps of constructing the multi-dimensional new energy consumption model comprise the following:

[0089] The objective function is to minimize the new energy curtailment amount, and the formula is:

[0090]

[0091] In the formula, N w is the number of new energy stations connected to the system; T represents the total length of the dispatching time; is the power cut of the new energy station i at the time t.

[0092] In the method, the formula for defining the new energy consumption space is:

[0093]

[0094] In the formula, E a is the new energy consumption space; E1 is the load amount; E t,D is the planned power transmission amount (positive for transmission); β is the average peak shaving depth of the conventional unit in the system; λ 1+D is the load rate of the equivalent load (considering the load and the power transmission); T is a new energy consumption period; R + is the positive reserve capacity after considering the participation of the new energy in balancing.

[0095] The constraint conditions are as follows:

[0096] (1) New energy power cut constraint

[0097]

[0098] t = 1, 2,..., T, is the power prediction value of the new energy station i at the time t.

[0099] (2) Main transformer capacity limited constraint

[0100]

[0101] (3) Active power balance constraint

[0102]

[0103] is the output planning value of the conventional unit k at time t; N g is the number of conventional units; N T is the number of main transformers in the system connected to the new energy station; N L is the number of bus loads.

[0104] (4) Conventional unit output constraint

[0105]

[0106] and are the maximum and minimum technical output of the conventional unit k, respectively

[0107] (5) Conventional unit ramp constraint

[0108]

[0109] and are the maximum up-ramp and down-ramp capabilities of the conventional unit k within 1 min, respectively; T x is the time interval.

[0110] (6) System positive reserve constraint

[0111]

[0112] (7) System negative reserve constraint

[0113]

[0114] and are the minimum requirements of the system for positive spinning reserve and negative spinning reserve at time t, respectively.

[0115] (8) Line power flow constraint

[0116]

[0117] is the power flow value of line n at time t.

[0118] Preferably, the dimensionality reduction calculation of the new energy consumption model by using LASSO regression comprises:

[0119] simulate the new energy consumption model to obtain basic data;

[0120] regularize the basic data to obtain processed data;

[0121] Determine the influence characteristics of the new energy consumption capacity by the plurality of uncertain factors according to the processed data;

[0122] Based on the influence characteristics, a new energy consumption capacity model suitable for a multi-dimensional space theory is obtained by using LASSO regression.

[0123] Specifically, LASSO is a compression estimate, which obtains a more refined model by constructing a penalty function, so that it compresses some regression coefficients, that is, forces the sum of the absolute values of the coefficients to be less than a fixed value; at the same time, some regression coefficients are set to zero. Therefore, it retains the advantage of subset shrinkage and is a biased estimate for handling data with multicollinearity.

[0124] The new energy consumption capacity is restricted by multiple factors. With the growth of data dimensions, the number of samples in the analysis space will increase exponentially. In this case, analyzing the problem will make the data very sparse and easily lead to the curse of dimensionality, making prediction very difficult. At the same time, it will also lead to data overfitting, so that the conclusion cannot be generalized to new research samples. Therefore, regularization is used to process the data. In regularization, all feature variables will be retained, but the order of magnitude of the feature variables will be reduced. Based on the influence characteristics of the new energy consumption capacity by the plurality of uncertain factors, a new energy consumption capacity model suitable for a multi-dimensional space theory is obtained by using LASSO regression.

[0125]

[0126] is a regularization term, and λ is a regularization parameter. θ (x) is a response variable, and y is a covariate

[0127] The gradient descent method is used to minimize the loss function to find the parameters of the objective function

[0128]

[0129] X i is an influencing factor, and F is a new energy utilization rate.

[0130] Preferably, the multi-dimensional model is represented by a factor function relationship, which comprises:

[0131] The least squares method is used for polynomial fitting, and the cftool command of MATLAB is used to make the residual error 0 and minimize the sum of squared errors to obtain the optimal matching function;

[0132] According to the optimal matching function, a first three-dimensional model between variables is constructed; the variables include installed capacity, power load and resource hours; and the first three-dimensional model is used to represent the multi-dimensional factor new energy consumption capacity model by color dimension.

[0133] Optionally, when the historical data is sufficient, the relationship between factors can be determined, so the least square fitting method is used, which finds the best function matching of data by minimizing the sum of squares of errors. Using the least square method can easily obtain unknown data, and make the sum of squares of errors between the obtained data and the actual data minimum. Using the cftool command of MATLAB to fit the function makes the residual error 0, the average value of the fitted value equal to the average value of the original data and the sum of squares of errors minimum. A three-dimensional model between installed capacity, power load and resource hours is constructed, and the color dimension is used to represent the multi-dimensional factor new energy consumption capacity model.

[0134] The least square fitting formula is:

[0135]

[0136] The theoretical fitting function is f(x1,x2,x3) The function after disassembly is f(x1,x2,x3)

[0137] The residual error formula is:

[0138]

[0139] y i The actual fitting function is f(x1,x2,x3)

[0140] The function relationship between factors obtained from the above formula is f(x1,x2,x3)

[0141] F=f n (X1,X2,X3)

[0142] Preferably, the multi-dimensional model is represented by the Peaks function, which includes:

[0143] The new energy consumption model is plotted on MATLAB to obtain a density map.

[0144] The normality of the density map is tested by Kolmogorov-Smirnov test to obtain a new energy consumption model conforming to normal distribution.

[0145] The Peaks function is used as a representative fitting of the normal function to fit the new energy consumption model conforming to normal distribution.

[0146] Specifically, the Peaks function is a typical multivariate function, which is essentially a density function of a two-dimensional Gaussian distribution, and the Peaks function has a concave and convex surface with three maximum points and three minimum points. Due to the randomness of the new energy consumption model sample space and the good fitting with the Gaussian function, the Peaks function is adopted, and by randomly taking values of the factors affecting the new energy consumption capacity, the color dimension is added on the basis of the three-dimensional space to represent the multi-dimensional factor new energy consumption capacity model.

[0147]

[0148] The form of the Gaussian function is a, b, and c are real constants and a>0. In the case of only considering installed capacity, power load and resource hours, the new energy consumption model is plotted on MATLAB to obtain the density map and use Kolmogorov-Smirnov test to test its normality, and it is found that the new energy consumption model conforms to the normal distribution, and the Peaks function is used as a representative of the normal function to fit the new energy consumption model.

[0149] In the actual application process, the original data is as shown in Table 1, and Table 1 is the original data record table.

[0150] Table 1

[0151]

[0152] The positive and negative indicators are standardized by using the range standardization formula, as shown in Table 2.

[0153] Table 2

[0154]

[0155] The new energy consumption model is obtained by using SPSS to perform LASSO regression analysis on the new energy consumption data, as shown in Figure 3

[0156] When the data has the characteristics of random distribution, the Peaks function is used to represent the multi-dimensional model, and 8 points are taken on the model according to different new energy utilization rates, as shown in Table 2, and the relationship between the new energy utilization rate and the three factors is analyzed, as shown in Figure 4 and Figure 5 Table 3 is the Peaks function point calculation table.

[0157] Table 3

[0158]

[0159]

[0160] ​Wherein x represents the installed capacity, y represents the wind resource hours, z represents the electricity load, and F represents the new energy utilization rate by color depth.

[0161] Through the above table analysis, it is obtained that the installed capacity and the wind resource factor are negatively correlated with the new energy consumption, and the electricity load is positively correlated with the new energy consumption, and the new energy utilization rate is 97.53% when each variable is the initial value, which is the same as the analysis result of the new energy consumption simulation software.

[0162] Considering the uncertainty of the sample space of the influencing factors, based on historical cases, the new energy utilization rate fitted by the spatial model is randomly obtained, as shown in Table 4, and the fitting result is less than 1% different from the historical case value, which can represent the multi-dimensional model.

[0163] Table 4

[0164]

[0165]

[0166] When the data has a functional relationship, the new energy consumption model is analyzed by using the factor function relationship, and the new energy consumption model is analyzed by using the factor function relationship. Figure 6 According to different new energy utilization rates, 9 points are selected on the factor function multi-dimensional model, as shown in Table 5, and the influence law of the change of the three factors on the new energy utilization rate is analyzed. Figure 7

[0167] Table 5

[0168]

[0169] It is obtained that the installed capacity and the wind resource factor are negatively correlated with the new energy consumption, and the electricity load is positively correlated with the new energy consumption, and the new energy utilization rate is 97.53% when each variable is the initial value, which is the same as the analysis result of the new energy consumption model.

[0170] Considering the high and low levels of new energy consumption capacity, two groups of new energy utilization rate sample sets are randomly selected, and the results are shown in Table 6, and the fitting result is less than 1.5% different from the historical case value, which can realize multi-dimensional model representation.

[0171] Table 6

[0172]

[0173] In this embodiment, a system for determining a new energy consumption model based on LASSO regression is also disclosed, comprising:

[0174] The construction module is configured to construct a new energy consumption model.

[0175] ​A judgment module is configured to perform dimension reduction calculation on the new energy consumption model by using LASSO regression, and determine whether there is a mathematical relationship between factors.

[0176] A first representation module is configured to represent the multi-dimensional model by a factor function relationship when the result of the judgment module is yes.

[0177] A second representation module is configured to represent the multi-dimensional model by using Peaks function when the result of the judgment module is no.

[0178] Preferably, the construction module specifically comprises:

[0179] A target function unit is configured to take the minimum new energy curtailment as a target function.

[0180] A constraint condition unit is configured to construct a constraint condition set; the constraint condition set comprises a new energy power curtailment constraint, a main transformer capacity constraint, an active power balance constraint, a conventional unit output constraint, a conventional unit ramping constraint, a system positive reserve constraint, a system negative reserve constraint, and a line power flow constraint.

[0181] A model determination unit is configured to determine the new energy consumption model according to the target function and the constraint condition set.

[0182] Preferably, the judgment module specifically comprises:

[0183] A simulation unit is configured to simulate the new energy consumption model to obtain basic data.

[0184] A preprocessing unit is configured to perform regularization processing on the basic data to obtain processed data.

[0185] A feature determination unit is configured to determine the influence characteristics of multiple uncertainty factors on new energy consumption capacity according to the processed data.

[0186] A model regression unit is configured to obtain a new energy consumption capacity model applicable to multi-dimensional space theory by using LASSO regression based on the influence characteristics.

[0187] Preferably, the first representation module specifically comprises:

[0188] A first fitting unit is configured to perform polynomial fitting by using the least square method, and make the residual error 0 and the sum of square errors minimum by using the cftool command of MATLAB, to obtain an optimal matching function.

[0189] A first three-dimensional model between variables is constructed according to the optimal matching function; the variables include installed capacity, electricity load, and resource hours; and the first three-dimensional model is configured to represent the multi-dimensional factor new energy consumption capacity model by color dimension.

[0190] Preferably, the second characterization module specifically comprises:

[0191] a plotting unit, configured to plot the new energy consumption model on MATLAB to obtain a density map;

[0192] a verification unit, configured to verify normality of the density map by using Kolmogorov-Smirnov test to obtain a new energy consumption model conforming to normal distribution;

[0193] a second fitting unit, configured to fit the new energy consumption model conforming to normal distribution by using a Peaks function as a representative of normal function.

[0194] The present application has the following beneficial effects:

[0195] The multi-dimensional new energy response surface model based on LASSO regression can visually depict the change of new energy utilization rate with factors, and is of great significance for studying the influence of multiple uncertain factors on new energy utilization rate.

[0196] The embodiments in the specification are described in a progressive manner, and each embodiment focuses on the difference from other embodiments, and the same or similar parts between the embodiments can be referred to each other. For the system disclosed by the embodiments, since it corresponds to the method disclosed by the embodiments, the description is relatively simple, and the related parts can be referred to the method part.

[0197] The principles and implementation manners of the present application are described by using specific examples in the present application, and the above embodiment description is only used to help understand the method of the present application and its core idea; meanwhile, for the general technical personnel in the art, according to the idea of the present application, the specific implementation manner and application range can be changed. In conclusion, the content of the specification should not be understood as the limitation of the present application.

Claims

1. A method for determining a new energy consumption model based on LASSO regression, characterized in that, The method comprises the following steps: constructing a new energy consumption model; performing dimensionality reduction calculation on the new energy consumption model by using LASSO regression, and determining whether there is a mathematical relationship between factors: if yes, representing the multi-dimensional model by a factor function relationship; if no, representing the multi-dimensional model by a Peaks function; the dimensionality reduction calculation on the new energy consumption model by using LASSO regression comprises: simulating the new energy consumption model to obtain basic data; performing regularization processing on the basic data to obtain processed data; determining the influence characteristics of multiple uncertainty factors on new energy consumption capacity according to the processed data; based on the influence characteristics, using LASSO regression to obtain a new energy consumption capacity model suitable for multi-dimensional space theory; the representation of the multi-dimensional model by the factor function relationship comprises: performing polynomial fitting by using the least square method, and making the residual error 0 and the sum of squared errors minimum by using the cftool command of MATLAB to obtain an optimal matching function; constructing a first three-dimensional model between variables according to the optimal matching function; the variables include installed capacity, power load and resource hours; the first three-dimensional model is used to represent the multi-dimensional factor new energy consumption capacity model by color dimension; the representation of the multi-dimensional model by the Peaks function comprises: plotting the new energy consumption model on MATLAB to obtain a density plot; using Kolmogorov-Smirnov test to test the normality of the density plot to obtain a new energy consumption model conforming to normal distribution; fitting the new energy consumption model conforming to normal distribution by using a Peaks function as a representative normal function. 2.The method of claim 1, wherein, The construction of the new energy consumption model comprises: taking the minimum new energy curtailment as an objective function; constructing a constraint condition set; the constraint condition set includes new energy curtailment constraint, main transformer capacity constraint, active power balance constraint, conventional unit output constraint, conventional unit ramping constraint, system positive reserve constraint, system negative reserve constraint and line flow constraint; determining the new energy consumption model according to the objective function and the constraint condition set.

3. A system for determining a new energy consumption model based on LASSO regression, characterized in that, The system for determining a new energy consumption model based on LASSO regression comprises: a construction module for constructing a new energy consumption model; a judgment module for performing dimensionality reduction calculation on the new energy consumption model by using LASSO regression, and determining whether there is a mathematical relationship between factors: a first representation module for representing a multi-dimensional model by a factor function relationship when the result of the judgment module is yes; a second representation module for representing a multi-dimensional model by a Peaks function when the result of the judgment module is no. 4.The system of determining a new energy consumption model based on LASSO regression according to claim 3, wherein, The construction module specifically comprises: an objective function unit for taking the minimum new energy curtailment as an objective function; The constraint condition unit is configured to construct a constraint condition set, and the constraint condition set comprises a new energy power curtailment constraint, a main transformer capacity constraint, an active power balance constraint, a conventional unit output constraint, a conventional unit ramping constraint, a system positive reserve constraint, a system negative reserve constraint, and a line power flow constraint. The model determination unit is configured to determine the new energy consumption model according to the target function and the constraint condition set. 5.The system of determining a new energy consumption model based on LASSO regression according to claim 3, wherein, The judgment module specifically comprises: The simulation unit is configured to simulate the new energy consumption model to obtain basic data. The preprocessing unit is configured to perform regularization processing on the basic data to obtain processed data. The feature determination unit is configured to determine an influence feature of multiple uncertainty factors on the new energy consumption capacity according to the processed data. The model regression unit is configured to obtain a new energy consumption capacity model applicable to a multi-dimensional space theory by using LASSO regression based on the influence feature. 6.The system of determining a new energy consumption model based on LASSO regression according to claim 3, wherein, The first representation module specifically comprises: The first fitting unit is configured to perform polynomial fitting by using a least square method, and make the residual error 0 and the sum of square errors minimum by using a cftool command of MATLAB, so as to obtain an optimal matching function. A first three-dimensional model between variables is constructed according to the optimal matching function, and the variables comprise installed capacity, power load, and resource hours; and the first three-dimensional model is configured to represent the multi-dimensional factor new energy consumption capacity model by color dimensions. 7.The system of determining a new energy consumption model based on LASSO regression according to claim 3, wherein, The second representation module specifically comprises: The plotting unit is configured to plot a density map on MATLAB by using the new energy consumption model. The verification unit is configured to verify normality of the density map by using Kolmogorov-Smirnov, and obtain a new energy consumption model conforming to normal distribution. The second fitting unit is configured to fit the new energy consumption model conforming to normal distribution by using a Peaks function as a representative of a normal function.

Images