CN114519449A - Operation optimization method for park energy system - Google Patents

Abstract

The invention relates to the technical field of new energy optimization, and provides a method for optimizing operation of a park energy system, which comprises the following steps: dividing a park energy system into a supply side and a user side; considering the output uncertainty of the photovoltaic unit and the wind turbine unit at the supply side, and establishing a hub energy flow model and a carbon emission flow model at the supply side; considering the randomness of electric vehicle access at a user side, and establishing an energy hub matrix model at the user side; calculating a total cost of the supply side, the total cost including an operating cost, a carbon emission cost and a wind and light abandonment cost, the objective function being that the total cost is minimal; and solving the objective function to obtain a plurality of scene comprehensive analysis solutions. According to the invention, the uncertainty of user energy behaviors and the random access of the electric automobile are considered at a user side, and the uncertainty of the output of the fan and the photovoltaic unit is considered at a supply side; the total cost is the lowest as an optimization target, and the obtained solution ensures the full utilization of renewable energy sources and reduces carbon emission; the operation optimization model has robustness.

Description

A method for optimizing the operation of the energy system in the park

technical field

The invention relates to the technical field of new energy optimization, in particular to a method for optimizing the operation of a park energy system.

Background technique

With the popularization of electric vehicles and the substantial increase in the installed capacity of renewable energy, the grid connection of electric vehicles and renewable energy has brought great challenges to the energy system.

There are many types of electric vehicles, and the existing research is limited to taking electric vehicles as an uncertain load on the consumer side, and does not consider the complex characteristics of actual electric vehicles.

The existing operation optimization methods do not take into account the access of electric vehicles, and the in-depth research on the utilization of renewable energy such as wind and light is not enough, which is not in line with the national strategic goal of energy conservation. In the face of enormous challenges, new energy optimization methods are urgently needed.

SUMMARY OF THE INVENTION

The purpose of the present invention is to overcome at least one of the deficiencies of the prior art, and to provide a method for optimizing the operation of the energy system in the park. And participate in power grid dispatching, when participating in dispatching, electric vehicles are regarded as consumer end, energy storage equipment and production capacity end. At the same time, this method fully considers the uncertainty of wind and solar renewable energy output, and takes the lowest total cost of integrated system operation cost, carbon emission cost, photovoltaic cost and user comfort cost as the optimization goal. The simulation results show that the operational optimization model proposed in this paper is robust and can improve the utilization of renewable energy and reduce carbon emissions.

The present invention adopts following technical scheme:

A method for optimizing the operation of a park energy system, comprising the following steps:

S1. Divide the energy system of the park into a supply side and a user side. The supply side includes gas boilers, CHP units, photovoltaic units and wind turbines. The user side includes energy storage equipment, electric heat pumps, absorption refrigeration units, and air conditioning equipment. and electric vehicles;

S2. On the supply side, considering the output uncertainty of photovoltaic units and wind turbines, establish a hub energy flow model and a carbon emission flow model on the supply side; on the user side, consider the randomness of electric vehicle access , establish the energy hub matrix model of the user side;

S3. According to the supply-side hub energy flow model and carbon emission flow model obtained in step S2, and the user-side energy hub matrix model, calculate the total cost of the supply side. The total cost includes operation cost, carbon emission cost and Abandoning the cost of wind and solar, the objective function of the park energy system operation optimization is to minimize the total cost;

S4. Solve the objective function in step S3 to obtain an optimal solution, and optimize the operation of the park energy system according to the optimal solution.

Any of the above-mentioned possible implementations further provides an implementation, and step S2 is specifically:

The energy hub model on the supply side described in S2.1 is as follows:

η_GB分别为变压器效率、CHP机组制电效率、CHP机组制热效率和燃气锅炉效率；

分别为CHP机 组燃气分配系数、燃气锅炉机组燃气分配系数，E^e、E^g、E^h分别为园区能源系统向上级网 络购买的电能、气量、热能，P_wt、P_pv分别为风电实际使用值和光电实际使用值；In the formula, ^Pe and ^Ph are the power and heat of the output node on the supply side, respectively; η _T ,

η _GB is the transformer efficiency, the CHP unit electricity production efficiency, the CHP unit heating efficiency and the gas boiler efficiency;

are the gas distribution coefficient of CHP units and gas boiler units, respectively, E ^e , E ^g , and E ^h are the electric energy, gas volume, and thermal energy purchased by the park energy system from the upper-level network, respectively, P _wt , P _pv are the actual use values of wind power, respectively and the actual use value of optoelectronics;

The energy hub matrix model of the S2.2 user side is as follows:

分别为变压器能量分配系数、电热泵能量 分配系数、空调能量分配系数、换热器能量分配系数、吸收式制冷机能量分配系数；η_EHP、η_AC、η_HE、η_AF分别为电热泵效率、空调效率、换热效率、吸收式制冷机效率。In the formula: L ^e and L ^h are the electricity and heat of the input node on the user side, respectively; ^De , D ^h , and D ^c are the electricity load, heat load and cooling load on the user side, respectively;

are the energy distribution coefficient of the transformer, the energy distribution coefficient of the electric heat pump, the energy distribution coefficient of the air conditioner, the energy distribution coefficient of the heat exchanger, and the energy distribution coefficient of the absorption chiller; η _EHP , η _AC , η _HE , and η _AF are the electric heat pump efficiency, Air conditioning efficiency, heat exchange efficiency, absorption chiller efficiency.

Any of the above-mentioned possible implementations further provides an implementation. In step S2.1,

The forecast value of wind power and photovoltaic output is obtained according to the historical output data of wind turbines and photovoltaics;

The error of wind power and photovoltaic output forecast value obeys a normal distribution with a mean value of 0 and the standard deviation as follows:

分别为风电出力预测值和光电出力预测值；In the formula: σ _wt , σ _pv are the standard deviation of the wind power output forecast value error and the standard deviation of the photovoltaic output forecast value error, respectively; λ _wt , λ _pv are the standard deviation coefficient of the wind power output forecast value error and the photovoltaic output forecast value error, respectively The standard deviation coefficient of ;

are the wind power output forecast value and the photovoltaic output forecast value respectively;

The grid space of renewable energy is limited, so the actual use value of wind power P _wt and the actual use value of photovoltaic power P _pv fluctuate within the interval less than the predicted value plus the prediction error;

In the formula: δ _wt , δ _pv are the wind power forecast value error and the photovoltaic forecast value error, respectively.

Any of the above-mentioned possible implementations further provides an implementation. In step S2.2, the user-side load is obtained through historical data; the user-side load includes electrical load, heating load and cooling load;

The user-side load error follows a normal distribution with mean 0 and standard deviation as follows:

In the formula: σ ^e , σ ^h , σ ^c are the standard deviation of the electrical load error, the standard deviation of the heating load error and the standard deviation of the cooling load error, respectively; λ _ed , λ _hd , λ _cd are the standard deviation of the electrical load error, respectively coefficient, standard deviation coefficient of heating load error and standard deviation coefficient of cooling load error;

According to the demand characteristics of electric vehicles, electric vehicles are divided into rigid EVs and flexible EVs, and flexible EVs are further divided into fast-charging flexible EVs and slow-charging flexible EVs; a one-dimensional matrix is used to describe the electric vehicle model:

Ω=[LGS _n S _e T _s T _c ];

In the formula: L represents the load type of the electric vehicle, and the numbers 1, 2, and 3 represent the rigid EV load, the slow-charging flexible EV charging and discharging load, and the fast-charging flexible EV charging and discharging load; G represents the electric vehicle charging and discharging mark, which is in the charging mode. is 1, the discharge mode is -1, and the rest time is 0; _Sn and _{S e} represent the charge amount when the EV is stopped and the user's expected charge amount when the EV is off- _grid , respectively; car off-grid time;

Rigid EVs do not participate in grid scheduling; flexible EVs participate in grid scheduling;

Build a flexible EV charging and discharging model:

The internal charge of the flexible EV satisfies the following expression:

SOC _t =SOC _t-1 +P _c -P _d ;

In the formula: P _c is the charging power of the flexible EV; P _d is the discharging power of the flexible EV; SOC _t is the electric charge of the flexible EV at time t, SOC _t-1 is the electric charge of the flexible EV at time t-1, and the time t-1 represents The last metering time at time t.

Any of the above-mentioned possible implementations further provides an implementation. In step S3,

S3.1 The calculation method of carbon emission cost on the supply side of the park energy system is:

S3.1.1 Use the baseline method to determine the free carbon emission quota on the supply side of the energy system in the park:

In the formula: CE′ is the free carbon emission quota for the energy system of the park; β _CHPE is the carbon emission quota obtained by the unit power supply of the CHP unit, β _CHPH and β _GB are the carbon emission quota obtained by the unit heat supply of the CHP unit and the gas boiler unit, respectively ; β _EGrid is the carbon emission quota obtained by purchasing a unit of electricity from the upper power grid;

S3.1.2 When the carbon emission of the energy system in the park is higher than the free carbon emission quota, the carbon emission right is purchased from the carbon trading market. The higher carbon emission range corresponds to the higher carbon trading price;

S3.1.3 Calculate the carbon emission cost C _co2 on the supply side of the energy system in the park: the carbon emission cost is calculated by using the stepped carbon price;

The carbon emissions of the park energy system are calculated by the following formula:

CE=β _e E ^e +β _g E ^g

In the formula: CE is the carbon emission of the energy system of the park, β _e is the carbon emission produced by the unit power generation of the upper power grid, and β _g is the carbon emission produced by the unit of natural gas consumption;

The carbon emission cost C _co2 on the supply side of the energy system in the park is:

c ^co2 is the carbon price per unit of carbon emission;

S3.2 The calculation method of the operating cost C _b on the supply side of the energy system in the park is:

S3.3 The calculation method of the abandoned wind and solar power cost C _p on the supply side of the park energy system is:

分别为 风电预测值和光电预测值。In the formula: C _b and C _p are the operating cost and the cost of abandoning wind and solar, respectively; c ^e , c ^g , and ^ch are the electricity price, gas price, and heat price, respectively; c _p is the unit cost of abandoning the wind and wind; P _wt , P _pv are respectively The actual use value of wind power and the actual use value of photovoltaic,

are the wind power forecast value and the photovoltaic forecast value, respectively.

Any of the above-mentioned possible implementation manners further provides an implementation manner. In step S3, the total cost on the supply side also includes the user comfort cost, and the user comfort cost is the actual electrical load and thermal load of the user. , The cost incurred when there is an offset between the cooling load and the energy provided by the park energy system; offset refers to the difference between the energy supply and the load caused by the park energy system failing to meet the actual load demand of the user.

Any of the above-mentioned possible implementations further provides an implementation. In step S4, the salps group algorithm is used to solve the objective function.

Any of the above-mentioned possible implementations further provides an implementation, and the specific method for solving the objective function using the salps group algorithm is:

S4.1 Let the search space be an N×D Euclidean space, D is the space dimension, and N is the number of populations; the nth population X _n =[X _n1 ,X _n2 ,…,X _nD ] ^T represents the nth species The output of each equipment on the supply side and the user side in the scenario, including the electrical output of the CHP unit, the thermal output of the CHP unit, the output of the GB unit, the output of the electric heat pump, the output of the air conditioner, and the output of the absorption refrigeration unit; _Fn represents the total output of the nth population Xn. Cost; n=1,2,3,...,N; the upper bound of the search space is ub=[ub ₁ ,ub ₂ ,...,ub _D ], and the lower bound lb=[lb ₁ ,lb ₂ ,...,lb _D ] ; ub ₁ , ub ₂ , ..., ub _n are the upper limit values of the output of the equipment on the supply side and the user side, respectively; lb ₁ , lb ₂ , ..., lb _n are the lower output values of the equipment on the supply side and the user side, respectively limit value;

S4.2 Initialize the population; according to the upper and lower bounds of each dimension of the search space, initialize a salps group with a scale of N×D;

S4.3 Calculate the total cost Fn of various groups Xn; select food: sort the population Xn according to the value of Fn from small to large, the group with the smallest total cost Fn in the first place is set as the current food, and X ₀ is the current food Food; selected leaders and followers: In addition to the current food X ₀ , there are N-1 remaining populations in the population, according to the order of the salps populations, the first half of the populations are regarded as leaders, and the rest are regarded as followers By;

S4.4: Leader position update

During the movement and foraging of the salp chain, the leader's position update is expressed as:

x_0d分别是第n个群体第d维的设备出力和食物第d维的设备出力；ub_d和lb_d分别是对应的第d维的设备出力的上界和下界；c₂、c₃是控制参数；c₁是优化算法中的收敛 因子；where:

x _0d are the equipment output of the nth dimension of the d-th dimension and the equipment output of the d-th dimension of food respectively; ub _d and lb _d are the upper and lower bounds of the corresponding equipment output of the d-th dimension; c ₂ , c ₃ are Control parameters; c ₁ is the convergence factor in the optimization algorithm;

_The expression for c1 is:

where: iter is the current number of iterations; maxiter is the maximum number of iterations;

S4.5 Follower Location Update

During the movement and foraging of the salps chain, the followers move forward in a chain-like manner through the mutual influence between the front and rear individuals; their displacement conforms to Newton's law of motion, and the movement displacement of the follower is:

In the formula: a is the acceleration, and the calculation formula is a=v _final /iter; and

After simplification, it is expressed as:

分别是更新前彼此紧连的两个追随者的第d维的设备出力；

为更新后追随者第d维中的设备出力；where:

are the device outputs of the d-th dimension of the two followers that are closely connected to each other before the update;

Contribute to the updated follower's device in the d-th dimension;

S4.6 Calculate the total cost of various groups after the update, and compare the total cost of each group after the update with the total cost of the current food. If the total cost of a certain group after the update is less than the total cost of the current food, take the total cost of The lowest cost species as new food;

S4.7 Repeat steps S4.4-S4.6 until a certain number of iterations is reached or the total cost reaches the termination threshold. After the termination condition is met, the current food corresponds to the optimal solution with the smallest total cost.

On the other hand, the present invention also provides an information data processing terminal for realizing the above-mentioned method for optimizing the operation of a park energy system.

In another aspect, the present invention also provides a computer-readable storage medium, comprising instructions, when executed on a computer, to cause the computer to execute the above-mentioned method for optimizing the operation of a park energy system.

The beneficial effects of the present invention are:

1. The park energy system of the present invention considers the uncertainty of the user's energy consumption behavior and the random access of electric vehicles on the user side, and considers the uncertainty of the output of fans and photovoltaic units on the supply side.

2. The operation optimization strategy proposed by the present invention combines the energy flow and the carbon emission flow model, and takes into account carbon emission, operation cost, abandoned wind and solar cost, and user comfort cost, and takes the lowest total cost as the optimization goal. Make full use of renewable energy and reduce carbon emissions.

3. The present invention uses the salps swarm algorithm to solve the model, and proposes a comprehensive analysis solution for multiple scenarios. The simulation results show that the operation optimization model proposed in this paper is robust.

Description of drawings

FIG. 1 shows an algorithm flow chart of the Ascidian swarm algorithm in the embodiment.

Figure 2 shows the electricity price change diagram (within 24 hours a day) in the embodiment.

FIG. 3 shows the curve of the electricity price change in the embodiment.

FIG. 4 shows the wind and solar forecasting output diagram in the embodiment.

FIG. 5 shows the predicted value of the electric heating load in the embodiment.

FIG. 6 is a schematic diagram of electric power balance in scenario 2 in the embodiment.

FIG. 7 is a schematic diagram showing the thermal power balance of scenario 2 in the embodiment.

FIG. 8 shows the structure diagram of the comprehensive energy system of the park in the embodiment.

Detailed ways

Hereinafter, specific embodiments of the present invention will be described in detail with reference to the accompanying drawings. It should be noted that the technical features or combinations of technical features described in the following embodiments should not be considered isolated, and they can be combined with each other to achieve better technical effects.

An embodiment of the present invention is a method for optimizing the operation of a park energy system. The structure of the park energy system is shown in Figure 8, and the method includes:

S1. Divide the energy system of the park into the supply side and the user side. The users are multi-energy users. In order to increase the user's ability to respond to energy demand, the user side is equipped with energy storage equipment, electric heat pumps, absorption refrigeration units, air conditioning equipment and Electric vehicles are equipped with gas boilers, CHP units, photovoltaic units and wind turbines on the supply side, and can interact with the upper-level power grid at the same time;

S2. On the supply side, considering the output uncertainty of photovoltaic units and wind turbines, establish a hub energy flow model and a carbon emission flow model on the supply side; on the user side, consider the randomness of electric vehicle access , establish the energy hub matrix model of the user side;

S3. According to the supply-side hub energy flow model and carbon emission flow model obtained in step S2, and the user-side energy hub matrix model, calculate the total cost of the supply side. The total cost includes operation cost, carbon emission cost and Abandoning the cost of wind and solar, the objective function of the park energy system operation optimization is to minimize the total cost; the total cost can also include the cost of user comfort;

S4. Solve the objective function in step S3 to obtain an optimal solution, and optimize the operation of the park energy system according to the optimal solution.

In a specific embodiment, the specific steps of step S2 are:

S2.1 By purchasing electricity, heat, natural gas and other energy sources from superiors, and managing energy conversion equipment to meet the multi-energy load of users; the hub energy flow model and carbon emission flow model on the supply side are as follows:

η_GB分别为变压器效率、CHP机组制电效率、CHP机组制热效率和燃气锅炉效率；

分别为CHP 机组燃气分配系数、燃气锅炉机组燃气分配系数，E^e、E^g、E^h分别为园区能源系统向上级 网络购买的电、气、热能量，P_wt、P_pv分别为风电实际使用值和光电实际使用值；In the formula, P ^e and ^Ph are the electrical power and thermal power of the output node on the supply side, respectively; η _T ,

η _GB is the transformer efficiency, the CHP unit electricity production efficiency, the CHP unit heating efficiency and the gas boiler efficiency;

are the gas distribution coefficient of CHP units and gas boiler units, respectively, E ^e , E ^g , and E ^h are the electricity, gas, and heat energy purchased by the park energy system from the upper-level network, respectively, and P _wt and P _pv are the actual use of wind power, respectively value and photoelectric actual use value;

S2.2 On the energy demand side, the user purchases electrical and thermal energy resources from the system operator, and uses various energy-consuming equipment to meet the energy demand for electricity, heat and cooling. The user's energy hub matrix model can be established as follows:

分别为变压器能量分配系数、电热泵能量 分配系数、空调能量分配系数、换热器能量分配系数、吸收式制冷机能量分配系数；η_EHP、η_AC、η_HE、η_AF分别为电热泵效率、空调效率、换热效率、吸收式制冷机效率。In the formula: L ^e and L ^h are the power and heat of the input node on the user side, respectively; ^De , D ^h , and D ^c are the power demand, heat demand and cooling demand on the user side, respectively;

are the energy distribution coefficient of the transformer, the energy distribution coefficient of the electric heat pump, the energy distribution coefficient of the air conditioner, the energy distribution coefficient of the heat exchanger, and the energy distribution coefficient of the absorption chiller; η _EHP , η _AC , η _HE , and η _AF are the electric heat pump efficiency, Air conditioning efficiency, heat exchange efficiency, absorption chiller efficiency.

In a specific embodiment, in step S2.1, considering the prediction error of wind and solar output and the prediction error of load, consumed wind and wind output≤prediction value+prediction error, load=prediction value of load+prediction error;

The predicted value of wind and solar output can be obtained from historical output data of wind turbines and photovoltaic units. In a specific embodiment, the predicted value of wind and solar output is shown in FIG. 4 .

The errors of wind and light output prediction values obey the normal distribution with mean 0 and standard deviation as follows:

分别为风电出力预测值和光电出力预测值；In the formula: σ _wt , σ _pv are the standard deviation of the wind power output forecast value error and the standard deviation of the photovoltaic output forecast value error, respectively; λ _wt , λ _pv are the standard deviation coefficient of the wind power output forecast value error and the photovoltaic output forecast value error, respectively The standard deviation coefficient of ;

are the wind power output forecast value and the photovoltaic output forecast value respectively;

The grid space of renewable energy is limited, so the actual use value of wind power P _wt and the actual use value of photovoltaic power P _pv fluctuate within the interval less than the predicted value plus the prediction error;

In the formula: δ _wt , δ _pv are the wind power forecast value error and the photovoltaic forecast value error, respectively.

In step S2.2, the load on the user side is obtained from historical data; the load on the user side includes electric load, heating load and cooling load; for example, in a specific embodiment, the electricity, heating and cooling loads are predicted and obtained from historical load. The load is shown in Figure 5.

The user-side load error follows a normal distribution with a mean of 0 and a standard deviation as follows:

In the formula: σ ^e , σ ^h , σ ^c are the standard deviation of the electrical load error, the standard deviation of the heating load error and the standard deviation of the cooling load error, respectively; λ _ed , λ _hd , λ _cd are the standard deviation of the electrical load error, respectively coefficient, standard deviation coefficient of heating load error and standard deviation coefficient of cooling load error.

Due to the volatility and randomness of wind power output, large-scale wind power access brings great challenges to the optimal dispatching of the power generation side of the power system. According to the difference of electric vehicle charging demand, this paper divides electric vehicle EVs into three categories: the first category is rigid EVs, the average daily driving time is relatively long, and the charging speed and charging time are relatively high, so they cannot participate in grid interaction; the second category is rigid EVs. The type is fast charging and flexible EV, which has fast scheduling response and can reduce the load of the regional power grid in a short time. However, the fast charging mode will have a huge current impact on the power battery pack, and the battery will heat up seriously during charging, which will reduce the cycle life of the power battery pack. . At the same time, it is necessary to use the operator's dedicated charging pile, and the charging fee will increase the service fee of the charging pile operator on the basis of the industrial electricity price; the third type is the slow charging flexible EV, which can transfer a large number of EV charging time periods. Small, it can reduce the heating phenomenon when the battery is charging, thereby prolonging the battery life; EV slow charging period generally occurs during the low-peak period of electricity consumption, using residential charging piles or private charging piles, and the charging electricity price is the residential electricity price;

This application describes the demand characteristics of electric vehicles in detail, and divides electric vehicles into rigid EVs, fast-charging flexible EVs, and slow-charging flexible EVs. A one-dimensional matrix is used to describe the EV model:

Ω=[LGS _n S _e T _s T _c ]

In the formula: L represents the load type of the electric vehicle, and the numbers 1, 2, and 3 represent the rigid EV load, the slow-charging flexible EV charging and discharging load, and the fast-charging flexible EV charging and discharging load; G represents the electric vehicle charging and discharging mark, which is in the charging mode. is 1, the discharge mode is -1, and the rest time is 0; Sn and _Se represent the state of charge when the EV is stopped and the user _'s expected charge when it is off-grid, respectively; network time;

Among them, rigid vehicles are vehicles that do not participate in grid dispatching. By analyzing the driving data of private vehicles in the United States, the data is fitted to obtain the predicted probability density curve of the daily mileage of electric vehicles, and the probability density curve of its history approximates a normal distribution. Taking the end of driving time as the charging time, the probability density curve of the time to connect to the power grid is obtained. The rigid EV model is obtained using Monte Carlo sampling. For the flexible EV charging and discharging model, the time when the flexible EV finishes driving is taken as the time when it is connected to the power grid, and the internal state of charge of the electric vehicle satisfies the following expression:

SOC _t =SOC _t-1 +P _c -P _d

where P _c is the charging power of the flexible EV; P _d is the discharging power of the flexible EV; SOC _t is the electric charge of the EV at time t, SOC _t-1 is the electric charge of the EV at time t-1, and time t-1 represents time t the last measurement time.

In a specific embodiment, in step S3,

The calculation method of carbon emission cost on the supply side of the energy system in the park is as follows:

S3.1.1 Use the baseline method to determine the free carbon emission quota on the supply side of the energy system in the park:

In the formula: CE′ is the free carbon emission quota of the energy system of the park; β _CHPE is the carbon emission quota obtained by the unit power supply of the CHP unit, β _CHPH and β _GB are the carbon emission quota obtained by the unit heat supply of the CHP unit and the gas boiler unit; β _EGrid is a carbon emission allowance obtained by purchasing a unit of electricity from the upper power grid.

S3.1.2 When the carbon emission of the energy system in the park is higher than the carbon emission quota allocated for free, the carbon emission right is purchased from the carbon trading market. The higher carbon emission range corresponds to the higher carbon trading price; When the carbon emission quota is lower than the free allocation, a certain subsidy is given (or no subsidy, only free);

S3.1.3 The carbon emission cost C _co2 on the supply side of the energy system in the park, and the carbon emission cost is calculated by using the stepped carbon price; this paper assumes that the electricity purchased from the upper power grid comes from thermal power, and the carbon emission of the park mainly comes from the consumption of natural gas and the purchase of electricity. The carbon emission of the power generation side, the carbon emission of the park is calculated by the following formula:

CE=β _e E ^e +β _g E ^g

In the formula: CE is the carbon emission of the park, β _e is the carbon emission generated by the unit power generation of the upper-level power grid, and β _g is the carbon emission generated by the unit of natural gas consumption.

The carbon emission cost C _co2 is:

c ^co2 is the carbon price per unit of carbon emission. When the user participates in the response, there is an offset between the actual electricity, heat, and cooling energy loads and the energy supplied by the system. In order to compensate the user, the user comfort cost is defined according to the user energy offset.

The mathematical description of the operating cost and the cost of abandoning the scenery is shown in the following formula.

分别为风电、光电预测值。In the formula: C _b and C _p are the operating cost and the cost of abandoning wind and solar, respectively, E ^e , E ^g , and E ^h are the purchase of electricity, gas, and heat, respectively, and c ^e , c ^g , and c ^h are the price of electricity and gas, respectively , heat price; c _p is the unit cost of abandoning wind and solar power; P _wt , P _pv are the actual use values of wind power and photovoltaics,

are the forecast values of wind power and photovoltaics, respectively.

In a specific embodiment, in step S4, the salps group algorithm is used to solve the objective function.

S4.1 Let the search space be an N×D Euclidean space, D is the space dimension, and N is the number of populations; the nth population X _n =[X _n1 ,X _n2 ,…,X _nD ] ^T represents the nth species The output of each equipment on the supply side and the user side in the scenario, including the electrical output of the CHP unit, the thermal output of the CHP unit, the output of the GB unit, the output of the electric heat pump, the output of the air conditioner, and the output of the absorption refrigeration unit; _Fn represents the total output of the nth population Xn. Cost; n=1,2,3,...,N; the upper bound of the search space is ub=[ub ₁ ,ub ₂ ,...,ub _D ], and the lower bound lb=[lb ₁ ,lb ₂ ,...,lb _D ] ; ub ₁ , ub ₂ , ..., ub _n are the upper limit values of the output of the equipment on the supply side and the user side, respectively; lb ₁ , lb ₂ , ..., lb _n are the lower output values of the equipment on the supply side and the user side, respectively limit value;

S4.2 Initialize the population; according to the upper and lower bounds of each dimension of the search space, initialize a salps group with a scale of N×D;

S4.3 Calculate the total cost Fn (population fitness) of various groups Xn; select food: sort the population Xn according to the value of Fn from small to large, and the group with the smallest total cost Fn in the first place is set as the current food, Denote X ₀ as the current food; select leaders and followers: in addition to the current food X ₀ , there are N-1 remaining populations in the population, according to the order of the salps populations, the first half of the populations are regarded as leaders, The rest of the population are considered followers;

S4.4: Leader position update

During the movement and foraging of the salp chain, the leader's position update is expressed as:

x_0d分别是第n个群体第d维的设备出力和食物第d维的设备出力；ub_d和lb_d分别是对应的第d维的设备出力的上界和下界；c₂、c₃是控制参数；c₁是优化算法中的收敛 因子；where:

x _0d are the equipment output of the nth dimension of the d-th dimension and the equipment output of the d-th dimension of food respectively; ub _d and lb _d are the upper and lower bounds of the corresponding equipment output of the d-th dimension; c ₂ , c ₃ are Control parameters; c ₁ is the convergence factor in the optimization algorithm;

_The expression for c1 is:

where: iter is the current number of iterations; maxiter is the maximum number of iterations;

S4.5 Follower Location Update

During the movement and foraging of the salps chain, the followers move forward in a chain-like manner through the mutual influence between the front and rear individuals; their displacement conforms to Newton's law of motion, and the movement displacement of the follower is:

In the formula: a is the acceleration, and the calculation formula is a=v _final /iter; and

After simplification, it is expressed as:

分别是更新前彼此紧连的两个追随者的第d维的设备出力；

为更新后追随者第d维中的设备出力；where:

are the device outputs of the d-th dimension of the two followers that are closely connected to each other before the update;

Contribute to the updated follower's device in the d-th dimension;

S4.6 Calculate the total cost of various groups after the update, and compare the total cost of each group after the update with the total cost of the current food. If the total cost of a certain group after the update is less than the total cost of the current food, take the total cost of The lowest cost species as new food;

S4.7 Repeat steps S4.4-S4.6 until a certain number of iterations is reached or the total cost reaches the termination threshold. After the termination condition is met, the current food corresponds to the optimal solution with the smallest total cost.

In order to verify the operation optimization method of the park energy system proposed in this application, two scenarios are proposed below for comparison and verification.

The initial data are shown in Figures 2-5.

This example uses Monte Carlo sampling to sample 100 electric vehicles, and sets two scenarios:

Scenario 1 is that the electric vehicle does not participate in the scheduling plan, that is, when the power of the car owner at the end of the trip is less than the expected power of the second trip, the electric car is charged, and the charging start time is the time when the owner ends the trip, and the end time is when the charging reaches the expected power time or the start time of the owner's second trip.

Scenario 2 is that electric vehicles participate in the scheduling plan. That is, the owner of the electric vehicle starts to connect to the grid at the end of the trip, and the off-grid time is the start time of the second trip. Regardless of whether the remaining power of the car is greater or less than the expected power, the electric vehicle will participate in the scheduling and act as a virtual energy storage. The constraint is to ensure that each vehicle When off-grid, the power is greater than or equal to the expected power of the car owner.

Figure 3 shows the electricity price change in scenario 2. After the electric vehicle participates in the dispatch, it is optimized on the basis of the original electricity price.

Figure 6 shows the electric power balance curve of scenario 2. It can be seen that most of the electric vehicles are charged from 23:00 to 5:00 in the morning. At this time, the electricity price is relatively cheap, the user's electricity consumption is small, the output of renewable energy is high, and the electric vehicle is charged. Can effectively consume renewable energy. In the afternoon, the user load is high, and the discharge of the electric vehicle meets the user's requirements. The electric vehicle plays the role of energy storage and reduces the operating cost.

Figure 7 shows the thermal load power balance curve of scenario 2, in which the thermal load mainly depends on the supply of the upper-level heating network, gas boilers and CHP units. The output of the CHP unit is affected by the electrical load. When the electrical load is low, the power supply of the CHP unit At the same time, the heat supply is also reduced, and the heat load mainly depends on the upper-level supply. When the electrical load is high, the output of the CHP unit increases, and the heating capacity of the CHP unit increases at the same time, and the heat load mainly depends on the supply of the CHP unit.

The data of the running results are shown in Table 1.

Table 1 Operating costs of each scenario

Scenes energy purchase cost Abandoning the cost of scenery comfort cost carbon cost total cost scene 1 4511 35 0 507 5053 scene 2 4110 0 0 441 4551

It can be seen from Table 1 that the total cost of Scenario 2 is significantly lower than that of Scenario 1, which indicates that the participation of electric vehicles in dispatching will increase the capacity of the system's virtual energy storage system and improve the system's consumption of renewable energy, thereby reducing the cost of purchasing energy. and reduce carbon emissions.

Although several embodiments of the present invention have been presented herein, those skilled in the art should understand that changes may be made to the embodiments herein without departing from the spirit of the present invention. The above-mentioned embodiments are only exemplary, and the embodiments herein should not be construed as limiting the scope of the rights of the present invention.

Claims (10)

1. A method for optimizing operation of a park energy system, the method comprising the steps of:

s1, dividing a park energy system into a supply side and a user side, wherein the supply side comprises a gas boiler, a CHP unit, a photovoltaic unit and a wind turbine unit, and the user side comprises energy storage equipment, an electric heat pump, an absorption refrigeration unit, air conditioning equipment and an electric automobile;

s2, at the supply side, considering output uncertainty of a photovoltaic unit and a wind turbine unit, and establishing a hub energy flow model and a carbon emission flow model of the supply side; on the user side, considering the randomness of electric automobile access, and establishing an energy hub matrix model on the user side;

s3, calculating the total cost of the supply side according to the supply side hub energy flow model and the carbon emission flow model obtained in the step S2 and the user side energy hub matrix model, wherein the total cost comprises the operation cost, the carbon emission cost and the wind and light abandoning cost, and the objective function of the operation optimization of the park energy system is the minimum total cost;

s4, solving the objective function in the step S3 to obtain an optimal solution, and optimizing the operation of the energy system of the park according to the optimal solution.

2. The operation optimization method for the park energy system according to claim 1, wherein the step S2 is specifically:

s2.1 the energy junction model of the supply side is as follows:

in the formula, P^e、P^hRespectively outputting the electric quantity and the heat quantity of a node at the supply side; eta_T、

η_GBThe efficiency of the transformer, the heating efficiency of the CHP unit and the efficiency of the gas boiler are respectively;

respectively a CHP unit gas distribution coefficient, a gas boiler unit gas distribution coefficient, E^e、E^g、E^hElectric energy, air quantity, heat energy P purchased from park energy system to superior network_wt、P_pvRespectively representing the actual wind power utilization value and the actual photoelectric utilization value;

s2.2 the energy hub matrix model at the user side is as follows:

in the formula: l is^e、L^hAre respectively provided withInputting the electric quantity and the heat quantity of the node for a user side; d^e、D^h、D^cRespectively user side electrical load, thermal load and cold load;

the energy distribution coefficients of the transformer, the electric heat pump, the air conditioner, the heat exchanger and the absorption refrigerator are respectively; eta_EHP、η_AC、η_HE、η_AFRespectively the efficiency of an electric heat pump, the efficiency of an air conditioner, the efficiency of heat exchange and the efficiency of an absorption refrigerator.

3. The park energy system operation optimization method of claim 2, wherein in step S2.1, the wind power actual usage value P_wtAnd a photoelectric actual use value P_pvFluctuating in an interval less than the predicted value plus the prediction error:

in the formula:

respectively obtaining a wind power output predicted value and a photoelectric output predicted value; delta_wt、δ_pvRespectively a wind power predicted value error and a photoelectric predicted value error;

the wind power output predicted value and the photoelectric output predicted value are respectively obtained according to historical output data of the wind turbine generator and the photovoltaic generator;

the obeying mean value of the wind power output predicted value error and the photoelectric output predicted value error is 0, and the standard deviation is normally distributed as follows:

in the formula: sigma_wt、σ_pvAre respectively windStandard deviation of the error of the predicted value of the electric output and standard deviation of the error of the predicted value of the photoelectric output; lambda [ alpha ]_wt、λ_pvThe standard deviation coefficient of the wind power output predicted value error and the standard deviation coefficient of the photoelectric output predicted value error are respectively.

4. The park energy system operation optimization method of claim 2, wherein in step S2.2, the user side load is derived from historical data; the user side loads comprise an electric load, a heat load and a cold load;

the user side load error follows a normal distribution with a mean value of 0 and standard deviation as follows:

in the formula: sigma^e、σ^h、σ^cRespectively, the standard deviation of the electric load error, the standard deviation of the heat load error and the standard deviation of the cold load error; lambda [ alpha ]_ed、λ_hd、λ_cdRespectively is a standard deviation coefficient of an electric load error, a standard deviation coefficient of a heat load error and a standard deviation coefficient of a cold load error;

according to the demand characteristics of the electric automobile, the electric automobile is divided into a rigid EV and a flexible EV, wherein the flexible EV is divided into a fast charging flexible EV and a slow charging flexible EV; the electric automobile model is described by adopting a one-dimensional matrix:

Ω＝[L G S_n S_e T_s T_c]；

in the formula: l represents the type of the electric vehicle load, and numbers 1,2 and 3 represent rigid EV load, slow charge flexible EV charge-discharge load and fast charge flexible EV charge-discharge load respectively; g represents a charging and discharging identifier of the electric automobile, the charging mode is 1, the discharging mode is-1, and the rest moments are 0; s_nAnd S_eRespectively representing the charge quantity when the EV stops and the charge quantity expected by a user when the EV leaves the network; t is_sAnd T_cRespectively representing the network access time and the network leaving time of the electric automobile;

the rigid EV does not participate in power grid dispatching; flexible EV participates in power grid dispatching;

establishing a flexible EV charge-discharge model:

the internal charge amount of the flexile EV satisfies the following expression:

SOC_t＝SOC_t-1+P_c-P_d；

in the formula: p_cCharging power for a flexible EV; p_dDischarge power for flexible EV; SOC_tAmount of charge, SOC, for Flexible EV at time t_t-1For the amount of charge of the agile EV at time t-1, time t-1 represents the last metering time at time t.

5. The campus energy system operation optimizing method of claim 2 wherein, in step S3,

s3.1, the method for calculating the carbon emission cost at the supply side of the energy system of the park comprises the following steps:

s3.1.1, determining the uncompensated carbon emission quota of the supply side of the park energy system by adopting a reference line method:

in the formula: CE' is the uncompensated carbon emission quota of the park energy system; beta is a_CHPECarbon emission quota, beta, obtained for a unit supply of the CHP unit_CHPH、β_GBCarbon emission quotas obtained for the unit heat supply of the CHP unit and the gas boiler unit respectively; beta is a_EGridA carbon emission quota obtained for purchasing a unit of electricity to an upper-level power grid;

s3.1.2 when the carbon emission of the park energy system is higher than the uncompensated carbon emission quota, purchasing the carbon emission right to the carbon trading market, wherein the larger the carbon emission is, the higher the corresponding carbon trading price is;

s3.1.3 calculating the carbon emission cost C of the park energy system supply side_co2: the carbon emission cost is calculated by adopting a step carbon price;

the carbon emissions of the park energy system are calculated by the following formula:

CE＝β_eE^e+β_gE^g；

in the formula: CE is the carbon emission, beta, of the energy system of the park_eCarbon emission, beta, generated for the unit generation of the upper level grid_gCarbon emissions per unit of natural gas consumed;

carbon emission cost C of supply side of park energy system_co2Comprises the following steps:

c^co2carbon price per carbon emission;

s3.2 operating cost C of energy supply side of park_bThe calculation method comprises the following steps:

s3.3 abandoned wind and light cost C of energy system supply side of park_pThe calculation method comprises the following steps:

in the formula: c_b、C_pRespectively the running cost and the wind and light abandoning cost; c. C^e、c^g、c^hRespectively the electricity price, the gas price and the heat price; c. C_pAbandoning the wind and light cost for a unit; p_wt、P_pvRespectively a wind power actual use value and a photoelectric actual use value,

respectively a wind power predicted value and a photoelectric predicted value.

6. The campus energy system operation optimizing method of claim 2, wherein in step S3, the total cost of the supply side further includes a user comfort cost, the user comfort cost is a cost generated when a deviation occurs between an actual electric load, a thermal load, a cold load and energy supplied from the campus energy system; skew refers to the difference in energy supply and load that the park energy system fails to meet the actual load demand of the user.

7. The method for optimizing operation of a park energy system according to claim 2, wherein in step S4, the objective function is solved using the zun sea squirt group algorithm.

8. The park energy system operation optimization method of claim 7, wherein the specific method for solving the objective function by using the zun sea squirt group algorithm is as follows:

s4.1, setting a search space to be an Euclidean space of NxD, wherein D is a space dimension and N is the number of the populations; nth population X_n＝[X_n1,X_n2,…,X_nD]^TRepresenting the output of each device at the supply side and the user side in the nth scene, including the electrical output of a CHP unit, the thermal output of the CHP unit, the output of a GB unit, the output of an electric heat pump, the output of an air conditioner and the output of an absorption refrigeration unit; f_nRepresenting the total cost of the nth population Xn; n-1, 2,3, …, N; the upper bound of the search space is ub ═ ub₁,ub₂,…,ub_D]Lower bound lb ═ lb₁,lb₂,…,lb_D]；ub₁、ub₂、……、ub_nRespectively the upper limit value of the output of each device at the supply side and the user side; lb₁、lb₂、……、lb_nRespectively the lower limit value of the output of each device at the supply side and the user side;

s4.2, initializing a population; initializing a goblet sea squirt group with the scale of NxD according to the upper bound and the lower bound of each dimension of the search space;

s4.3, calculating the total cost Fn of each group Xn; selecting food: sorting the populations Xn from small to large according to the value of Fn, setting the population with the smallest total cost Fn ranked at the top as the current food, and recording X₀Is the current food; selecting a leader and a follower: except for current food X₀In addition, the rest N-1 groups are arranged according to the goblet sea squirt groupThe first half of the population is regarded as a leader, and the rest of the population is regarded as followers;

s4.4: leader location update

During the movement and foraging of the chain of goblet sea squirts, the position update of the leader is expressed as:

in the formula:

x_0drespectively the equipment output of the d-th dimension of the nth group and the equipment output of the d-th dimension of the food; ub_dAnd lb_dUpper and lower bounds for the corresponding d-th dimension of device contribution, respectively; c. C₂、c₃Is a control parameter; c. C₁Is a convergence factor in the optimization algorithm;

c₁the expression of (a) is:

wherein iter is the current iteration number; maximum is the maximum number of iterations;

s4.5 follower location update

In the process of moving and foraging of the goblet sea squirt chain, the followers sequentially advance in a chain shape through mutual influence between the front and the rear individuals; their displacement follows the newton law of motion, the displacement of the follower is:

wherein a is acceleration, and the calculation formula is a ═ v_finalA/iter; and is and

after simplification, the expression is:

in the formula:

respectively the d-dimension equipment output of two followers which are closely connected with each other before updating;

the output of the equipment in the d dimension of the updated follower is obtained;

s4.6, calculating the total cost of each updated population, comparing the total cost of each updated population with the total cost of the current food, and if the total cost of a certain population is less than the total cost of the current food after updating, taking the population with the minimum total cost as the new current food;

s4.7, repeating the steps S4.4-S4.6 until a certain iteration number is reached or the total cost reaches a termination threshold, and after the termination condition is met, the current food corresponds to the optimal solution with the minimum total cost.

9. An information data processing terminal for implementing the method for optimizing the operation of a park energy system according to any one of claims 1 to 8.

10. A computer-readable storage medium comprising instructions which, when executed on a computer, cause the computer to perform the method of optimizing the operation of a park energy system according to any one of claims 1 to 8.

Images