A method for reducing the load risk of a park based on the capacity configuration of PIES devices

Abstract

The present invention discloses a method for reducing the campus load risk based on the capacity configuration of PIES devices. The method includes: establishing an energy price scenario generation model, mainly based on the Copula function and the DCC model to model the seasonal joint probability distribution of natural gas prices and electricity prices for generating energy price scenarios; according to the established energy price scenario generation model, bringing the generated energy price scenarios into the PIES device capacity configuration model to establish a CVaR-based PIES risk avoidance control model; using the Linprog function in MATLAB to call the simplex algorithm to solve the PIES risk avoidance control model to obtain the PIES device capacity configuration schemes at different confidence levels. The present invention solves the problem of the loss of load shortage risk caused by the lack of consideration of the seasonal correlation between natural gas prices and electricity prices in the prior art, and solves the technical problem of avoiding the load shortage risk from the aspect of PIES device capacity configuration.

Description

Technical Field

[0001] The present invention relates to the field of integrated energy systems, and particularly to a method for reducing the load risk of a park based on the capacity configuration of PIES devices. Background Art

[0002] The park-level integrated energy system (PIES) uses natural gas and electricity purchased from the external power grid as fuels, and meets the internal heating and electrical load demands of the park through energy conversion devices. The strong volatility characteristics of natural gas prices and electricity prices significantly affect the economy of PIES energy supply. In severe cases, it will lead to the risk loss of load shortage due to insufficient equipment configuration capacity of PIES. Therefore, at the stage of PIES device capacity configuration, it is necessary to take into account the strong uncertainty of natural gas prices and electricity prices, and avoid the risk loss of load shortage caused by energy price uncertainty by jointly optimizing the device configuration capacity and operation strategy of PIES.

[0003] Currently, in the research on PIES operation optimization, existing research has taken renewable energy as a risk factor and used the Conditional value-at-risk (CVaR) method to measure the risk loss of load shortage. There is little research taking energy prices as a risk factor. However, energy prices are an important risk source faced by PIES and will bring serious risk losses of load shortage to PIES. In addition, when dealing with energy uncertainty, current research independently establishes the probability distribution of energy price uncertainty, and then simulates its uncertainty by sampling discrete probability scenarios. However, when establishing the energy price probability distribution, these research works ignore the temporal relationship and correlation between energy prices, which will significantly affect the operation boundary of PIES, and then lead to deviations in the device capacity configuration of PIES, affecting the reliability of energy supply. The above work only conducts risk avoidance research from the aspect of operation strategy, and there is little research on avoiding the risk loss of load shortage in the aspect of PIES device capacity configuration. Summary of the Invention

[0004] The present invention provides a method for reducing the load risk of a park based on the capacity configuration of PIES devices. The present invention solves the problem of risk loss of load shortage caused by not considering the seasonal correlation between natural gas prices and electricity prices in the prior art, and solves the technical problem of avoiding the risk of load shortage from the level of PIES device capacity configuration, as described in detail below:

[0005] A method for reducing the load risk of a park based on the capacity configuration of PIES devices, the method includes:

[0006] Build an energy price scenario generation model. The specific modeling steps are to conduct seasonal joint probability distribution modeling on natural gas prices and electricity prices based on the Copula function and the DCC model, and then use Monte Carlo sampling to generate energy price scenarios;

[0007] According to the established energy price scenario generation model, bring the generated energy price scenarios into the PIES device capacity configuration model to establish a CVaR-based PIES risk avoidance control model;

[0008] Use the Linprog function in MATLAB to call the simplex algorithm to solve the PIES risk avoidance control model, and obtain the PIES device capacity configuration plan at different confidence levels.

[0009] Among them, the seasonal joint probability distribution modeling of natural gas prices and electricity prices based on the Copula function and the DCC model is as follows:

[0010] Select the bivariate t-Copula function as the connection function:

[0011]

[0012] In the formula, represents the correlation coefficient between natural gas prices and electricity prices in season w; υ represents the degree of freedom; and represent the auxiliary variables of the bivariate integral operation in season w; represents the inverse operation of the t-distribution with degree of freedom υ; x gas,w and y elec,w are random variables;

[0013]

[0014] Among them,

[0015]

[0016]

[0017]

[0018] In the formula, q gase l e,w represents the covariance between natural gas prices and electricity prices in season w; q gas,w and q elec,w represent the autocovariances of natural gas prices and electricity prices in season w respectively; W is the total number of seasonal cycles; ε gas,w and ε elec,wThey respectively represent the prediction errors of natural gas price and electricity price in season w; γ and τ respectively represent the regression coefficients to be estimated.

[0019] Furthermore, the CVaR-based PIES risk avoidance control model is as follows:

[0020]

[0021] In the formula, VaR α represents the risk value of the total cost at the confidence level α; represents the total cost in scenario s; ρ rs represents the probability of the occurrence of scenario s; S is the set of joint price scenarios.

[0022] Among them, the operation constraint conditions of the CVaR-based PIES risk avoidance control model are: the upper and lower limit constraints of the output power of device i and the power balance constraint of the jth type of energy:

[0023]

[0024]

[0025] In the formula, represents the jth type of energy input into the ith device at time t in season w in scenario s; represents the load demand corresponding to the jth type of energy at time t in season w, represents the output power of the ith device at time t in season w in scenario s, Q i represents the configured capacity of the ith device; η i represents the energy conversion efficiency of the ith device.

[0026] The beneficial effects of the technical solution provided by the present invention are as follows:

[0027] 1. The present invention takes into account the seasonal correlation between natural gas price and electricity price, and establishes a seasonal joint probability distribution model based on the Copula function and the Dynamic Conditional Correlation (DCC) model to capture their correlation;

[0028] 2. The present invention applies the joint probability distribution model to the PIES risk avoidance device capacity configuration model based on the CVaR method, and establishes a PIES device capacity configuration method for avoiding the risk loss of load shedding. The risk loss of load shedding can be avoided by coordinating the configured capacities and operation strategies of each device of the PIES, and the obtained device capacity configuration plan can provide a basis for engineering personnel to overhaul the park load, improving the safety of the park. Description of the Drawings

[0029] Figure 1 This is a schematic diagram of the typical structure of the integrated energy system in the park in the present invention;

[0030] Figure 2 This is a framework diagram of the method for reducing the park load risk based on the PIES device capacity configuration in the present invention;

[0031] Figure 3 This is a schematic diagram of the electric - heat load and light intensity curves of the example system of the integrated energy system in the park in the present invention;

[0032] Figure 4 This is a scatter plot of NGP and EP on typical days in different seasons considering and not considering seasonal correlation in the present invention;

[0033] Figure 5 This is a schematic diagram of the operation curves of each device of PIES on a typical winter day in the present invention;

[0034] Figure 6 This is a schematic diagram of the change curve of CVaR with the number of different scenarios in the present invention. Detailed implementation manners

[0035] To make the objectives, technical solutions and advantages of the present invention clearer, the following further describes the embodiments of the present invention in detail.

[0036] Example 1

[0037] The embodiment of the present invention provides a method for reducing the park load risk based on the PIES device capacity configuration, and the method includes the following steps:

[0038] First, considering the seasonal correlation between natural gas price and electricity price, a seasonal joint probability distribution model based on the Copula function and DCC model is established to capture their correlation; subsequently, the seasonal joint probability distribution model is applied to the PIES risk - aversion device capacity configuration model based on the CVaR method, and a risk - aversion control method is established, which can avoid the loss of load - shortage risk by coordinating the configuration capacity and operation strategy of each device of PIES. The method proposed in the embodiment of the present invention can provide a device capacity configuration scheme for PIES and can avoid the loss of load - shortage risk.

[0039] Step 1: Establish an energy price scenario generation model;

[0040] I. Modeling of the marginal distributions of natural gas price and electricity price

[0041] Generally, the prediction error of energy prices is assumed to follow a normal distribution. The sum of the prediction error of energy prices and the mean represents the actual energy price. Therefore, the actual energy price and the prediction error of energy prices follow a probability distribution of the same shape. Accordingly, the marginal distributions of natural gas prices and electricity prices are modeled using a normal distribution, and their probability density functions are shown in Equations (1) and (2).

[0042]

[0043]

[0044] where x gas,w and y elec,w represent the random variables of natural gas prices and electricity prices in season w, respectively; f(x gas,w ) and g(y elec,w ) represent the probability density functions of natural gas prices and electricity prices, respectively; μ gas,w and μ elec,w represent the means of natural gas prices and electricity prices in season w, respectively; σ gas,w and σ elec,w represent the standard deviations of natural gas prices and electricity prices in season w, respectively.

[0045] II. Modeling of the Seasonal Joint Probability Distribution of Natural Gas Prices and Electricity Prices Based on the Copula Function and the DCC Model;

[0046] Due to the seasonal differences in energy supply and demand, energy prices also have seasonal characteristics. Considering the interdependence of the energy market and the fact that the correlation between natural gas prices and electricity prices changes in different seasons when transitioning from one season to another, showing seasonal correlation. The embodiments of the present invention use the Copula function and the Dynamic Conditional Correlation (DCC) model to establish the seasonal joint probability distribution between natural gas prices and electricity prices to generate joint price scenarios for different seasons, so as to simulate their uncertainty.

[0047] Suppose H(x gas,w , y elec,w ) represents the joint probability distribution function of the random variables x gas,w and y elec,w . There exists a connection function L(·) that satisfies Equation (3).

[0048] H(x gas,w , y elec,w ) = L(F(x gas,w ), G(y elec,w )) (3)

[0049] where F(xgas,w ) and G(y elec,w ) represent the cumulative probability distributions of the natural gas price and the electricity price in season w, respectively.

[0050] Due to the superiority of the bivariate t-Copula function in capturing the tail correlation caused by extreme events (such as extremely cold weather, emergency policy situations, etc.), the embodiment of the present invention selects the bivariate t-Copula function as the copula function, and the bivariate t-Copula function is shown in Equation (4).

[0051]

[0052] In the formula, represents the correlation coefficient between the natural gas price and the electricity price in season w; υ represents the degree of freedom; and represent the auxiliary variables of the bivariate integral operation in season w; represents the inverse operation of the t-distribution with degree of freedom υ.

[0053] In Equation (4), due to the advantage of parameter estimation, the DCC model is used to calculate The DCC model uses the regression method to obtain the dynamic covariance and autocovariance of the natural gas price and the electricity price to calculate the correlation coefficient in season w, as shown in Equation (5).

[0054]

[0055] Among them,

[0056]

[0057]

[0058]

[0059] In the formula, q gase l e,w represents the covariance between the natural gas price and the electricity price in season w; q gas,w and q elec,w represent the autocovariances of the natural gas price and the electricity price in season w, respectively; W is the total number of season cycles; ε gas,w and ε elec,w represent the prediction errors of the natural gas price and the electricity price in season w, respectively; γ and τ represent the regression coefficients to be estimated, satisfying γ + τ < 1, γ > 0 and τ > 0, and γ and τ can be estimated by using the maximum likelihood estimation method according to the historical data of the natural gas price and the electricity price.

[0060] Equations (6)-(8) are the recursive expressions for the covariance and autocovariance of natural gas prices and electricity prices. That is, the correlation between natural gas prices and electricity prices in season w is determined by the covariance, autocovariance, and prediction error in the previous season w, where the prediction error is obtained by neural network prediction technology.

[0061] According to the established joint probability distribution function (4) considering seasonal correlation, a set of joint price scenarios S in season w is generated using Monte Carlo simulation. N represents the number of joint price scenarios in season w, and the method for determining the value of N will be introduced in the next section.

[0062] Step 2: Establish a risk aversion control model for PIES based on CVaR

[0063] 1. The PIES device model is as follows:

[0064] 1) Combined heat and power unit

[0065] The CHP produces electrical energy and heat by consuming natural gas, and the model is shown in Equations (9) and (10):

[0066]

[0067]

[0068] In the equations, and represent the electrical output power and heat output power of the CHP at time t, respectively; G CHP (t) represents the natural gas power consumed by the CHP at time t; and represent the electrical output efficiency and heat output efficiency of the CHP, respectively.

[0069] 2) Gas boiler

[0070] The GB is another heating device that provides heat by burning natural gas, and the model is shown in Equation (11):

[0071]

[0072] In the equations, represents the heat output power of the GB at time t; G GB (t) represents the natural gas power consumed by the GB at time t; represents the heat output efficiency of the GB.

[0073] 3) Electric boiler

[0074] The EB produces heat by consuming electrical energy, and the model is shown in Equation (12):

[0075]

[0076] In the formula, represents the thermal output power of EB at time t; represents the electric power consumed by EB at time t; represents the thermal output efficiency of EB.

[0077] 4) Photovoltaic

[0078] The electrical output power of PV is mainly determined by factors such as the light intensity irradiated on the PV surface and the physical parameters of PV. The input-output relationship of PV is shown in Equation (13):

[0079]

[0080] In the formula, P PV (t) represents the electrical output power of the photovoltaic at time t; Q PV represents the peak capacity (kWp) of PV; f PV represents the power derating factor of the photovoltaic, which is used to characterize the output power reduction caused by dirt and aging on the PV surface, and generally takes 0.9; G T (t) represents the actual light intensity (kW / m2) at time t; G T,STC represents the light intensity under standard test conditions, and generally takes 1 kW / m2.

[0081] The risk aversion device capacity configuration model based on CVaR established in the embodiments of the present invention aims to minimize the CVaR of the total cost considering the load shedding amount, and determines the PIES control scheme at different confidence levels by optimizing all generated joint price scenarios with seasonal correlations.

[0082] 2. Objective function of the risk aversion control model

[0083] The objective function of the established risk aversion device capacity configuration model is shown in Equation (14).

[0084]

[0085]

[0086]

[0087]

[0088]

[0089]

[0090]

[0091] VaR a = min{ζ: Pr(Π Tot ≤ ζ) ≥ α} (21)

[0092] Where VaR α represents the risk value of the total cost at the confidence level α, which refers to the minimum total cost ζ that does not exceed the probability of α; Pr(·) represents the probability measure; represents the total cost in scenario s; Π Inv represents the investment cost; represents the fuel cost in scenario s; represents the maintenance cost in scenario s; represents the load shedding penalty cost in scenario s; r represents the discount rate; ρ rs represents the probability of the occurrence of scenario s; l i represents the service life of the i-th device; Ω represents the set of candidate devices; c inv,i represents the unit capacity price of the i-th device; Q i represents the configured capacity of the i-th device; c main,i represents the maintenance cost per unit power of the i-th device; represents the output power of the i-th device at time t in season w under scenario s; T represents the total annual operating time, which is 8760 hours; represents the load shedding amount at time t in season w under scenario s; △t represents the system operating time interval, with a value of 1 hour; and respectively represent the natural gas purchase volume and the external power purchase volume of the system at time t in season w under scenario s; and respectively represent the natural gas price and the electricity price at time t in season w under scenario s; c Loss represents the unit penalty cost for load shedding; α is the confidence level.

[0093] Since equation (14) is a non-linear optimization problem, an auxiliary variable is introduced for linearization. After linearization, equation (14) can be reformulated as equation (22) and two linear inequalities, as shown in equations (23) and (24).

[0094]

[0095]

[0096]

[0097] The constraint conditions include investment constraints and operation constraints. The investment constraints are mainly the installation capacity constraints of equipment i, as shown in Equation (25). The operation constraints are mainly the upper and lower limits of the output power of equipment i and the power balance constraints of the jth type of energy (electric / thermal energy), as shown in Equations (26) and (27).

[0098]

[0099]

[0100]

[0101] In the formula, represents the maximum configuration capacity of the ith device; η i represents the energy conversion efficiency of the th device; represents the jth type of energy input into the ith device at the tth moment of the wth season in the s scenario; represents the corresponding load demand of the jth type of energy at the tth moment of the wth season.

[0102] Therefore, the risk - aversion device capacity configuration model based on CVaR proposed in the present invention, as shown in Equations (14)-(27), is a typical linear optimization problem, and the simplex algorithm is used for solution by calling the Linprog function in MATLAB.

[0103] In the joint price scenario set S, the number of scenarios N affects the accuracy of the device capacity configuration result, and N is determined by the convergence criterion of CVaR, as shown in Equations (28) and (29).

[0104]

[0105] Z k+1 =Z k +△N (29)

[0106] In the formula, CVaR(Z k ) represents the CVaR value when the number of scenarios is Zk; Z k represents the number of joint price scenarios at the kth iteration; ε represents the convergence error; △N represents the iteration step size; if Equation (28) is satisfied, the calculation of Equation (14) is terminated, and the value of N is set to Nk. Furthermore, the device capacity configuration scheme of PIES under different confidence levels can be obtained.

[0107] Application example

[0108] Taking Figure 1 the PIES shown as an example as the test system to verify the effectiveness of the risk - aversion device capacity configuration method proposed by this method. According to the limitation of the equipment installation space, the maximum installation capacity of each device Set it to 5000 kW and the discount rate r to 8%. The example uses a typical seasonal day to represent the annual load and light intensity, so as to reduce the total optimization variables of the equipment capacity configuration model, that is, select the typical day load and light intensity in winter, summer and spring / autumn to represent the corresponding seasonal characteristics. The electricity / heat load curves and light intensity curves of the typical days in winter, summer and spring / autumn are as Figure 3 shown. The maximum electricity load is 1248 kW, the maximum heat load is 2100 kW, △t is set to 1 hour, and c Loss is set to 10 $ / kWh. The economic and technical parameters of the equipment are shown in Table 1.

[0109] Table 1 Economic and Technical Parameters of Equipment

[0110]

[0111] 1 Scenario Generation

[0112] Based on the historical data of natural gas prices and electricity prices from 2000 to 2019 provided by the US Energy Information Administration, the means and standard deviations of natural gas prices and electricity prices on typical days in different seasons can be obtained through statistical analysis. As shown in Table 2, the estimated values of γ and τ in equations (6)-(8) are 0.0258 and 0.664 respectively, and the estimated value of υ is 13. The correlation coefficients in winter, summer and spring / autumn are 0.107, 0.12 and 0.11 respectively.

[0113] Table 2 Means and Standard Deviations of Natural Gas Prices and Electricity Prices on Typical Days in Different Seasons

[0114]

[0115] According to equation (27), N is set to 2000, and the convergence curve of CVaR with respect to Zk is shown later. Furthermore, 2000 natural gas price and electricity price scenarios considering seasonal correlation and not considering seasonal correlation are generated using the Monte Carlo method, and the probability of each scenario occurring is 0.0005.

[0116] To illustrate the influence on scenario generation, Figure 4 two-dimensional scatter plots of natural gas prices and electricity prices considering seasonal correlation and not considering correlation on typical days in different seasons are given.

[0117] In Figure 4 , the distribution of scenarios considering seasonal correlation is more discrete, and more extreme scenarios are sampled than when not considering correlation. The reason is it increases the probability of extreme scenarios being sampled. Figure 4 The differences in the price scenario distributions in

[0118] 2. Equipment Capacity Configuration Results and Discussion

[0119] To verify the effectiveness of the method of the present invention, two scenarios are constructed to compare and illustrate the effectiveness of the equipment capacity configuration method:

[0120] Scenario I: Risk - aversion Equipment Capacity Configuration Considering Seasonal Correlation between Natural Gas Price and Electricity Price

[0121] Scenario II: Risk - aversion Equipment Capacity Configuration without Considering Correlation between Natural Gas Price and Electricity Price

[0122] Optimize the equipment capacities of PIES at different confidence levels in Scenario I and Scenario II. The capacity configuration results are shown in Table 3.

[0123] Table 3 Capacity Configuration Results at Different Confidence Levels in Scenarios I and II

[0124]

[0125] As can be seen from Table 3, as the confidence level increases, the capacities of CHP, EB, and PV increase, while the capacity of GB decreases. In Scenarios I and II, at a confidence level of 90% - 95%, the capacity configuration of EB remains unchanged, the capacity configurations of GB decrease by 7 kW and 22 kW respectively, and the capacity configurations of CHP increase by 4 kW and 13 kW respectively. This shows that since CHP can provide both electric / thermal power, part of the thermal output power of GB is replaced by CHP.

[0126] By comparing Scenario I and Scenario II, at the same confidence level, the capacity configurations of CHP, EB, and PV in Scenario I are greater than or equal to those in Scenario II. The capacity configuration of GB in Scenario I is less than or equal to that of GB in Scenario II. Since the equipment capacity configuration of PIES needs to meet the maximum output power condition during operation, therefore, combined with the optimal operation results of PIES, analyze the reasons for the difference in equipment capacity configuration. Taking the 85% confidence level as an example, the operation curves of Scenarios I and II on a typical winter day are as Figure 5 shown.

[0127] Since Scenario I considers the seasonal correlation of energy prices, the energy price scenarios in Scenario I are more dispersed, with more extreme price scenarios as Figure 3 shown, resulting in a stronger risk loss in Scenario I than in Scenario II. To cope with the stronger risk loss, PIES reduces the purchase of electricity from the external grid during the period of 18:00 - 22:00, as Figure 5 (e) shows. To maintain the electric power balance, the maximum output power of CHP increases during the period of 18:00 - 22:00, as Figure 5 (a) shows. InFigure 5 (d) From 8:00 to 16:00, the output power of PV increases to improve the risk aversion ability of the system. Therefore, the capacity ratio of CHP and PV in Scenario I is larger than that in Scenario II. Due to the thermoelectric ratio limit of CHP, CHP can only supply part of the heat load. To achieve thermal power balance, the maximum output power of EB increases from 8:00 to 16:00, as shown in Figure 5 (c). The maximum output power of GB decreases from 8:00 to 16:00, as shown in Figure 5 (b), so as to reduce the natural gas purchase volume and alleviate the risk loss faced by PIES. Therefore, compared with Scenario II, the capacity configuration of EB increases, while the capacity configuration of GB decreases. Table 4 lists the total costs and CVaR of Scenario I and Scenario II at different confidence levels.

[0128] Table 4 Total costs and CVaR of Scenario I and Scenario II at different confidence levels (unit: $10,000)

[0129]

[0130] As can be seen from Table 4, at confidence levels of 80%, 85%, 90%, 95% and 99%, the total costs of Scenario I and Scenario II increase by 4.42%, 5.10%, 5.85%, 6.83% and 8.17% respectively. While the CVaR values decrease by 12.22%, 17.84%, 27.01%, 36.91% and 71.41% respectively. The magnitude of the CVaR value reflects the level of the load shedding risk, further indicating that PIES faces a relatively high load shedding risk when pursuing a low total cost. In addition, it can be found that at the same confidence level, the total cost of Scenario I is higher than that of Scenario II, while the CVaR value of Scenario I is lower than that of Scenario II. For example, at a 95% confidence level, the total cost of Scenario I is $137,500 higher than that of Scenario II, while the CVaR value is $278,400 less than that of Scenario II, indicating that in order to avoid an additional load shedding risk of $278,400, Scenario I has to pay an additional total cost of $137,500 compared with Scenario II. The main reason is that Scenario I takes into account the seasonal correlation of energy prices and requires a higher total cost to avoid the additional load shedding risk brought by the seasonal correlation. While Scenario II ignores the impact of the correlation between energy prices on the equipment capacity configuration of PIES during the equipment capacity configuration, resulting in a higher load shedding risk for Scenario II.

[0131] Taking the calculation of CVaR under deterministic conditions in Scenario I as an example to illustrate the impact of the number of scenarios on CVaR. Let ε be 1% and △N be 500. The convergence curves of CVaR under different numbers of scenarios are as shown in Figure 6 shown.

[0132] As shown in Figure 6As shown, when the number of scenarios is greater than or equal to 2000, the convergence curve of CVaR tends to be stable. The percentage of the difference in CVaR between the 2000-scenario and the 2500-scenario is 0.14%, meeting the convergence criterion. Therefore, when calculating the capacity configuration of the PIES device in the present invention, N is set to 2000.

[0133] In summary, the present invention proposes a risk avoidance control method for PIES based on CVaR and Copula theory, and the conclusions are as follows:

[0134] 1) Based on the CVaR risk measurement method, the risk avoidance control method proposed by the present invention can avoid the risk loss of load shedding caused by insufficient capacity configuration by coordinating the capacity configuration of each device and the operation strategy at the device capacity configuration level.

[0135] 2) Through the case study, it can be concluded that as the confidence level increases, the capacity configuration of the CHP and the photovoltaic system increases, while the capacity configuration of the GB decreases, indicating that increasing the configuration capacity of the CHP and the photovoltaic helps to avoid the risk of system load shedding and improve the energy supply reliability of the system.

[0136] For the models of each device in the embodiments of the present invention, except for those with special descriptions, the models of other devices are not limited, as long as the devices can perform the above functions.

[0137] Those skilled in the art can understand that the drawings are only schematic diagrams of a preferred embodiment, and the serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0138] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for reducing the load risk in a park based on the capacity configuration of PIES devices, characterized in that, The method includes: Establish an energy price scenario generation model. The specific modeling steps are to model the seasonal joint probability distribution of natural gas prices and electricity prices based on the Copula function and the DCC model, and then use Monte Carlo sampling to generate energy price scenarios; According to the established energy price scenario generation model, bring the generated energy price scenarios into the PIES device capacity configuration model to establish a CVaR-based PIES risk aversion control model; Use the Linprog function in MATLAB to call the simplex algorithm to solve the PIES risk aversion control model, and obtain the PIES device capacity configuration scheme at different confidence levels; The seasonal joint probability distribution modeling of natural gas prices and electricity prices based on the Copula function and the DCC model is as follows: Select the bivariate t-Copula function as the linking function: ； In the formula, represents the correlation coefficient between the natural gas price and the electricity price in season w; represents the degree of freedom; and represent the auxiliary variables for the double integral operation in season w; represents the inverse operation of the t-distribution with a degree of freedom of ; and are random variables; ； Where, ； ； ； In the formula, represents the covariance between the natural gas price and the electricity price in season w; and represent the autocovariances of the natural gas price and the electricity price in season w respectively; W is the total number of season cycles; and represent the prediction errors of the natural gas price and the electricity price in season w respectively; and represent the regression coefficients to be estimated respectively.

2. The method for reducing the campus load risk based on the capacity configuration of the PIES device according to claim 1, wherein The CVaR-based PIES risk aversion control model is: ； In the formula, represents the risk value of the total cost at the confidence level ; represents the total cost in scenario s; represents the probability of the occurrence of scenario s. S is the set of joint price scenarios.

3. A method for reducing the campus load risk based on the capacity configuration of PIES devices according to claim 1, characterized in that, The operating constraint conditions of the CVaR-based PIES risk aversion control model are: the upper and lower limit constraints of the output power of device i and the power balance constraint of the jth type of energy: ； ； Wherein, represents the j-th type of energy input into the i-th device at the t-th moment of the w-th season under the s scenario; represents the corresponding load demand of the j-th type of energy at the t-th moment of the w-th season, represents the output power of the i-th device at the t-th moment of the w-th season under the s scenario, represents the configured capacity of the i-th device; represents the energy conversion efficiency of the i-th device.

Images