Micro-grid optimization scheduling method considering system operation risk and user satisfaction

Abstract

The invention discloses a micro-grid optimization scheduling method considering system operation risk and user satisfaction, which comprises the steps of generating a power prediction curve of wind power and photovoltaic new energy output in a micro-grid, obtaining a prediction error range by adopting a Monte Carlo simulation method, and generating a typical scene set, obtaining a wind and light prediction error probability density distribution function through a data fitting method; micro-grid operation risks increased by high or low output caused by wind power and photovoltaic power are quantified by adopting the value-at-risk and conditional value-at-risk theories; obtaining the scheduling cost of the adjustable load in the micro-grid system; taking the minimum sum of the conditional value-at-risk cost and the operating cost of the micro-grid system as an objective function, taking the user satisfaction level under demand side response as a constraint, and establishing a micro-grid optimal scheduling model considering the operating risk and the user satisfaction; solving the model by adopting a genetic algorithm, and optimizing the optimal scheduling result of the grid-connected micro-grid system; and the safety, economy, stability and reliability of the micro-grid are enhanced.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power grid dispatching, and particularly relates to a micro-grid optimal dispatching method considering system operation risk and user satisfaction. BACKGROUND

[0002] Micro-grid has become an important part of the development of intelligent power in China due to its environmental protection, flexibility, high efficiency and high proportion of renewable energy access. However, the output of wind and solar distributed energy has randomness, volatility and uncertainty, which brings great risk to the safe, stable, economic and reliable operation of the system. How to effectively characterize and quantify the operation risk of wind and solar uncertainty to the micro-grid needs to be studied. At the same time, large-scale renewable energy access leads to difficulty in peak shaving of micro-grid system, and flexible adjustment resources on the power side and grid side are also relatively scarce, which will inevitably bring great risk to the optimal dispatching of micro-grid system. Fully exploring the flexible adjustment capacity of the load side of micro-grid can effectively reduce the system operation risk and improve the risk control ability.

[0003] At present, the main method for solving wind and solar uncertainty in the economic optimal dispatching research of micro-grid is stochastic optimization and robust optimization, but the risk cost is not included in the economic model of micro-grid operation. The uncertainty risk of renewable energy output will have a certain impact on the operation of micro-grid, so the risk cost should be included in the economic model. Although flexible load can effectively improve the flexibility of the system by accepting micro-grid regulation, it will also bring certain dispatching cost and change the user's own power consumption habits, affecting the user's power experience. The existing research does not comprehensively consider the influence of demand side response on the economy, flexibility and user satisfaction of micro-grid operation. SUMMARY

[0004] The technical problem to be solved by the present application is to provide a micro-grid optimal dispatching method considering system operation risk and user satisfaction, so as to solve the problem that although flexible load can effectively improve the flexibility of the system by accepting micro-grid regulation, it will also bring certain dispatching cost and change the user's own power consumption habits, affecting the user's power experience, and the existing research does not comprehensively consider the influence of demand side response on the economy, flexibility and user satisfaction of micro-grid operation.

[0005] The technical scheme of the present application is:

[0006] A micro-grid optimal dispatching method considering system operation risk and user satisfaction, comprising:

[0007] Step 1, generating the power prediction curve of wind power and photovoltaic new energy output in the micro-grid, adopting the Monte Carlo simulation method to obtain the prediction error range to generate a typical scenario set, and obtaining the wind and light prediction error probability density distribution function through data fitting method;

[0008] Step 2, the risk value and conditional risk value theory is used to quantify the increased micro-grid operation risk caused by the high or low output of wind power and photovoltaic;

[0009] Step 3, the demand side response model of transferable load and interruptible load is established based on time-of-use price and incentive mechanism, and the scheduling cost of adjustable load in the micro-grid system is obtained;

[0010] Step 4, the sum of the conditional risk value cost and the micro-grid system operation cost is minimized as the objective function, and the user satisfaction level under the demand side response is used as the constraint to establish the micro-grid optimization scheduling model considering operation risk and user satisfaction;

[0011] Step 5, the genetic algorithm is used to solve the model, and the optimal result of the grid-connected micro-grid system containing gas turbine, wind power, photovoltaic, energy storage and adjustable load is obtained.

[0012] The method for obtaining the probability density distribution function of wind and light prediction error comprises:

[0013] Step 1.1, in the actual operation process, the change relationship between the power generation of wind turbine and the wind speed is as follows:

[0014]

[0015] In the formula, P w,t represents the output power of the wind turbine at time t; v t represents the wind speed at time t; v in represents the cut-in wind speed; v out represents the cut-out wind speed; v e represents the rated wind speed; P WN represents the rated output power of the wind turbine;

[0016] The change relationship between the power generation of photovoltaic turbine and the solar radiation intensity, ambient temperature is as follows:

[0017]

[0018] In the formula, P pv,t represents the output power of the photovoltaic turbine at time t; P stc is the output under standard conditions, I r,t is the actual solar radiation intensity at time t; I stc = 1 kW / m 2 , a T is the power temperature coefficient of the photovoltaic panel; T t is the temperature of the photovoltaic panel at time t, T stc = 25℃;

[0019] According to the predicted values of the day-ahead wind speed, solar radiation intensity and temperature meteorological conditions, the wind power and photovoltaic average power prediction curves are generated, denoted as X=(X w , pv ), the prediction period is T, the sampling interval is Δt, X w and X pv are numerical vectors of 1×(T / Δt);

[0020] Step 1.2, the wind power and photovoltaic new energy power generation has randomness, volatility and uncertainty, the wind and light prediction error uncertainty is described by the following formula:

[0021]

[0022] In the formula, is the predicted output of wind and light, P k,t is the actual output of wind and light; ΔP k,t is the prediction error of wind and light, and are the upper and lower limit values of the wind and light output prediction error; rut and rlt respectively represent the 0-1 variables of the wind and light output being too high or too low;

[0023] The Monte Carlo simulation method is used to randomly obtain the wind and light prediction error to generate m typical scene sets, denoted as Y=(Y w , Y pv ), Y w and Y pv are numerical vectors of m×(T / Δt);

[0024] The probability density distribution function of wind and light prediction error is obtained by mathematical modeling method and data fitting technology, and the power prediction error of wind power is fitted by skew distribution; the power prediction error of photovoltaic output is fitted by standard normal distribution;

[0025]

[0026]

[0027] In the formula, x is the power prediction error of wind power, μ w , σ w and λ w are the location parameter, scale parameter and skewness parameter of the skew distribution respectively; y is the power prediction error of photovoltaic output, μ pv and σ pv are the mean and variance of the normal distribution respectively.

[0028] The risk value and conditional risk value theory are used to quantify the increased microgrid operation risk caused by the wind power and photovoltaic output being too high or too low, including:

[0029] Step 2.1, the form of penalty is used to quantify the increased risk cost caused by the high or low output of wind and light, and the risk measure is realized by risk value VaR and conditional risk value CvaR, VaR is the maximum risk loss expected at a certain confidence level, that is, the risk loss is considered to not exceed the VaR value at the confidence level probability;

[0030] For the loss function f(x, y), x and y are respectively the optimization decision variable and the random variable that determines the system risk loss, the probability density function of y is p(y), and it is assumed that a is the boundary value of f(x, y), and the distribution function of the loss function f(x, y) not greater than the boundary value a is:

[0031]

[0032] The VaR value at a given confidence level β ∈ (0, 1) is expressed as: β (x) represents:

[0033] VaR = a β (x) = min{a ∈ R; ψ(x, a) ≥ β};

[0034] CVaR measures the average loss exceeding VaR, which includes quantile and tail risk, and reflects the portfolio risk; the definition of CVaR is the conditional mean of the loss exceeding VaR at the same confidence level β, which is expressed as φ β (x) represents, and the basic calculation method is as follows:

[0035]

[0036] For convenience of solution, the CVaR value is represented by the function F β (x, a), that is:

[0037]

[0038] In the formula, a is the VaR value at the confidence level β; [f(x, y)-a] + represents max[0, f(x, y)-a];

[0039] The above formula is estimated and discretized by using the historical data of the random variable y:

[0040]

[0041] In the formula, is the estimated value of CVaR, y k is the kth group of sample data of y, and there are m groups;

[0042] Step 2.2. Quantify the increased microgrid operation risk caused by overestimation or underestimation of wind and PV output using CVaR theory. If the actual output of wind and PV is overestimated, it is called underestimation risk, which may lead to curtailment of wind and PV output due to oversupply;

[0043]

[0044]

[0045]

[0046]

[0047]

[0048] wherein, and respectively represent the risk cost of wind and PV curtailment at time t, and α W and α PV are the critical values of wind curtailment risk and PV curtailment risk loss, respectively, represents the new energy curtailment conditional risk cost considering the uncertainty of wind and PV prediction error; and respectively represent the wind curtailment and PV curtailment penalty fees, and respectively represent the wind curtailment and PV curtailment penalty coefficients, and respectively represent the wind curtailment power and curtailment power at time t under scenario m;

[0049] If the actual output of wind and PV is underestimated, it is called overestimation risk, which may lead to load shedding due to undersupply:

[0050]

[0051]

[0052] wherein, represents the risk cost of microgrid system load shedding, and α LD is the critical value of load shedding risk loss; is the load shedding penalty fee, is the load shedding penalty coefficient, represents the load shedding power at time t under scenario m;

[0053] The total operation risk cost CVaR of the microgrid system can be represented as

[0054]

[0055] Ut +S t ≤1

[0056] In the formula, U t and S t respectively represent 0-1 state variables indicating whether the micro-grid system is in the risk of power curtailment or load shedding at time t.

[0057] The demand side response model of transferable load and interruptible load is established based on time-of-use electricity price and incentive mechanism, and includes:

[0058] Step 3.1, the load in the micro-grid system mainly includes two categories of rigid load and flexible load, the rigid load cannot accept the regulation of the micro-grid, and is denoted as P basi The flexible load is divided into two categories of transferable load P tran and interruptible load P disr , and the total load P load in the micro-grid system is represented as:

[0059] P load =P basi +P tran +P disr ;

[0060] Establishing a transferable load demand response model based on time-of-use electricity price

[0061]

[0062]

[0063] In the formula, P tran,i and P ij respectively represent the power before and after the transferable load demand response, T tran represents a transfer coefficient, λ tran represents a scheduling cost coefficient of load transfer, and C disr,i represents a scheduling cost of transferable load demand response.

[0064] Establishing an interruptible load demand response model based on incentive policy

[0065]

[0066]

[0067] In the formula, P disr,i and P disr respectively represent the power before and after the interruptible load demand response, ΔP disr represents the actual interruptible power of the interruptible load at time i, λ DR represents a scheduling cost coefficient of load interruption, and C DR represents a scheduling cost of interruptible load demand response.The scheduling cost representing interruptible load demand response;

[0068] Total load in the micro-grid system under demand side response and total scheduling cost C DR Can be respectively expressed as:

[0069]

[0070] C DR = C tran + C disr .

[0071] The micro-grid optimization scheduling model considering operation risk and user satisfaction is established, including:

[0072] Step 4.1, taking the sum of the conditional risk value cost and the micro-grid system operation cost as the minimum as the objective function

[0073] ; min f = C OPE + CVaR + C GRI DR

[0074]

[0075]

[0076] In the formula, C OPE is the fuel cost of gas turbine operation, P gas,t is the power generation of gas turbine, η gas is the power generation efficiency, C gas is the market price of unit cubic meter of natural gas, H is the heat value of unit cubic meter of natural gas, unit is kWh / m 3 ; C GRI is the cost of power interaction of the micro-grid from the main grid, λ buy,t and λ sell,t respectively represent the purchase power price and the sale power price, P buy,t and P sell,t respectively represent the purchase power and the sale power; CVaR represents the total operation risk cost of the micro-grid system considering the wind and light uncertainty; C DR represents the total scheduling cost of the adjustable load in the micro-grid system under demand side response.

[0077] The constraint conditions include system power balance constraint, unit output constraint, micro-grid and main grid power interaction constraint, demand response constraint and user satisfaction constraint;

[0078] Power balance constraint

[0079] ​

[0080] where P g is the output power of the gas turbine, and are the charge and discharge power of the electrochemical energy storage, respectively;

[0081] unit output power constraint

[0082] gas turbine output power constraint

[0083]

[0084]

[0085] renewable energy output power constraint

[0086]

[0087]

[0088] energy storage system output power constraint

[0089]

[0090]

[0091]

[0092] E min ≤ E ESS,t ≤ E max

[0093] E ESS,0 = E ESS,T

[0094] where is the maximum output power of the gas turbine; ΔP g,t is the variation of the output power of the gas turbine, are the upper and lower limits of the ramp rate of the gas turbine, respectively; are the maximum output power of the wind and photovoltaic power at time period t, respectively; are the maximum charge and discharge power limits of the energy storage system, respectively; are 0-1 variables indicating whether the energy storage system is charging or discharging, respectively; E ESS,t is the amount of electricity stored in the energy storage system at time t, σ is the self-discharge rate of the battery device, η ch and η dis are the charge and discharge efficiencies, respectively; E min and E max are the lower and upper limits of the amount of electricity allowed to be stored in the energy storage system, respectively; E ESS,0 denotes the initial amount of electricity stored in the energy storage system, E ESS,Trepresents the storage power at the last moment of the storage period;

[0095] Microgrid and main grid power interaction constraints

[0096]

[0097]

[0098] wherein, respectively represent the upper limit of the IES power purchase and power sale from the grid; respectively represent the 0-1 variable of whether the microgrid system is purchasing power or selling power;

[0099] Demand response constraints

[0100] 0≤-T ii ≤η max

[0101]

[0102]

[0103]

[0104] wherein, T ii and numerically represent the actual transfer ratio and the maximum transfer ratio of the transferable load at i moment, and the total load amount needs to be kept unchanged during the transfer process; represents the maximum interruption power of the interruptible load at i moment;

[0105] User satisfaction constraints

[0106]

[0107] wherein, the user satisfaction level ζ is introduced to measure the user power experience, θ cu numerically measures the worst experience acceptable to the user;

[0108] By analyzing the change of the microgrid operation risk cost CvaR under different user satisfaction levels, the effectiveness of the demand side response in reducing the power system operation risk is evaluated.

[0109] The method for evaluating the effectiveness of the demand side response in reducing the power system operation risk comprises the following steps:

[0110] Step 5.1, population initialization; a population with n individuals is generated by using a random simulation method, the output prediction curves of the gas turbine, wind power and photovoltaic are input, the performance parameters of the wind power, photovoltaic, energy storage and tie line equipment in the micro-grid system are input, the category, adjustable proportion and time-of-use electricity price of the load are input, and the boundary range of the population individuals is determined;

[0111] Step 5.2, population evaluation; the objective function is taken as the fitness function, the fitness of n individuals is calculated, and each individual is evaluated according to the size of the fitness to judge the quality;

[0112] Step 5.3, population selection; according to the fitness of the individuals in the population, the individuals with high fitness are selected from the current population by roulette and the like; wherein the roulette is to determine the number of individuals in the next generation population in proportion to the fitness;

[0113] Step 5.4, population crossover; new individuals are generated according to a certain crossover probability and a crossover method;

[0114] Step 5.5, population mutation; new individuals are generated according to a certain mutation probability and a mutation method;

[0115] Step 5.6, produce a new generation of population; a new generation of population is generated according to the population selection, crossover and mutation operation, whether the end condition of iteration number or accuracy setting is met is judged, if met, the best individual and the optimal solution are output and the process is ended, otherwise, steps 5.2-5.6 are repeated.

[0116] The present application has the following advantages:

[0117] The present application is aimed at the risk of load shedding and power curtailment caused by the prediction error uncertainty of renewable energy in the operation of the micro-grid system, the risk cost is quantitatively characterized by using the risk value theory VaR and the conditional risk value theory CVaR, and the risk cost is included in the cost model of the micro-grid optimization scheduling; improving the flexibility of the system is an effective way to cope with the prediction uncertainty risk of new energy, considering the demand side response can effectively reduce the operation risk of the micro-grid system, the adjustable load scheduling cost is included in the economic model of the micro-grid, and the user's power experience is also considered, a micro-grid optimization scheduling method considering the system operation risk and user satisfaction is proposed, and the characterization and control of the operation risk of the micro-grid system are realized.

[0118] The present application has the following advantages:

[0119] The present application considers the prediction error uncertainty of wind, light and other renewable energy, quantitatively characterizes the risk size of the uncertainty factors of the renewable energy output by using the CVaR theory, thereby a risk cost function is established, and the idealized situation of the economic optimization scheduling result is avoided.

[0120] The application fully taps the flexibility adjustment capability of the load side, introduces the adjustable load scheduling cost into the optimization objective function, introduces the user satisfaction into the constraint condition, and simultaneously considers the influence of the demand side response on the operation economy, flexibility and user power consumption experience of the micro-grid.

[0121] The application guides users to participate in the demand side response based on the time-of-use electricity price and incentive mechanism to reduce the power consumption cost and risk cost, effectively balances the operation cost and operation risk of the micro-grid, improves the risk control capability of the micro-grid, provides a more flexible optimization scheduling scheme for the micro-grid, and enhances the safety, economy, stability and reliability of the micro-grid. BRIEF DESCRIPTION OF DRAWINGS

[0122] Figure 1 It is a flowchart of the application;

[0123] Figure 2 It is a flowchart of the micro-grid optimization scheduling model solving process of the application. DETAILED DESCRIPTION

[0124] 1. A micro-grid optimization scheduling method considering operation risk and user satisfaction, comprising the following steps:

[0125] Step 1. Generating the power prediction curve of the wind power and photovoltaic new energy output in the micro-grid, adopting the Monte Carlo simulation method to obtain the prediction error range to generate a typical scenario set, and obtaining the wind and light prediction error probability density distribution function through the data fitting method.

[0126] Step 2. Quantifying the increased micro-grid operation risk caused by the high or low output of the wind power and photovoltaic by the risk value and conditional risk value theory.

[0127] Step 3. Establishing the demand side response model of the transferable load and interruptible load based on the time-of-use electricity price and incentive mechanism to obtain the scheduling cost of the adjustable load in the micro-grid system.

[0128] Step 4. Taking the sum of the conditional risk value cost and the micro-grid system operation cost as the objective function, and taking the user satisfaction level under the demand side response as the constraint to establish the micro-grid optimization scheduling model considering the operation risk and user satisfaction.

[0129] Step 5. Solving the model by adopting the genetic algorithm to obtain the optimal result of the grid-connected micro-grid system optimization scheduling containing the gas turbine, wind power, photovoltaic, energy storage and adjustable load.

[0130] In step 1, day-ahead wind power and photovoltaic power prediction curves are generated according to meteorological conditions such as wind speed, solar radiation intensity, etc., a typical scenario set is obtained by using a Monte Carlo simulation method to obtain a prediction error range, and a wind-solar prediction error probability density distribution function is obtained by using a data fitting method, including the following steps.

[0131] Step 1.1 In actual operation, the relationship between the power generation of a wind turbine generator and the change of wind speed is shown in the following formula:

[0132]

[0133] In the formula, P w,t represents the output power of the wind turbine generator at time t; v t represents the wind speed at time t; v in represents the cut-in wind speed; v out represents the cut-out wind speed; v e represents the rated wind speed; P WN represents the rated output power of the wind turbine generator.

[0134] The relationship between the power generation of a photovoltaic generator and the change of solar radiation intensity and ambient temperature is shown in the following formula:

[0135]

[0136] In the formula, P pv,t represents the output power of the photovoltaic generator at time t; P stc is the output under standard conditions, I r,t is the actual solar radiation intensity at time t; I stc = 1 kW / m 2 , a T is the power temperature coefficient of the photovoltaic panel; T t is the temperature of the photovoltaic panel at time t, and T stc = 25℃.

[0137] According to the predicted values of day-ahead wind speed, solar radiation intensity, temperature and other meteorological conditions, wind power and photovoltaic average power prediction curves can be generated, denoted as X = (X w , X pv ), the prediction period is T, the sampling interval is Δt, X w and X pv are numerical vectors of 1 × (T / Δt).

[0138] Step 1.2 Wind power and photovoltaic new energy power generation have randomness, volatility and uncertainty, and the wind-solar prediction error uncertainty is described by the following formula:

[0139]

[0140] In the formula, P is the predicted output of wind and solar power k,t P is the actual output of wind and solar power; ΔP k,t is the prediction error of wind and solar power, and r is the upper and lower limit of wind and solar power output prediction error; ut and r lt respectively represent the 0-1 variable of wind and solar power output being too high or too low.

[0141] The Monte Carlo simulation method is used to randomly obtain wind and solar power prediction error to generate m typical scenario sets, denoted as Y = (Y w ,Y pv ), Y w and Y pv are numerical vectors of m x (T / Δt).

[0142] The mathematical modeling method and data fitting technology are used to fit the probability density distribution function of wind and solar power prediction error. The power prediction of wind turbine output presents nonlinearity, and the distribution presents multi-peak, skewness and heavy tail. Skewness distribution is used to fit the power prediction error of wind turbine output; standard normal distribution is used to fit the power prediction error of photovoltaic output.

[0143]

[0144]

[0145] In the formula, x is the power prediction error of wind turbine output, μ w , σ w and λ w are the location parameter, scale parameter and skewness parameter of the skewness distribution respectively; y is the power prediction error of photovoltaic output, μ pv and σ pv are the mean and variance of the normal distribution respectively.

[0146] In step 2, the risk value and conditional risk value theory are used to quantify the increased microgrid operation risk caused by the high or low output of wind and solar power, which includes the following steps:

[0147] Step 2.1 In the microgrid system, the uncertainty of renewable energy output power will cause fluctuations, which will cause certain risk to the stable operation of the power grid. Therefore, the form of punishment is used to quantify the risk cost increased by the high or low output of wind and solar power. Risk value VaR and conditional risk value CVaR are commonly used risk measurement methods. VaR is the maximum risk loss expected at a certain confidence level, that is, the risk loss is considered to exceed the VaR value with the probability of the confidence level. The basic calculation method is as follows:

[0148] For a loss function f(x, y), x and y are the optimization decision variable and the random variable that determines the system risk loss, respectively, and the probability density function of y is ρ(y). Assuming that α is the boundary value of f(x, y), the distribution function of the loss function f(x, y) not greater than the boundary value α is:

[0149]

[0150] The VaR value under a given confidence level β ∈ (0, 1) is denoted as α β (x) :

[0151] VaR = α β (x) = min{α ∈ R; ψ(x, α) ≥ β}

[0152] Compared with VaR, which only considers the risk information at the quantile point and ignores the risk information in the tail of the quantile point, CVaR can better reflect the portfolio risk by measuring the average loss exceeding VaR and including the quantile point and the tail risk. The definition of CVaR is the conditional mean of the loss exceeding VaR under the same confidence level β, denoted as φ β (x), and the basic calculation method is as follows:

[0153]

[0154] In order to facilitate the solution, another relatively simple function F β (x, α) is often used to represent the CVaR value, that is:

[0155]

[0156] In the formula, α is the VaR value under the confidence level β; [f(x, y) - α] + represents max[0, f(x, y) - α].

[0157] In general, the probability density function ρ(y) is difficult to obtain, and the above formula can be estimated and discretized by using the historical data of the random variable y:

[0158]

[0159] In the formula, is the estimated value of CVaR, y k is the kth group of sample data of y, and there are m groups.

[0160] Step 2.2 uses the CVaR theory to quantify the increased microgrid operation risk caused by the high or low output of wind power and photovoltaic power. If the actual output of wind power and photovoltaic power is high, it is called underestimation risk, which may lead to more supply than demand and result in wind curtailment and light curtailment:

[0161]

[0162]

[0163]

[0164]

[0165]

[0166] wherein, and denote the risk cost of wind and PV curtailment at time t, respectively, and W and PV are the threshold values of wind and PV curtailment risk loss, respectively, denotes the risk cost of new energy curtailment considering the uncertainty of wind and PV forecast error; and denote the penalty cost of wind and PV curtailment, respectively, and denote the penalty coefficient of wind and PV curtailment, respectively, and denote the wind curtailment power and curtailment power at time t under scenario m, respectively.

[0167] If the actual output of wind and PV is low, it is called overestimation risk, which may lead to load shedding:

[0168]

[0169]

[0170] wherein, denotes the risk cost of microgrid system load shedding, and LD is the threshold value of load shedding risk loss; is the penalty cost of load shedding, is the penalty coefficient of load shedding, denotes the load shedding power at time t under scenario m.

[0171] The total operation risk cost CVaR of the microgrid system can be expressed as

[0172]

[0173] U t +S t ≤1

[0174] wherein, U t and S tThese are 0-1 state variables representing whether the microgrid system is at risk of power curtailment or load shedding at time t.

[0175] Step 3 involves establishing demand-side response models for transferable and interruptible loads based on time-of-use pricing and incentive mechanisms. This includes the following steps:

[0176] Step 3.1 The loads in a microgrid system mainly include two categories: rigid loads and flexible loads. Rigid loads cannot be controlled by the microgrid and are denoted as P. basi Flexible loads are divided into transferable loads P tran Interruptible load P disr Two main categories: the total load P within the microgrid system. load It can be represented as:

[0177] P load =P basi +P tran +P disr

[0178] Establish a demand response model for transferable load based on time-of-use pricing.

[0179]

[0180]

[0181] In the formula, P tran,i and T represents the power before and after the demand response of transferable load, respectively. ij λ represents the transfer coefficient. tran C represents the scheduling cost coefficient for load transfer. tran This represents the scheduling cost of responding to transferable load demand.

[0182] Establish a demand response model for interruptible loads based on incentive policies.

[0183]

[0184]

[0185] In the formula, P disr,i and These represent the power before and after interruptible load demand response, ΔP and ΔP, respectively. disr,i λ represents the actual interruptible power of the load at time i; disr C represents the scheduling cost coefficient for load interruption. disr This represents the scheduling cost of responding to interruptible load demand.

[0186] Total load in a microgrid system under demand-side response And total scheduling cost CDR may be expressed as:

[0187]

[0188] C DR = C tran + C disr

[0189] In step 4, the sum of the conditional value at risk cost and the micro-grid system operation cost is minimized as the objective function, and the user satisfaction level under demand side response is taken as the constraint to establish a micro-grid optimization scheduling model considering operation risk and user satisfaction, which includes the following steps:

[0190] Step 4.1 takes the sum of the conditional value at risk cost and the micro-grid system operation cost as the objective function

[0191] min f = C OPE + C GRI + CVaR + C DR

[0192]

[0193]

[0194] In the formula, C OPE is the fuel cost of gas turbine operation, P gas,t is the power generation of gas turbine, η gas is the power generation efficiency, C gas is the market price of unit cubic meter of natural gas, H is the heat value of unit cubic meter of natural gas, unit is kWh / m 3 ; C GRI is the cost of power interaction of micro-grid from the main grid, λ buy,t and λ sell,t respectively represent the purchase power price and the sale power price, P buy,t and P sell,t respectively represent the purchase power and the sale power; CVaR represents the total operation risk cost of micro-grid system considering wind and light uncertainty; C DR represents the total scheduling cost of adjustable load in micro-grid system under demand side response.

[0195] Step 4.2 constraints mainly include system power balance constraint, unit output constraint, micro-grid and main grid power interaction constraint, demand response constraint, user satisfaction constraint, etc.

[0196] ①Power balance constraint

[0197]

[0198] where P g is the output power of the gas turbine, and are the charge and discharge power of the electrochemical energy storage, respectively.

[0199] ② Unit output constraint

[0200] Gas turbine output constraint

[0201]

[0202]

[0203] Renewable energy output constraint

[0204]

[0205]

[0206] Energy storage system output constraint

[0207]

[0208]

[0209]

[0210] E min ≤ E ESS,t ≤ E max

[0211] E ESS,0 = E ESS,T

[0212] where is the maximum output power of the gas turbine; ΔP g,t is the change in the output power of the gas turbine, are the upper and lower limits of the gas turbine ramp rate, respectively; are the maximum values of the wind power and photovoltaic power in time period t, respectively; are the maximum charge power limit and the maximum discharge power limit of the energy storage system, respectively; are 0-1 variables indicating whether the energy storage system is charging or discharging, respectively; E ESS,t is the amount of electricity stored in the energy storage system at time t, and σ is the self-discharge rate of the battery device, and η ch and η dis are the charge and discharge efficiencies, respectively; E min and E max are the lower and upper limits of the amount of electricity allowed to be stored in the energy storage system, respectively; EESS,0 E0represents the storage power at the initial time of energy storage ESS,T E1represents the storage power at the final time of energy storage.

[0213] ③Microgrid and main grid power interaction constraints

[0214]

[0215]

[0216] wherein, respectively, are the upper limits of the IES power purchase and power sale from the grid; respectively, are 0-1 variables indicating whether the microgrid system is purchasing power or selling power.

[0217] ④Demand response constraints

[0218] 0≤-T ii ≤η max

[0219]

[0220]

[0221]

[0222] wherein, T ii and numerically represent the actual transfer ratio and the maximum transfer ratio of the transferable load at time i, and the total load amount needs to be kept unchanged during the transfer process; represents the maximum interruption power of the interruptible load at time i.

[0223] ⑤User satisfaction constraints

[0224]

[0225] wherein, the user satisfaction level ζ is introduced to measure the user power experience, θ cu numerically measures the worst experience acceptable to the user.

[0226] In step 5, the genetic algorithm is used to solve the model to obtain the optimal scheduling results of the grid-connected microgrid system containing a gas turbine, wind power, photovoltaic, energy storage and adjustable load, and the changes of the microgrid operation risk cost CvaR under different user satisfaction levels are analyzed to evaluate the effectiveness of demand side response in reducing the operation risk of the power system, including the following steps:

[0227] Step 1 population initialization. A population with n individuals is generated by using a random simulation method, the output prediction curves of the gas turbine, wind power and photovoltaic are input, the performance parameters of the wind power, photovoltaic, energy storage and tie line in the micro-grid system are input, the category of the load, adjustable proportion and time-of-use electricity price are input, and the boundary range of the population individuals is determined.

[0228] Step 2 population evaluation. The objective function is taken as the fitness function, the fitness of n individuals is calculated, and each individual is evaluated according to the size of the fitness to judge its quality.

[0229] Step 3 population selection. According to the fitness of the individuals in the population, the individuals with high fitness are selected from the current population by roulette and the like. The roulette is used to determine the number of individuals in the next generation population in proportion to the fitness.

[0230] Step 4 population crossover. According to certain crossover probability and crossover method, new individuals are generated.

[0231] Step 5 population mutation. According to certain mutation probability and mutation method, new individuals are generated.

[0232] Step 6 new generation population. The new generation population is generated according to the population selection, crossover and mutation operations. Whether the end condition of iteration number or accuracy setting is met is judged, if met, the best individual and the optimal solution are output and the process is ended, otherwise, steps 2-6 are repeated.

[0233] Characteristics of the application:

[0234] 1. The application establishes a micro-grid risk assessment model considering the uncertainty of wind and light prediction error based on the risk value VaR and the conditional risk value CVaR, and the micro-grid operation risk cost mainly includes two types of underestimation risk and overestimation risk: ① if the actual output of wind power and photovoltaic is high, it is called underestimation risk, and the underestimation risk may lead to surplus and wind and light abandonment; ② if the actual output of wind power and photovoltaic is low, it is called overestimation risk, and the overestimation risk may lead to insufficient supply and load shedding; the risk is quantitatively characterized.

[0235] 2. The application considers the uncertainty of wind and light output on one hand, and the operation risk cost of the micro-grid system is included in the optimization objective function, on the other hand, the demand side response is introduced to improve the system flexibility, the adjustable load scheduling cost is included in the optimization objective function, and the user satisfaction is introduced into the constraint condition; the operation cost and operation risk of the micro-grid system are fully considered, and the risk is flexibly controlled.

Claims

1. A microgrid optimal scheduling method considering system operation risk and user satisfaction, characterized in that: The method includes: Step 1: Generate power prediction curves for wind power and photovoltaic new energy output within the microgrid, use Monte Carlo simulation method to obtain prediction error range and generate typical scenario set, and obtain wind and solar prediction error probability density distribution function through data fitting method. Step 2: Use the Value at Risk (VaR) and Conditional Value at Risk (CVR) theories to quantify the increased microgrid operation risk caused by excessively high or low power output from wind and solar power. Step 3: Establish demand-side response models for transferable and interruptible loads based on time-of-use pricing and incentive mechanisms to obtain the scheduling cost of adjustable loads within the microgrid system; Step 4: Minimize the sum of conditional risk value cost and microgrid system operating cost as the objective function, and use the user satisfaction level under demand-side response as a constraint to establish a microgrid optimal scheduling model that considers operating risk and user satisfaction. Step 5: Use a genetic algorithm to solve the model and obtain the optimal scheduling results for the grid-connected microgrid system containing gas turbines, wind power, photovoltaics, energy storage, and adjustable loads.

2. The microgrid optimal scheduling method considering system operation risk and user satisfaction according to claim 1, characterized in that: Methods for obtaining the probability density distribution function of wind and solar prediction errors include: Step 1.1: During actual operation, the relationship between the power generation of the wind turbine and the wind speed is shown in the following formula: In the formula, P w,t v represents the output power of the wind turbine generator at time t; t The wind speed at time t is represented by v. in Indicates the cut-in wind speed; v out Indicates the cut-out wind speed; v e Indicates the rated wind speed; P WN This indicates the rated output power of the wind turbine generator set; The relationship between the power generation of a photovoltaic unit and the changes in solar radiation intensity and ambient temperature is shown in the following formula: In the formula, P pv,t P represents the output power of the photovoltaic generator set at time t; stc For the output under standard conditions, I r,t Let I be the actual solar radiation intensity at time t; stc =1kW / m 2 a T T represents the power temperature coefficient of the photovoltaic panel. t Let T be the temperature of the photovoltaic panel at time t. stc =25℃; Based on the predicted values ​​of wind speed, solar radiation intensity, and temperature, average power forecast curves for wind power and photovoltaic power are generated, denoted as X = (X... w ,X pv The prediction period is T, the sampling interval is Δt, and X w and X pv These are numerical vectors of 1×(T / Δt); Step 1.2: The output of wind power and photovoltaic new energy power generation has randomness, volatility, and uncertainty. The uncertainty of wind and solar forecasting error is described by the following formula: In the formula, P contributes to the forecasting of scenery k,t The actual contribution to the scenery; ΔP k,t For the prediction error of scenery, and These are the upper and lower limits of the prediction error for wind and solar power output; r ut and r lt These represent 0-1 variables indicating whether the wind and solar power output is too high or too low. The Monte Carlo simulation method is used to randomly obtain the wind and light prediction error and generate m typical scene sets, denoted as Y = (Y w ,Y pv ), Y w and Y pv These are numerical vectors of size m × (T / Δt); The probability density distribution function of wind and solar power prediction errors is adopted using mathematical modeling methods and data fitting techniques. A skewed distribution is used to fit the power prediction error of wind power output, and a standard normal distribution is used to fit the power prediction error of photovoltaic output. In the formula, x represents the power prediction error of wind power output, and μ w σ w and λ w These represent the location, scale, and skewness parameters of the skewed distribution, respectively; y represents the power prediction error of the photovoltaic output, and μ represents the skewness parameter. pv and σ pv These are the mean and variance of the normal distribution, respectively.

3. The microgrid optimal scheduling method considering system operation risk and user satisfaction according to claim 1, characterized in that: The increased microgrid operational risks caused by excessively high or low power output from wind and solar power, quantified using value-at-risk and conditional value-at-risk theories, include: Step 2.1: Use penalties to quantify the increased risk cost caused by excessively high or low power output due to wind and solar power. Risk measurement is achieved using Value at Risk (VaR) and Conditional Value at Risk (CvaR). VaR is the maximum expected risk loss at a certain confidence level, that is, the probability that the risk loss will not exceed the VaR value at this confidence level. For the loss function f(x,y), where x and y are the optimization decision variable and the random variable determining the system risk loss, respectively, and the probability density function of y is ρ(y), assuming α is the boundary value of f(x,y), the distribution function of the loss function f(x,y) not greater than the boundary value α is: The VaR value at a given confidence level β∈(0,1) is expressed as α β (x) represents: VaR=α β (x)=min{α∈R;ψ(x,α)≥β}; CVaR measures the average loss exceeding VaR, encompassing quantile and tail risk, and reflects portfolio risk. CVaR is defined as the loss exceeding the conditional mean of VaR at the same confidence level β, expressed as φ. β (x) represents the basic calculation method as follows: To facilitate the solution, the function F is used. β (x, α) represents the CVaR value, i.e.: In the formula, α is the VaR value at confidence level β; [f(x,y)-α] + It represents max[0,f(x,y)-α]; Estimate the above formula using historical data of the random variable y and then discretize it: In the formula, y is an estimate of CVaR. k Let y be the kth sample data group, with a total of m groups; Step 2.2: Use CVaR theory to quantify the increased microgrid operation risk caused by the higher or lower output of wind power and photovoltaic power. If the actual output of wind power and photovoltaic power is higher, it is called underestimation of risk. Underestimation of risk may lead to oversupply and curtailment of wind and solar power. In the formula, and Let α represent the risk costs of wind power and solar power curtailment at time t, respectively. W and α PV These are the critical values ​​for wind curtailment risk and solar curtailment risk losses, respectively. This indicates the risk and cost of renewable energy curtailment due to the uncertainty of wind and solar forecasting errors; and These represent the penalties for wind and solar power curtailment, respectively. and These represent the penalty coefficients for wind curtailment and solar curtailment, respectively. and Let them represent the wind curtailment power and the electricity curtailment power at time t under scenario m, respectively; If the actual output of wind and solar power is lower than expected, this is called an overestimation risk. Overestimation risk may lead to supply shortages and load shedding. In the formula, α represents the risk cost incurred by load shedding in a microgrid system. LD This is the critical value for the risk of loss due to load shedding; To incur penalties for load shedding, This is the load shedding penalty factor. This represents the load shedding power at time t under scenario m; The total operating risk cost (CVaR) of a microgrid system can be expressed as: IN t +S t ≤1 In the formula, U t and S t These are 0-1 state variables representing whether the microgrid system is at risk of power curtailment or load shedding at time t.

4. The microgrid optimal scheduling method considering system operation risk and user satisfaction according to claim 1, characterized in that: Demand-side response models for transferable and interruptible loads based on time-of-use pricing and incentive mechanisms include: Step 3.1: The loads within a microgrid system mainly consist of two categories: rigid loads and flexible loads. Rigid loads cannot be controlled by the microgrid and are denoted as P. basi Flexible loads are divided into transferable loads P tran and interruptible load P disr Two main categories: the total load P within the microgrid system. load Represented as: P load =P basi +P tran +P disr ; Establish a demand response model for transferable load based on time-of-use pricing. In the formula, P tran,i and T represents the power before and after the demand response of transferable load, respectively. ij λ represents the transfer coefficient. tran C represents the scheduling cost coefficient for load transfer. tran Represents the scheduling cost of responding to transferable load demand; Establish a demand response model for interruptible loads based on incentive policies. In the formula, P disr,i and ΔP represents the power before and after interruptible load demand response. disr,i λ represents the actual interruptible power of the load at time i; disr C represents the scheduling cost coefficient for load interruption. disr This represents the scheduling cost of responding to interruptible load demand. Total load in a microgrid system under demand-side response And total scheduling cost C DR They can be represented as: C DR =C tran +C disr 。 5. A microgrid optimal scheduling method considering system operation risk and user satisfaction according to claim 1, characterized in that: Establishing a microgrid optimal scheduling model that considers operational risks and user satisfaction includes: Step 4.1: Minimize the sum of conditional risk value cost and microgrid system operating cost as the objective function. min f=C OPE +C GRI +CVaR+C DR In the formula, C OPE P represents the fuel cost of operating a gas turbine. gas,t η represents the power generation capacity of the gas turbine. gas For power generation efficiency, C gas H represents the market price per cubic meter of natural gas, and H represents the calorific value per cubic meter of natural gas, expressed in kWh / m³. 3 C GRI λ represents the cost of power exchange between the microgrid and the main grid. buy,t and λ sell,t P represents the electricity purchase price and the electricity sales price, respectively. buy,t and P sell,t These represent the purchased power and the sold power, respectively; CVaR represents the total operating risk cost of the microgrid system considering the uncertainties of wind and solar power; C DR This represents the total dispatch cost of adjustable loads within a microgrid system under demand-side response.

6. A microgrid optimal scheduling method considering system operation risk and user satisfaction according to claim 5, characterized in that: The constraints include system power balance constraints, unit output constraints, microgrid and main grid power interaction constraints, demand response constraints, and user satisfaction constraints. Power balance constraints In the formula, P g For the output of the gas turbine, and These are the charging and discharging powers of electrochemical energy storage, respectively. Unit output constraints Gas turbine output constraints Renewable energy output constraints Energy storage system output constraints AND min ≤E ESS,t ≤E max AND ESS,0 =And ESS,T In the formula, ΔP is the maximum output power of the gas turbine. g,t This represents the change in the output power of the gas turbine. These are the upper and lower limits of the gas turbine ramp rate, respectively. These represent the maximum output values ​​of wind power and solar power during time period t, respectively. These are the maximum charging power limit and the maximum discharging power limit for energy storage systems, respectively. These represent 0-1 variables indicating whether the energy storage system is charging or discharging; E ESS,t Let η be the amount of electricity stored in the energy storage system at time t, σ be the self-discharge rate of the battery device, and η be the energy stored in the system at time t. ch and η dis These are the charge and discharge efficiencies, respectively; E min and E max These represent the lower and upper limits, respectively, for the allowable storage of electrical energy by the energy storage system. E ESS,0 E represents the amount of electricity stored at the initial moment of energy storage. ESS,T This indicates the amount of electricity stored at the end of the energy storage cycle; Microgrid and main grid power interaction constraints In the formula, These represent the upper limits of the IES's power purchase from the grid and power sold to the grid, respectively. These are 0-1 variables representing whether the microgrid system is purchasing or selling electricity; Demand response constraints 0≤-T ii ≤η max In the formula, T ii and Numerically, it represents the actual transfer ratio and the maximum transfer ratio of the load that can be transferred at time i. During the transfer process, it is necessary to ensure that the total load remains constant. This represents the maximum interruptible power of the load at time i; User satisfaction constraints In the formula, user satisfaction level ζ is introduced to measure the user's electricity experience, θ cu Numerical measure of the worst acceptable user experience when using electricity.

7. A microgrid optimal scheduling method considering system operation risk and user satisfaction according to claim 1, characterized in that: By analyzing the changes in the microgrid operation risk cost (CvaR) under different user satisfaction levels, the effectiveness of demand-side response in reducing power system operation risk is evaluated.

8. A microgrid optimal scheduling method considering system operation risk and user satisfaction according to claim 7, characterized in that: Methods for assessing the effectiveness of demand-side response in reducing power system operational risks include: Step 5.1 Population initialization: Generate a population of n individuals using a random simulation method. Input the output prediction curves of gas turbine, wind power, and photovoltaic, input the performance parameters of wind power, photovoltaic, energy storage, and tie-line equipment in the microgrid system, input the load type, adjustable ratio, and time-of-use electricity price, and determine the boundary range of the population individuals. Step 5.2, Population Evaluation: Using the objective function as the fitness function, calculate the fitness of n individuals, evaluate each individual based on its fitness value, and determine its quality. Step 5.3, Population Selection: Based on the fitness of individuals in the population, individuals with high fitness are selected from the current population through methods such as roulette wheel selection; roulette wheel selection determines the number of each individual that is passed on to the next generation with a probability proportional to fitness. Step 5.4, Population crossover: Generate new individuals according to a certain crossover probability and crossover method; Step 5.5, Population Mutation: Generate new individuals according to a certain mutation probability and mutation method; Step 5.6: Generate a new generation population; Generate a new generation population based on population selection, crossover, and mutation operations, and determine whether it meets the termination conditions set by the number of iterations or precision. If it does, output the best individual and its optimal solution and end the process; otherwise, repeat steps 5.2 to 5.6.

Images