Risk assessment method and equipment for integrated energy system based on R-vine copula
Through the R-vine copula method, the multi-dimensional risk factor joint probability distribution model is constructed, combined with Monte Carlo sampling and clustering technology, the problems of risk factor correlation and scenario randomness are ignored in the existing technology, and more accurate comprehensive energy Systematic risk assessment.
Patent Information
- Application Number
- CN202411164442.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-23
- Publication Date
- 2024-11-12
AI Technical Summary
The existing comprehensive energy system risk assessment method lacks consideration of risk factor correlation, cannot accurately evaluate the risk magnitude of multi-energy flow solutions and specific nodes and lines, and ignores the randomness of the operating scenario when modeling uncertain scenarios. .
Using the integrated energy system risk assessment method based on R-vine copula, a multidimensional risk factor joint probability distribution model is constructed by fitting the edge probability distribution of risk factors through non-parametric kernel density estimation and normal distribution, combining Monte Carlo sampling and clustering technology, obtain typical operating risk scenarios, and use the AHP-Topsis method to quantify the operating risk value.
Improve the comprehensiveness, accuracy and reliability of the risk assessment of comprehensive energy system, can refine the multi-energy flow solution and perceive the risk size of specific nodes and lines, and consider the randomness of multiple scenarios, providing quantitative and accurate risks. Evaluation results.
Smart Images

Figure CN118941092A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of risk assessment of integrated energy systems, and in particular relates to a risk assessment method and device for an integrated energy system based on R-vinecopula. Background Art
[0002] The "dual carbon" strategic goal has put forward requirements for the continuous transformation of the energy structure to green and low-carbon. The Integrated Energy System (IES) can integrate multiple clean energy sources and realize the flexible transformation of heterogeneous energy sources, which is conducive to promoting the construction of a clean, low-carbon, safe and efficient energy supply system. IES involves multiple links such as the production, transmission, storage, use and conversion of multiple energy subsystems such as electricity, gas and heat. Since the output of new energy such as wind turbines (WT) and photovoltaics (PV) has strong volatility, and the demand for electricity, gas and heat loads is easily affected by user behavior and climate, it also has strong randomness. These multiple uncertainties from the energy supply side and the load side will increase the operation risk of the integrated energy system. Therefore, it is necessary to accurately and comprehensively evaluate the operation risk of IES and establish a risk assessment system to identify the weak links in the operation process of IES, which is of great significance to guiding the safe, stable, economical and efficient operation of IES.
[0003] For IES risk assessment, in most studies, risk indicators usually point to the safety or economy of IES in a single way, lacking comprehensive consideration of risk assessment; and most of them are carried out under certain scenarios, ignoring the impact of the randomness of the operating scenario on risk assessment. For uncertainty scenario modeling, in most studies, the scenario set generates probability distribution based on historical data, uses Monte Carlo simulation (MCS), Latin hypercube sampling and other methods to generate uncertainty scenarios, and uses backward reduction, clustering and other methods to reduce the number of scenarios. Chinese patent application 202310636627.5 discloses a high-proportion renewable energy risk assessment method under severe weather conditions, extracts the state of a high-proportion new energy system according to the non-sequential Monte Carlo method, and performs risk assessment through optimal power flow calculation, which provides effective guidance for the planning and scheduling of new energy grid connection and resilience improvement strategies. However, the above method lacks consideration of the correlation of risk factors, and only considers the IES source-load power balance, without refining the IES multi-energy flow solution, and cannot perceive the risk size of specific nodes and lines. Therefore, it is necessary to design a comprehensive energy system risk assessment method to further improve the comprehensiveness, accuracy and reliability of the comprehensive energy system risk assessment results. Summary of the invention
[0004] The purpose of the present invention is to overcome the defects of the above-mentioned prior art and to provide a comprehensive energy system risk assessment method and equipment based on R-vinecopula, so as to further improve the comprehensiveness, accuracy and reliability of the risk assessment results of the comprehensive energy system.
[0005] The purpose of the present invention can be achieved by the following technical solutions:
[0006] The present invention provides a comprehensive energy system risk assessment method based on R-vine copula, comprising:
[0007] Establish a comprehensive energy system risk assessment indicator system;
[0008] The marginal probability distribution of risk factors was fitted by nonparametric kernel density estimation method and normal distribution. According to the marginal probability distribution of risk factors, the R-vine copula method was used to construct a joint probability distribution model of multidimensional risk factors.
[0009] Monte Carlo sampling is performed on the basis of the multi-dimensional risk factor joint probability distribution model, and risk factor values are obtained by inverse probability transformation and then clustering is performed to obtain typical operational risk scenarios of the integrated energy system;
[0010] According to the risk assessment index system of the integrated energy system and the typical operation risk scenarios of the integrated energy system, the AHP-Topsis method is adopted to quantitatively calculate the operation risk value of the integrated energy system.
[0011] Furthermore, the comprehensive energy system risk assessment index system includes safety indicators;
[0012] The safety indicators include power grid safety risk indicators, gas grid safety risk indicators and heat network safety risk indicators, wherein the power grid safety risk indicators are calculated based on the electric load weight coefficient and the power exceeding limit degree of each branch of the power grid, the gas grid safety risk indicators are calculated based on the gas load weight coefficient and the gas pressure exceeding limit degree of each node of the gas grid, and the heat network safety risk indicators are calculated based on the heat load weight coefficient and the temperature exceeding limit degree of each node of the heat network.
[0013] Furthermore, the comprehensive energy system risk assessment index system includes reliability index;
[0014] The reliability indicators include a power grid reliability risk indicator, a gas grid reliability risk indicator and a heat grid reliability risk indicator, wherein the power grid reliability risk indicator is calculated based on the load weight coefficient of each node in the power grid and the energy shortage ratio, the gas grid reliability risk indicator is calculated based on the load weight coefficient of each node in the gas grid and the energy shortage ratio, and the heat grid reliability risk indicator is calculated based on the load weight coefficient of each node in the heat grid and the energy shortage ratio.
[0015] Furthermore, the comprehensive energy system risk assessment index system includes economic indicators;
[0016] The economic index is the proportion of risk cost to total operating cost, and the risk cost includes the cost caused by load interruption and waste of new energy.
[0017] Furthermore, the risk factors include wind power output, photovoltaic output, electric load demand, gas load demand and thermal load demand. The process of fitting the marginal probability distribution of the risk factors using the non-parametric kernel density estimation method and the normal distribution is as follows:
[0018] Based on Gaussian kernel function: Fitting the marginal probability distribution of wind power output and photovoltaic output at time t and
[0019]
[0020] in, and are the output of wind turbine and photovoltaic at time t, and are the output of wind turbine and photovoltaic power generation at time t on day d, h is the smoothing coefficient, and n is the sample size;
[0021] The marginal probability distribution of the electric load demand, gas load demand and heat load demand is described by normal distribution, where the marginal probability distribution of the electric load demand is f(L e ), the expression is as follows:
[0022]
[0023] Among them, μ e and σ e are the mean and variance of the historical data of electric load, and P is the electric load power.
[0024] Furthermore, the process of obtaining the typical operation risk scenario of the integrated energy system is as follows:
[0025] In the multi-dimensional risk factor joint probability distribution model, Monte Carlo random sampling is performed on the joint probability distribution of each period to obtain the cumulative probability distribution of the risk factor in each period;
[0026] When the risk factor is the electric load demand, gas load demand or heat load demand, the corresponding CDF of the cumulative probability distribution is inversely transformed to obtain the specific value of the risk factor in each period; when the risk factor is wind power output or photovoltaic output, the spline difference method is used to obtain the specific value of the risk factor in each period;
[0027] The K-means algorithm is used to reduce and cluster the specific values of the risk factors obtained to obtain the typical operation scenarios of the integrated energy system.
[0028] Furthermore, the specific process of constructing the multi-dimensional risk factor joint probability distribution model using the R-vine copula method is as follows:
[0029] Get the risk factor data set X = {X1, X2, ...X n}, set node N = X, calculate the Kendall rank correlation coefficient of the empirical Copula function in N, and use Prim's algorithm to build an R-vine model, where the R-vine model includes a series of maximum spanning trees, each edge E of each tree corresponds to a connected binary Copula function, and the weight of each edge is the Kendall rank correlation coefficient of the empirical Copula function;
[0030] The maximum likelihood estimation method is used to estimate the binary Copula function parameter values in each maximum spanning tree, and the Euclidean distance between the Kendall rank correlation coefficients of the binary Copula function to be determined and the empirical Copula function is compared. The optimal binary Copula function between the nodes in each tree is determined based on the AIC information criterion, and a multidimensional risk factor joint probability distribution model is constructed. The expression of the multidimensional risk factor joint probability distribution model is as follows:
[0031]
[0032] Among them, f(x1,x2,…x n ) is the joint probability distribution model of multi-dimensional risk factors, f k (x k ) is the risk factor x k The marginal probability distribution of E={E1,E2,…E n-1} is the edge set, c j(e),k(e)|D(e) is the binary Copula function corresponding to edge e, j(e), k(e) are the nodes connected by edge e, D(e) is the condition, F(x j(e) |x D(e) ) and F(x k(e) |x D(e) ) are the probability distributions of the nodes connected by edge e under the condition D(e).
[0033] Furthermore, the calculation formula of the operation risk value of the integrated energy system is as follows:
[0034]
[0035] Among them, RISK is the operational risk value of the integrated energy system, π(s) is the probability of occurrence of a typical operational risk scenario s of the integrated energy system, R1, R2, …, R n T is the various indicators in the comprehensive energy system risk assessment indicator system. R (R1,R2,…,R n ) is a comprehensive evaluation function constructed using the AHP-Topsis method.
[0036] Furthermore, the comprehensive evaluation function T R (R1,R2,…,R n ) is constructed as follows:
[0037] Construct the weighted normalized matrix B of the comprehensive energy system risk assessment index system:
[0038]
[0039] The element b in the matrix ij =ω j z ij ,ω j is the weight of each indicator, obtained through the analytic hierarchy process AHP, z ij is the element in the risk index normalization matrix Z, i = 1, 2, ..., m, j = 1, 2, ..., n;
[0040] The comprehensive evaluation function T is constructed by using the superior and inferior solution distance method Topsis R (R1,R2,…,R n ), the specific process is: based on the weighted normalization matrix B, determine the positive and negative ideal solution sets B + and B - :
[0041]
[0042] Calculate the Euclidean distance between each element in the weighted normalized matrix B and the corresponding positive ideal solution and negative ideal solution and
[0043]
[0044] Calculate the closeness C of each element to the ideal solution i :
[0045]
[0046] Construct comprehensive evaluation function T R (R1,R2,…,R n ), T R (R1,R2,…,Rn )=1 / C i .
[0047] The present invention also provides an integrated energy system risk assessment device based on R-vine copula, comprising: a risk assessment index module, a multi-dimensional risk factor joint probability distribution model generation module, an integrated energy system typical operation risk scenario generation module and an integrated energy system operation risk value calculation module;
[0048] Wherein, the risk assessment indicator module is used to establish a comprehensive energy system risk assessment indicator system;
[0049] The multidimensional risk factor joint probability distribution model generation module is used to use a non-parametric kernel density estimation method and a normal distribution to fit the marginal probability distribution of risk factors, and to construct a multidimensional risk factor joint probability distribution model using an R-vinecopula method based on the marginal probability distribution of risk factors;
[0050] The typical operation risk scenario generation module of the integrated energy system is used to perform Monte Carlo sampling based on the multi-dimensional risk factor joint probability distribution model, and obtain the risk factor values through inverse probability transformation and then perform reduction clustering to obtain the typical operation risk scenario of the integrated energy system;
[0051] The integrated energy system operation risk value calculation module is used to quantitatively calculate the operation risk value of the integrated energy system using the AHP-Topsis method based on the integrated energy system risk assessment index system and the typical operation risk scenarios of the integrated energy system.
[0052] Compared with the prior art, the present invention has the following beneficial effects:
[0053] The present invention proposes a risk assessment method for an integrated energy system based on R-vine copula. First, a risk assessment index system for an integrated energy system is established, including safety index, reliability index and economic index, to ensure the comprehensiveness of the risk assessment results. The safety index and reliability index are calculated based on the operating conditions of each node and each branch in the integrated energy system, and the multi-energy flow solution of the integrated energy system is refined, so that the risk size of specific nodes and lines can be perceived. Then, a non-parametric kernel density estimation method and a normal distribution are used to fit the marginal probability distribution of risk factors. The risk factors specifically include wind power output, photovoltaic output, electric load demand, gas load demand and thermal load demand. According to the marginal probability distribution of the risk factors, R-vine is used to estimate the risk of wind power output, photovoltaic output, electric load demand, gas load demand and thermal load demand. The copula method can be used to construct a joint probability distribution model of multidimensional risk factors. The uncertainty and correlation between risk factors can be captured through tree structure and binary Copula function, which can avoid underestimation of risk assessment and obtain more accurate risk assessment results. Finally, Monte Carlo sampling is performed on the basis of the joint probability distribution model of multidimensional risk factors. The risk factor values are obtained by solving the inverse probability transformation and then clustered to obtain typical operation risk scenarios of the integrated energy system. The randomness of multiple scenarios is taken into account, and the complexity of subsequent calculations can be reduced. According to the risk assessment indicator system of the integrated energy system and the typical operation risk scenarios of the integrated energy system, the AHP-Topsis method is used to calculate the operation risk value of the integrated energy system, which can obtain quantitative and accurate risk assessment results of the integrated energy system. BRIEF DESCRIPTION OF THE DRAWINGS
[0054] Figure 1 Generate and evaluate flow charts for typical operational risk scenarios of IES integrated energy systems;
[0055] Figure 2 This is a schematic diagram of the structure of the electricity-gas-heat interconnected integrated energy system;
[0056] Figure 3 Build a process diagram for the joint probability distribution model of multi-dimensional risk factors;
[0057] Figure 4 This is a schematic diagram of the tree structure of the R-vine copula model.
[0058] Among them, R WT , R PV , R E , R G , R HThere are five risk factors, namely wind power output, photovoltaic output, electric load, gas load and thermal load; F, SG, t, C, J90 are Frank copula, Survival Gumbel copula, t-copula, Clayton copula and Joe copula rotated 90 degrees respectively;
[0059] Figure 5 The network topology of the integrated energy system IES is shown in Figure 1.
[0060] Among them, CHP is a combined heat and power model, P2G is a power-to-gas model, NGU is a gas-fired generator model, EB is an electric boiler model, GB is a gas boiler model, and CFU is a coal-fired unit model;
[0061] Figure 6 This is a typical scenario of wind turbine output and photovoltaic output in different seasons.
[0062] Among them, (6a) is the output of photovoltaic power in the heating season, (6b) is the output of photovoltaic power in other seasons, (6c) is the output of wind turbine in the heating season, and (6d) is the output of wind turbine in other seasons.
[0063] Figure 7 These are typical scenarios of electricity load, gas load and heat load in different seasons.
[0064] Among them, (7a) is the scenario 1 of the three loads in the heating season, (7b) is the scenario 2 of the three loads in the heating season, (7c) is the scenario 3 of the three loads in the heating season, (7d) is the scenario 1 of the three loads in other seasons, (7e) is the scenario 2 of the three loads in other seasons, (7f) is the scenario 3 of the three loads in other seasons, blue represents electric load, red represents gas load, yellow represents heat load, and the horizontal axis is time t / h;
[0065] Figure 8 This is a schematic diagram of the safety indicators of the integrated energy system.
[0066] Among them, (8a) is the risk sequence diagram of power over-limit of power grid branch, (8b) is the risk sequence diagram of gas grid node pressure over-limit, and (8c) is the risk sequence diagram of temperature over-limit of heating grid node (water supply network);
[0067] Fig. 9 Schematic diagram of reliability index of integrated energy system.
[0068] Among them, (9a) is the risk time series diagram of the energy shortage ratio of the power grid node, (9b) is the risk time series diagram of the energy shortage ratio of the gas grid node, and (9c) is the risk time series diagram of the energy shortage ratio of the heat grid node (water supply network);
[0069] Fig.10 It is a schematic diagram of the risk factor time series cross-correlation analysis process;
[0070] Fig.11 This is a schematic diagram of the time series cross-correlation of typical risk factors.
[0071] Among them, (11a) is the cross-correlation function diagram of photovoltaic output-electric load (other seasons), (11b) is the cross-correlation function diagram of electricity-heat load (heating season), (11c) is the cross-correlation function diagram of wind-solar output (other seasons, the sequence selects 7-20 periods), the horizontal axis is the time offset, and the vertical axis is the expected value of the correlation coefficient of each scenario in the corresponding season;
[0072] Fig.12 is the risk factor correlation matrix,
[0073] Wherein, (12a) is the risk factor correlation matrix of method A, and (12b) is the risk factor correlation matrix of method B;
[0074] Fig.13 Run comprehensive risk time series diagrams for IES;
[0075] Fig.14 Run composite risk for systems under different risk appetites. DETAILED DESCRIPTION
[0076] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. This embodiment is implemented based on the technical solution of the present invention, and provides a detailed implementation method and specific operation process, but the protection scope of the present invention is not limited to the following embodiments.
[0077] Example 1
[0078] This embodiment provides a comprehensive energy system risk assessment method based on R-vine copula, which is suitable for Figure 2 The electricity-gas-heat interconnected integrated energy system (IES) shown in the figure takes into account the randomness of the integrated energy system operation scenario and the correlation of risk factors, and improves the comprehensiveness, accuracy and reliability of the integrated energy system risk assessment results. Specifically, it includes the following steps:
[0079] S1. Obtain the basic data and economic and technical parameters of each energy coupling equipment in the electricity-gas-heat interconnected integrated energy system (IES), and establish an electricity-gas-heat interconnected integrated energy system model.
[0080] like Figure 2As shown in Figure 1, the electricity-gas-heat interconnected integrated energy system (IES) consists of three parts: the power network, the natural gas network and the thermal network, which are connected through energy coupling equipment. The energy coupling equipment models include: combined heat and power (CHP) model, power-to-gas (P2G) model, gas generator unit (NGU) model, electric boiler (EB) model, gas boiler (GB) model and coal-fired unit model (CFU). The expressions of each model are as follows:
[0081] Power network model:
[0082] P ij =(θ i -θ j ) / x ij (1)
[0083]
[0084] In formula (1) and (2), P ij is the branch power between nodes i and j, x ij is the reactance of the branch, θ i ,θ j are the phase angles of nodes i and j respectively. inj The node injection power matrix, θ is the node phase angle vector, b ii and b ij are the diagonal elements and other position elements of matrix B respectively.
[0085] In the thermal network, the heat load power Φ is expressed as:
[0086]
[0087] In formula (3), C p is the specific heat capacity of water, m i is the water injection volume of load node i, are the supply water temperature and return water temperature of node i respectively.
[0088] In order to save calculation cost, the delay effect of the heat network is not considered, and only the influence of ambient temperature and the material of the pipeline itself is considered. The water temperature at the pipeline outlet is:
[0089]
[0090] In formula (4), The outlet and inlet water temperatures of pipe b are, respectively, T air is the ambient temperature, λ b is the heat transfer coefficient of the pipe, L b is the length of the pipe.
[0091] The water flow in the thermal pipeline should satisfy the basic law, that is, the amount of water flowing into each node is equal to the amount of water flowing out:
[0092] A nb m=m q (5)
[0093] In formula (5), A nb is the thermal network association matrix, n and b are the number of nodes and pipes respectively, the elements in the matrix are 1 for inflow, -1 for outflow, 0 for no connection, m for the water flow of each pipe, m q Represents the water flow rate injected into the node from the heat source or supplied to the heat load by the node.
[0094] In the natural gas network, the Weymouth steady-state model is used to describe the relationship between natural gas pipeline flow and pressure, and the expression is:
[0095]
[0096] In formula (6), F k-n 、c k-n They represent the natural gas flow and pipeline transmission parameters in the pipeline connecting nodes k and n, respectively, and are related to the length, diameter, and temperature of the pipeline; k 、p n Respectively represent the air pressure of nodes k and n, sgn(p k ,p n ) is the direction parameter, indicating that natural gas always flows from nodes with higher pressure to nodes with lower pressure.
[0097]
[0098] The expression of the combined heat and power (CHP) model is:
[0099]
[0100] In formula (8), P chp is the total output power of the CHP unit, P MGT , P WHR are the electrical power and thermal power output of the CHP unit, They are the electrical conversion efficiency and thermal conversion efficiency of the CHP unit respectively.
[0101] The expression of the power-to-gas (P2G) model is:
[0102]
[0103] In formula (9), is the gas production of P2G during period t, η p2g is the electrical conversion efficiency of P2G, is the electric power consumed by P2G in period t.
[0104] The expression of the gas generator unit (NGU) model is:
[0105]
[0106] In formula (10), is the power generation of NGU in period t, η ngu is the gas-to-electricity conversion efficiency, is the gas consumption of NGU in period t.
[0107] The expression of the electric boiler (EB) model is:
[0108]
[0109] In formula (11), is the thermal power output by EB during period t, η eb is the electrothermal conversion efficiency of EB, is the electric power consumed by EB in period t.
[0110] The expression of the gas boiler (GB) model is:
[0111]
[0112] In formula (12), is the thermal power output of GB in period t, η gb is the gas-to-heat conversion efficiency, is the natural gas power consumed by GB in period t, is the natural gas consumption per unit time, V LHV It is the lower calorific value of natural gas.
[0113] The constraints of the electricity-gas-heat interconnected integrated energy system model include power balance constraints, equipment output and capacity constraints, ramp constraints and other constraints. Among them, the power balance constraints include grid drop constraints, gas network constraints and heat network constraints. The expressions of each constraint are as follows:
[0114] Electric network constraints:
[0115]
[0116] In formula (13), represents the association matrix between coal-fired unit i and the power grid, n is the number of power grid nodes, It represents the electric power produced by coal-fired unit i in period t. NGU, CHP, WT, PV, EB, ES are similar to the grid association matrix and electric power expression, so they will not be explained again; are the wind power and solar power abandonment in period t, Take -1 or 1, indicating the charging and discharging status of the battery, are the active load and load shedding of node n in period t respectively.
[0117] Gas network constraints:
[0118]
[0119] In formula (14), represents the association matrix between gas source i and gas network, m is the number of gas network nodes, the association matrix expressions of P2G, GB, NGU, GS, CHP and gas network are similar and will not be explained again; H l It represents the association matrix between natural gas pipelines and gas nodes. The elements in the matrix are positive for inflow and negative for outflow. l Represents the natural gas pipeline flow. They represent the natural gas output power of GW and P2G in period t respectively; They represent the gas load and load shedding of the node respectively; They represent the gas consumption of GB and NGU in period t respectively; is the power of the gas storage device during period t, Indicates the state of inflation and deflation, with a value of -1 or 1, representing the storage state and the release state respectively.
[0120] Thermal network constraints:
[0121]
[0122] In formula (15) and (16), is the heat load of the root node in the heat network during period t, are the heat load and load shedding at the heat node n in period t respectively; is the association matrix of the heat storage device with respect to the heat node, The values are -1 and 1, which represent the release state and storage state of the heat storage device respectively. is the power of the heat storage device, Represents the total load of heat load node n during period t.
[0123] Equipment output and capacity constraints:
[0124]
[0125] In formula (17), They represent the power and upper limit of the equipment in time period t, including NGU, CFU, CHP, EB, GB, P2G, and GC; They represent the capacity of energy storage equipment and its upper limit in period t, including ES, GS, and HS; Represents the operating power of the energy storage device during period t.
[0126] Hill climbing constraints:
[0127]
[0128] In formula (18), are the down-ramp rate and up-ramp rate of coal-fired generating unit i respectively.
[0129] Other constraints:
[0130]
[0131] In formula (19) and (20), are the load shedding amount and load value of the electric network node j in period t, respectively. The gas and heat network nodes are similar and will not be described in detail. They are respectively the abandoned wind power and power generation of wind turbine k in period t. Photovoltaic power generation is similar and will not be described in detail.
[0132] S2. Establish a risk assessment index system for the integrated energy system, including safety index, reliability index and economic index. The safety index and reliability index are calculated based on the operating conditions of each node and branch in the integrated energy system.
[0133] The safety index includes the power grid safety risk index, gas grid safety risk index and heat grid safety risk index. The specific expressions are as follows:
[0134] Grid security risk indicators
[0135]
[0136] In formula (21), is the power weight coefficient of branch i, is the severity of the power limit violation of branch i in period t, and its expression is as follows:
[0137]
[0138] In formula (22), is the power of branch i in period t, is the maximum safe power allowed for branch i.
[0139] Gas network safety risk indicators
[0140]
[0141] In formula (23), is the gas load weight coefficient of node j, is the severity of the air pressure exceeding the limit at node j during period t, and its expression is as follows:
[0142]
[0143] In formula (24), P j,t is the air pressure value of node j in period t, are the maximum and minimum safety air pressure of node j respectively.
[0144] Heating network safety risk indicators
[0145]
[0146] In formula (25), is the heat load coefficient of node k, is the severity of the temperature exceeding the limit at node k in period t, and its expression is as follows:
[0147]
[0148] In formula (26), T k,t is the air pressure value of node k during period t, are the maximum and minimum safety air pressure of node k respectively.
[0149] In order to evaluate the energy supply reliability of the integrated energy system, the energy shortage ratio is used as the reliability risk indicator, and its expression is:
[0150]
[0151] In formula (27), Respectively represent the reliability risk index of the electricity, gas and heat networks in period t, Respectively represent the load weight of electricity, gas and heat nodes in the energy network; Respectively represent the energy shortage ratio of electricity, gas, and heat nodes i, j, and k in period t, where The expression is:
[0152]
[0153] In formula (28), They are respectively the electric load demand and supply of grid node i in period t. The heat network and gas network are similar to the power grid and will not be elaborated on here.
[0154] Economic indicators Risk cost as a percentage of total operating cost The risk cost includes the cost caused by load interruption and waste of new energy. and The expression is:
[0155]
[0156] The calculation formula for each part of the cost is as follows:
[0157]
[0158] In formula (30), Respectively represent the weight coefficients of electricity, gas and heat loads, c E,loss 、c G,loss 、c H,loss Respectively represent the compensation costs for interruption of unit electricity, gas and heat load, are the load interruptions of the i-th, j-th, and k-th nodes in the electricity, gas, and heat networks in period t, respectively.
[0159]
[0160] In formula (31), N wt 、N pv are the number of wind turbines and photovoltaic power generation devices, c wt,abn 、c pv,abn They are the penalty fees for the waste of new energy for wind and solar units, are the abandoned power of wind power and photovoltaic power generation in period t respectively.
[0161]
[0162] In formula (32), are the coal burning cost, gas purchasing cost, and penalty term of the second-order cone relaxation of the Weymouth equation of the gas grid in period t, See formula (30) and (31). The calculation formulas for coal burning and gas purchasing costs are as follows:
[0163]
[0164] In formula (33) and (34), c i,1 、c i,2 、c i,3 is the fuel cost coefficient of the i-th coal-fired unit, P i,t is the power generation of unit i during period t, N coal is the number of units; is the natural gas price in period t, is the gas purchase volume of the gas source during period t, N gw is the number of gas sources.
[0165] S3. The nonparametric kernel density estimation method and normal distribution are used to fit the marginal probability distribution of risk factors. According to the marginal probability distribution of risk factors, the R-vine copula method is used to construct a multidimensional risk factor joint probability distribution model.
[0166] Risk factors include wind power output, photovoltaic output, electricity load demand, gas load demand and heat load demand. The above risk factors are not only random, but also correlated with each other. Therefore, the risk factors are not independent, and it is difficult to model the joint probability distribution. Figure 1 As shown, this embodiment uses the R-vine copula method to establish a joint probability distribution model for the five-dimensional risk factors, performs Monte Carlo sampling on the basis of the joint probability distribution model, obtains the risk factor values through inverse probability transformation, and then reduces the clusters, and finally obtains a typical risk scenario for the operation of the electricity-gas-heat interconnected integrated energy system. The specific process is as follows:
[0167] 1) Probability distribution model of wind and solar output
[0168] Based on the measured data of wind power and photovoltaic output, the non-parametric kernel density estimation method is used to fit the wind and light output probability model, and the Gaussian kernel function is selected:
[0169]
[0170] Based on the time series of wind and solar output, the probability density function of wind and solar output in each time period within 24 hours is obtained. and
[0171]
[0172] In formula (36), and are the output of wind turbine and photovoltaic at time t, and are the outputs of wind turbines and photovoltaics at time period t on day d, respectively. t=1,2,…,24 means 24 time periods in a day. h is the smoothing coefficient and n is the sample size of the time series.
[0173] 2) Probability distribution model of electricity, gas and heat load
[0174] The marginal probability distribution of the electric load demand, gas load demand and heat load demand is described by normal distribution, where the marginal probability distribution of the electric load demand is f(L e ), the expression is as follows:
[0175]
[0176] Among them, μ e and σ eare the mean and variance of the historical data of electric load respectively, P is the electric load value, and the gas and heat loads are similar to the electric load and will not be repeated here.
[0177] The specific process of using the R-vine copula method to construct a multi-dimensional risk factor joint probability distribution model is as follows: Figure 3 As shown, the following steps are included:
[0178] S301, obtain risk factor data set X = {X1, X2, ...X n}, set node N = X.
[0179] S302. Calculate the Kendall rank correlation coefficient of the empirical Copula function in N:
[0180]
[0181] In formula (38), x i and i are the samples corresponding to the binary copula, and n is the total number of samples in the risk factor data set.
[0182] S303. Use Prim (minimum spanning tree) algorithm to construct R-vine model. R-vine model includes a series of maximum spanning trees. Each edge E of each tree corresponds to a connected binary Copula function. The weight of each edge is the Kendall rank correlation coefficient of the empirical Copula function.
[0183] S304. Use the maximum likelihood estimation method to estimate the parameter values of the binary Copula function in each maximum spanning tree, compare the Euclidean distance between the Kendall rank correlation coefficients of the binary Copula function to be determined and the empirical Copula function, determine the optimal binary Copula function between the nodes in each tree based on the AIC information criterion, and construct a multidimensional risk factor joint probability distribution model.
[0184] For two marginal probability distribution functions U = F1(x), V = F2(y), the maximum likelihood estimate is:
[0185]
[0186] In formula (39), u i 、v i are the CDF values of the i-th sample in the sample set.
[0187] The expression of the multi-dimensional risk factor joint probability distribution model finally obtained is as follows:
[0188]
[0189] In formula (40), n-1 means there are n-1 layers of tree structure, f k (x k ) is the risk factor x k The marginal probability distribution of E={E1,E2,…E n-1} is the edge set, c j(e),k(e)|D(e) is the binary Copula function corresponding to edge e, j(e), k(e) are the nodes connected by edge e, D(e) is the condition, F(x j(e) |x D(e) ) and F(x k(e) |x D(e) ) are the probability distributions of the nodes connected by edge e under this condition.
[0190] S4. Monte Carlo sampling is performed on the basis of the joint probability distribution model of multi-dimensional risk factors. The risk factor values are obtained by solving the inverse probability transformation and then clustering is performed to obtain the typical operation risk scenario of the integrated energy system. The specific process is as follows:
[0191] S401. In the multi-dimensional risk factor joint probability distribution model, Monte Carlo random sampling is performed on the joint probability distribution of each time period to obtain the cumulative probability distribution of the risk factor in each time period.
[0192] S402. When the risk factor is electric load demand, gas load demand or heat load demand, the CDF corresponding to the cumulative probability distribution is inversely transformed to obtain the specific value of the risk factor in each period; when the risk factor is wind power output or photovoltaic output, since the probability distribution function of wind and photovoltaic output is obtained by non-parametric kernel density estimation, its inverse function is not easy to obtain, so the spline difference method is used to obtain the specific value of the risk factor in each period, and each probability interval is divided into d-1 small intervals. The interpolation polynomial is shown as follows:
[0193]
[0194] In formula (41), d = 1, 2, ..., n, t = 1, 2, ..., 24, the interpolation function is determined by solving the coefficients of the polynomial, and finally the sample value corresponding to the cumulative probability distribution is obtained.
[0195] S403. Due to the large scale of random sampling, the K-means algorithm is used to reduce and cluster the specific values of the risk factors obtained, obtain typical operation scenarios of the integrated energy system, and calculate the probability of each scenario.
[0196] S5. Based on the risk assessment index system of the integrated energy system and the typical operational risk scenarios of the integrated energy system, the AHP-Topsis method is used to quantitatively calculate the operational risk value RISK of the integrated energy system. The specific calculation formula is as follows:
[0197]
[0198] In formula (42), π(s) is the probability of occurrence of typical operation risk scenario s of the integrated energy system, R1, R2, …, R n T is the various indicators in the comprehensive energy system risk assessment indicator system. R (R1,R2,…,R n ) is a comprehensive evaluation function constructed using the AHP-Topsis method. The construction process is as follows:
[0199] S501. Construct the weighted normalized matrix B of the comprehensive energy system risk assessment index system:
[0200]
[0201] The element b in the matrix ij =ω j z ij ,ω j is the weight of each indicator, obtained through the analytic hierarchy process AHP, z ij is the element in the risk index normalization matrix Z, i = 1, 2, ..., m, j = 1, 2, ..., n;
[0202] S501. Construct a comprehensive evaluation function T through the superior and inferior solution distance method Topsis R (R1,R2,…,R n ), the specific process is: determine the positive and negative ideal solution sets B based on the weighted normalization matrix B + and B - :
[0203]
[0204] Calculate the Euclidean distance between each element in the weighted normalized matrix B and the corresponding positive ideal solution and negative ideal solution and
[0205]
[0206] Calculate the closeness C of each element to the ideal solution i :
[0207]
[0208] Construct comprehensive evaluation function T R (R1,R2,…,R n )=1 / C i .
[0209] To verify the effectiveness of the above method, this example uses an improved IEEE 30-node power network, a Belgian 20-node natural gas network, and a 6-node thermal network consisting of an electric-gas-heat integrated energy system for example analysis. Among them, wind power and photovoltaic power are connected to grid nodes 18 and 16 respectively, and energy storage devices are connected to the electric, gas, and thermal networks respectively. The network topology is as follows: Figure 5 shown.
[0210] Set the power limit of the power grid branch The gas network node pressure and the heat network node temperature are allowed to fluctuate 20% above and below the maximum and minimum safety values respectively. The simulation process is solved by mixed integer linear programming (MILP), and the objective function is the system operation cost. The hardware platform for the simulation is 64-bit Win10 PC, processor: 13th Gen Intel(R)Core(TM)i5-13490F, 2.50GHz, RAM: 32GB. The software platform for scenario generation and simulation is R-4.3.0.
[0211] Ⅰ. Modeling results of the joint probability distribution model of multi-dimensional risk factors
[0212] The wind turbine and photovoltaic output data of a certain region throughout the year as well as the historical data of electricity, gas and heat loads are selected, and a five-dimensional R-vine copula model is established based on the marginal probability distribution of risk factors. Taking a certain time period of a day as an example, the tree structure is as follows: Figure 4 As shown, R WT , R PV , R E , R G , R H There are five risk factors, namely wind power output, photovoltaic output, electric load, gas load and thermal load; F, SG, t, C, and J90 are Frank copula, Survival Gumbel copula, t-copula, Clayton copula and Joe copula rotated 90 degrees respectively.
[0213] The model estimation results are shown in Table 1. The nodes 1, 2, 3, 4, and 5 in the edge correspond to the risk factors R WT , R PV , R E , R G , R H ; cop is the copula function type between nodes; par1 and par2 are the parameters of the copula function; tau is the Kendall correlation coefficient; utd and ltd are the upper tail correlation coefficient and the lower tail correlation coefficient respectively.
[0214] Table 1 Risk factor R-vine copula modeling results
[0215]
[0216]
[0217] As can be seen from Table 1, there is a large correlation between wind power-photovoltaic output and wind power output-electric load, with Kendall correlation coefficients of 0.83 and 0.75 respectively. The latter has a large lower tail correlation, with a correlation coefficient of 0.88, indicating that when one is in a plummeting trend, the other will also plummet. For example, in extreme stormy weather, the power load drops sharply, and at the same time, due to the shutdown protection of the wind turbine, the wind turbine output will also drop sharply. It can be seen from T2 that under the condition of wind power output, the Kendall correlation coefficient between each node is positive, showing a large positive correlation, and the electric and thermal load has a great lower tail correlation. As the number of layers of the tree increases, the indirect infectiousness of the risk factor decreases layer by layer.
[0218] II. Results of generating typical operational risk scenarios for integrated energy systems
[0219] The historical data selected the wind turbine and photovoltaic output values and the electricity, gas and heat load values of a certain region throughout the year, and the Monte Carlo sampling scale was 10,000. Taking into account the seasonality of heat load, the whole year was divided into (heating season and other seasons), and finally the typical scenarios of risk factors in different seasons were obtained. The typical scenarios of wind power output and photovoltaic output in different seasons are as follows: Figure 6 As shown in Figure 2, typical scenarios of electricity load, gas load and heat load in different seasons are as follows: Figure 7 shown.
[0220] Depend on Figure 6 It can be seen that wind and solar power output has obvious seasonality and timing: the average daily duration and intensity of photovoltaic output in the heating season are smaller than those in other seasons; wind power output also shows a weaker trend in the heating season.
[0221] analyze Figure 7 It can be concluded from the scenarios in other seasons that: scenario 3 has a low probability and a large power load, showing the characteristics of a hot summer; scenario 1 has the highest probability and a small power load, showing the characteristics of spring and autumn; scenario 2 is the transition season between autumn and winter and spring and summer. Figure 6 and Figure 7 In certain periods, wind, light, and load show the same or opposite trends, showing a certain correlation. Therefore, the typical operation risk scenario of the integrated energy system generated by the method of this embodiment can better retain the uncertainty and correlation of risk factors.
[0222] III. Analysis of scenario generation results
[0223] Based on the joint probability distribution model of multi-dimensional risk factors, the safety and reliability indicators of the integrated energy system are as follows: Figure 8 , Fig. 9 shown. Figure 8 In the figure, (8a) is the risk sequence diagram of power over-limit of power grid branch, (8b) is the risk sequence diagram of gas grid node pressure over-limit, and (8c) is the risk sequence diagram of temperature over-limit of heating grid node (water supply network). Figure 8 From (8a), we can see that the risk is mainly concentrated on the branches connecting nodes 19, 10, and 27. This is because the above nodes are connected to EB, P2G, and CHP. The increase in line transmission power will lead to an increase in the risk of branch power exceeding the limit. Figure 8 From (8b), we can see that the risks are mainly concentrated in gas network nodes 1, 2, 7, and 10. This is because the above nodes are connected to gas-consuming equipment such as CHP, NGU, and GB. Among them, the risk of node 10 is relatively low because node 10 is also connected to P2G and can share part of the gas load. Figure 8 From (8c), we can see that the risk is mainly concentrated in node 2 and the risk size is positively correlated with the heat load size.
[0224] Fig. 9 In the figure, (9a) is the risk time series diagram of the energy shortage ratio of the power grid node, (9b) is the risk time series diagram of the energy shortage ratio of the gas grid node, and (9c) is the risk time series diagram of the energy shortage ratio of the heat grid node (water supply network). Fig. 9 From (9a), we can see that the risk is mainly concentrated in node 19. This is because node 19 is connected to EB, and in order to ensure its normal operation, part of the power load on the node is reduced. Fig. 9 From (9b), we can see that the risks are mainly concentrated in nodes 19, 20, and 9. The first two are remote nodes of the gas source to meet the power balance requirements, and the latter is to ensure the normal operation of GB. Fig. 9 From (9c), we can see that the risk is mainly concentrated in node 5. This is because the heat load of node 5 is large. In order to ensure power balance, the load shedding amount is positively correlated with the heat load demand.
[0225] IV. Impact of scenario generation methods on risk assessment
[0226] The integrated energy system risk assessment method based on R-vine copula proposed in this embodiment takes into account the randomness of the operation scenario and the correlation of the risk factors. The impact of the three scenario generation methods on risk assessment is compared as shown in Table 2:
[0227] Table 2 Risk scenario generation method
[0228]
[0229] Among them, method A is the scenario generation method proposed in this example; method B is to use the two-dimensional Frankcopula function to jointly model the load data of wind and solar output using Dest software to simulate, and sample and reduce it respectively to obtain the scenario probability; method C is to determine the scenario, without considering the scenario uncertainty of source and load risk factors, and the wind and solar power output and multi-energy load demand are regarded as fixed curves, which are obtained through Monte Carlo and clustering. The above three methods only consider two seasons, namely the heating season and other seasons, and the formula for calculating the number of scenarios is:
[0230]
[0231] In formula (47), Ω, Γ, and Λ are seasonal scenes. is the number of R-vine copula scenarios in season i, is the wind and solar power output and the number of electricity, gas and heat load scenarios in season j, is the number of scenarios in season k, and the risk scenario probability π(s) is the product of the probabilities of each scenario.
[0232] Considering that the correlation between risk factors may have a lag, such as Fig.10 As shown in Figure 1, a cross-correlation analysis is performed on the generated risk factor sequence, the time offset Δt is selected from -2 to 2 hours, and the Spearman correlation coefficient is introduced to represent the cross-correlation function value. The analysis results are shown in Figure 1. Fig.11 As shown, (11a) is the cross-correlation function diagram of photovoltaic output-electric load (other seasons), (11b) is the cross-correlation function diagram of electricity-heat load (heating season), (11c) is the cross-correlation function diagram of wind-solar output (other seasons, the sequence selects 7-20 periods), the horizontal axis is the time offset, and the vertical axis is the expected value of the correlation coefficient of each scene in the corresponding season. Fig.11 From (11c), we can see that method A and method B can better reflect the correlation between wind and solar energy. Fig.11 From (11a) and (11b), we can see that in terms of the correlation between PV output and electric load and the correlation between electric and thermal load, only method A has a better performance, and there is a certain lag in the correlation between PV output and electric load.
[0233] The correlation matrices between risk factors under Method A and Method B are as follows: Fig.12 As shown in (12a) and (12b), the maximum correlation coefficient of each lag time point is selected. It can be seen that method A can better retain the correlation of risk factors such as wind and solar output correlation, load correlation and source-load correlation.
[0234] The time series diagram of the system operation comprehensive risk value RISK under different risk scenario generation methods (assuming that the weights of each indicator are consistent) is as follows Fig.13As shown, it can be concluded that, compared with other methods, method A (i.e., the method proposed in this embodiment) can avoid underestimation of risk values. This is because there is a correlation between risk factors, which makes different types of risks likely to overlap in the same period. Fully considering the correlation between risk factors can avoid ignoring this overlapping effect, which is more in line with the actual system operation situation.
[0235] V. Impact of Coupling Methods on Risk Assessment
[0236] In order to analyze the impact of different coupling modes on the risk assessment of the electricity-gas-heat integrated energy system, this embodiment sets 6 coupling modes, and the comprehensive expected values of their operating risks are shown in Table 3:
[0237] Table 3 Impact of different coupling modes on risk assessment of integrated energy system
[0238]
[0239]
[0240] The following conclusions can be drawn from Table 3: 1) Case 1 has the most comprehensive energy coupling equipment and the lowest comprehensive operation risk. This is because through energy coupling equipment, the comprehensive energy system can integrate and utilize multiple energy resources, improve energy utilization efficiency, and reduce dependence on a single energy source, thereby reducing the overall system operation risk. 2) NGU, GB, EB, and P2G have a cumulative effect on reducing system operation risks, and the impact decreases in turn. This is because different coupling devices have different coupling effects in the system, and the coupling effects of the above devices decrease in turn in the system.
[0241] VI. The impact of risk preference on risk assessment
[0242] Consider the risk assessment results under different risk preferences, that is, the impact of safety, reliability, and economy indicators on risk assessment results under different weights. Consider only the weights of the above three indicators, assuming that the sub-indicators under each indicator have equal weights, the expected value of the comprehensive risk value of system operation in a day is as follows: Fig.14 shown.
[0243] Depend on Fig.14It can be seen that the economic index weight has a greater impact on the comprehensive risk value and shows a negative correlation; when the economic index weight is quantitative, the system reliability index weight and the comprehensive risk value show a positive correlation. Based on the above analysis, the higher the economic weight, the lower the system comprehensive operation risk evaluation value. This is because the probability flow objective function is the optimal operating cost, and the economic risk is lower than other risks. Therefore, increasing its proportion can reduce the comprehensive risk evaluation value; if the economic index weight remains unchanged, the higher the reliability index weight / the lower the safety index weight, the higher the system operation comprehensive risk evaluation value. However, this rule is not absolute. This is related to the load shedding penalty coefficient in the flow calculation and the size of the safety domain interval of the energy network lines and nodes.
[0244] In summary, this embodiment proposes a comprehensive energy system risk assessment method based on R-vine copula, which fully considers the multi-faceted operational risks brought to the comprehensive energy system by the volatility of new energy output and the uncertainty of load demand, and verifies the effectiveness of the proposed assessment method by performing case analysis on an improved IEEE 30-node power network, a Belgian 20-node natural gas network, and a 6-node heat network consisting of an electric-gas-heat comprehensive energy system. The results show that:
[0245] 1) Considering the randomness of the integrated energy system operation scenarios and the correlation of risk factors can avoid underestimation of risk assessment and improve the degree of risk control.
[0246] 2) The comprehensive evaluation value of the operation risk of the integrated energy system is affected by the energy coupling equipment and risk preference.
[0247] Example 2
[0248] This embodiment provides an integrated energy system risk assessment device based on R-vine copula, which can implement the method described in Example 1, including: a risk assessment indicator module, a multi-dimensional risk factor joint probability distribution model generation module, an integrated energy system typical operation risk scenario generation module and an integrated energy system operation risk value calculation module.
[0249] The risk assessment index module is used to establish a risk assessment index system for an integrated energy system. The multidimensional risk factor joint probability distribution model generation module is used to fit the marginal probability distribution of risk factors using a non-parametric kernel density estimation method and a normal distribution, and to construct a multidimensional risk factor joint probability distribution model based on the marginal probability distribution of risk factors using the R-vine copula method. The integrated energy system typical operation risk scenario generation module is used to perform Monte Carlo sampling based on the multidimensional risk factor joint probability distribution model, and to reduce and cluster the risk factor values obtained by solving the inverse probability transformation to obtain typical operation risk scenarios for the integrated energy system. The integrated energy system operation risk value calculation module is used to quantitatively calculate the operation risk value of the integrated energy system using the AHP-Topsis method based on the integrated energy system risk assessment index system and the typical operation risk scenarios of the integrated energy system.
[0250] The above description of the embodiments is to facilitate the understanding and use of the invention by those skilled in the art. It is obvious that those skilled in the art can easily make various modifications to these embodiments and apply the general principles described herein to other embodiments without creative work. Therefore, the present invention is not limited to the above embodiments, and improvements and modifications made by those skilled in the art based on the disclosure of the present invention without departing from the scope of the present invention should be within the scope of protection of the present invention.
Claims
1. A comprehensive energy system risk assessment method based on R-vine copula, characterized in that: include: Establish a comprehensive energy system risk assessment indicator system; The marginal probability distribution of risk factors was fitted by nonparametric kernel density estimation method and normal distribution. According to the marginal probability distribution of risk factors, the R-vine copula method was used to construct a joint probability distribution model of multidimensional risk factors. Monte Carlo sampling is performed on the basis of the multi-dimensional risk factor joint probability distribution model, and risk factor values are obtained by inverse probability transformation and then clustering is performed to obtain typical operational risk scenarios of the integrated energy system; According to the risk assessment index system of the integrated energy system and the typical operation risk scenarios of the integrated energy system, the AHP-Topsis method is adopted to quantitatively calculate the operation risk value of the integrated energy system.
2. The integrated energy system risk assessment method based on R-vine copula according to claim 1 is characterized in that: The comprehensive energy system risk assessment index system includes safety indicators; The safety indicators include power grid safety risk indicators, gas grid safety risk indicators and heat network safety risk indicators, wherein the power grid safety risk indicators are calculated based on the electric load weight coefficient and the power exceeding limit degree of each branch of the power grid, the gas grid safety risk indicators are calculated based on the gas load weight coefficient and the gas pressure exceeding limit degree of each node of the gas grid, and the heat network safety risk indicators are calculated based on the heat load weight coefficient and the temperature exceeding limit degree of each node of the heat network.
3. The integrated energy system risk assessment method based on R-vine copula according to claim 1 is characterized in that: The comprehensive energy system risk assessment index system includes reliability index; The reliability indicators include a power grid reliability risk indicator, a gas grid reliability risk indicator and a heat grid reliability risk indicator, wherein the power grid reliability risk indicator is calculated based on the load weight coefficient of each node in the power grid and the energy shortage ratio, the gas grid reliability risk indicator is calculated based on the load weight coefficient of each node in the gas grid and the energy shortage ratio, and the heat grid reliability risk indicator is calculated based on the load weight coefficient of each node in the heat grid and the energy shortage ratio.
4. The integrated energy system risk assessment method based on R-vine copula according to claim 1 is characterized in that: The comprehensive energy system risk assessment index system includes economic indicators; The economic index is the proportion of risk cost to total operating cost, and the risk cost includes the cost caused by load interruption and waste of new energy.
5. A comprehensive energy system risk assessment method based on R-vine copula according to any one of claims 1 to 4, characterized in that: The risk factors include wind power output, photovoltaic output, electric load demand, gas load demand and thermal load demand. The process of fitting the marginal probability distribution of risk factors using non-parametric kernel density estimation method and normal distribution is as follows: Based on Gaussian kernel function: Fitting the marginal probability distribution of wind power output and photovoltaic output at time t and in, and are the output of wind turbine and photovoltaic at time t, and are the output of wind turbine and photovoltaic power generation at time t on day d, h is the smoothing coefficient, and n is the sample size; The marginal probability distribution of the electric load demand, gas load demand and heat load demand is described by normal distribution, where the marginal probability distribution of the electric load demand is f(L e ), the expression is as follows: Among them, μ e and σ e are the mean and variance of the historical data of electric load, and P is the electric load value.
6. The integrated energy system risk assessment method based on R-vine copula according to claim 5 is characterized in that: The process of obtaining the typical operational risk scenario of the integrated energy system is as follows: In the multi-dimensional risk factor joint probability distribution model, Monte Carlo random sampling is performed on the joint probability distribution of each period to obtain the cumulative probability distribution of the risk factor in each period; When the risk factor is the electric load demand, gas load demand or heat load demand, the corresponding CDF of the cumulative probability distribution is inversely transformed to obtain the specific value of the risk factor in each period; when the risk factor is wind power output or photovoltaic output, the spline difference method is used to obtain the specific value of the risk factor in each period; The K-means algorithm is used to reduce and cluster the specific values of the risk factors obtained to obtain the typical operation scenarios of the integrated energy system.
7. A comprehensive energy system risk assessment method based on R-vine copula according to any one of claims 1 to 4, characterized in that: The specific process of constructing the multi-dimensional risk factor joint probability distribution model using the R-vine copula method is as follows: Get the risk factor data set X = {X1, X2, ...X n }, set node N = X, calculate the Kendall rank correlation coefficient of the empirical Copula function in N, and use Prim's algorithm to build an R-vine model, where the R-vine model includes a series of maximum spanning trees, each edge E of each tree corresponds to a connected binary Copula function, and the weight of each edge is the Kendall rank correlation coefficient of the empirical Copula function; The maximum likelihood estimation method is used to estimate the binary Copula function parameter values in each maximum spanning tree, and the Euclidean distance between the Kendall rank correlation coefficients of the binary Copula function to be determined and the empirical Copula function is compared. The optimal binary Copula function between the nodes in each tree is determined based on the AIC information criterion, and a multidimensional risk factor joint probability distribution model is constructed. The expression of the multidimensional risk factor joint probability distribution model is as follows: (F(x j(e) |x D(e) ),F(x k(e) |x D(e) )) Among them, f(x1, x2, ...x n ) is the joint probability distribution model of multi-dimensional risk factors, f k (x k ) is the risk factor x k The marginal probability distribution of E = {E1, E2, ... E n-1 } is the edge set, C j(e),k(e)|D(e) is the binary Copula function corresponding to edge e, j(e), k(e) are the nodes connected by edge e, D(e) is the condition, F(x j(e) |x D(e) ) and F(x k(e) |x D(e) ) are the probability distributions of the nodes connected by edge e under this condition.
8. The integrated energy system risk assessment method based on R-vine copula according to claim 1 is characterized in that: The calculation formula of the operation risk value of the integrated energy system is as follows: Among them, RISK is the operational risk value of the integrated energy system, π(s) is the probability of occurrence of a typical operational risk scenario s of the integrated energy system, R1, R2, ..., R n T is the various indicators in the comprehensive energy system risk assessment indicator system. R (R1, R2, ..., R n ) is a comprehensive evaluation function constructed using the AHP-Topsis method.
9. The integrated energy system risk assessment method based on R-vine copula according to claim 8 is characterized in that: The comprehensive evaluation function T R (R1, R2, ..., R n ) is constructed as follows: Construct the weighted normalized matrix B of the comprehensive energy system risk assessment index system: The element b in the matrix ij =ω j z ij ,ω j is the weight of each indicator, obtained through the analytic hierarchy process AHP, z ij is the element in the risk index normalization matrix Z, i = 1, 2, ..., m, j = 1, 2, ..., n; The comprehensive evaluation function T is constructed by using the superior and inferior solution distance method Topsis R (R1, R2, ..., R n ), the specific process is: based on the weighted normalization matrix B, determine the positive and negative ideal solution sets B + and B-: Calculate the Euclidean distance between each element in the weighted normalized matrix B and the corresponding positive ideal solution and negative ideal solution and Calculate the closeness C of each element to the ideal solution i : Construct comprehensive evaluation function T R (R1, R2, ..., R n ), T R (R1, R2, ..., R n )=1 / C i .
10. An integrated energy system risk assessment device based on R-vine copula, characterized in that: include: Risk assessment indicator module, multi-dimensional risk factor joint probability distribution model generation module, integrated energy system typical operation risk scenario generation module and integrated energy system operation risk value calculation module; Wherein, the risk assessment indicator module is used to establish a comprehensive energy system risk assessment indicator system; The multidimensional risk factor joint probability distribution model generation module is used to use a non-parametric kernel density estimation method and a normal distribution to fit the marginal probability distribution of risk factors, and to construct a multidimensional risk factor joint probability distribution model using an R-vine copula method based on the marginal probability distribution of risk factors; The typical operation risk scenario generation module of the integrated energy system is used to perform Monte Carlo sampling based on the multi-dimensional risk factor joint probability distribution model, and obtain the risk factor values through inverse probability transformation and then perform reduction clustering to obtain the typical operation risk scenario of the integrated energy system; The integrated energy system operation risk value calculation module is used to quantitatively calculate the operation risk value of the integrated energy system using the AHP-Topsis method based on the integrated energy system risk assessment index system and the typical operation risk scenarios of the integrated energy system.
Citation Information
Cited By
Regulation and control strategy of underwater wind-solar-storage integrated multi-energy complementary system in extreme high-temperature scene
CN119726743A
A water, wind, light and storage integrated multi-energy complementary system regulation strategy under an extremely high temperature scenario
CN119726743B
