CN115700655B - Data center micro-grid operation optimization method based on demand side response - Google Patents
Data center micro-grid operation optimization method based on demand side responseInfo
- Publication number
- CN115700655B CN115700655B CN202211436226.7A CN202211436226A CN115700655B CN 115700655 B CN115700655 B CN 115700655B CN 202211436226 A CN202211436226 A CN 202211436226A CN 115700655 B CN115700655 B CN 115700655B
- Authority
- CN
- China
- Prior art keywords
- load
- power
- stage
- period
- data center
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Landscapes
- Supply And Distribution Of Alternating Current (AREA)
Abstract
本发明公开了一种基于需求侧响应的数据中心微电网运行优化方法，属于数据中心微电网优化领域，包括：分别获取负荷侧和发电侧的日前报价；负荷侧的负荷提供商包括DCMG；基于CVaR建立两阶段电力市场出清模型并求解；第一阶段基于负荷侧和发电侧的日前报价实现社会福利最大化，决策变量包括出清后发电侧各机组各时段的发电计划，以及出清后负荷侧各时段的负荷削减计划；第二阶段进行场景采样后，求解第一阶段决策变量在各场景下的调整量，调整后综合得到第二阶段决策变量，并由此计算出清电价；以DCMG运行成本最小化为目的，对DCMG进行日前调度优化，得到其中各设备的出力计划。本发明能够使数据中心运行成本降低并减少系统不确定性带来的风险。
The present invention discloses a data center microgrid operation optimization method based on demand side response, which belongs to the field of data center microgrid optimization, including: respectively obtaining the day-ahead quotations of the load side and the power generation side; the load providers on the load side include DCMG; establishing and solving a two-stage power market clearing model based on CVaR; in the first stage, social welfare is maximized based on the day-ahead quotations of the load side and the power generation side, and the decision variables include the power generation plan of each unit in each time period on the power generation side after clearing, and the load reduction plan of each time period on the load side after clearing; in the second stage, after scene sampling, the adjustment amount of the decision variables in the first stage under each scene is solved, and the decision variables in the second stage are obtained after adjustment, and the clearing electricity price is calculated accordingly; with the purpose of minimizing the operation cost of DCMG, the DCMG is optimized for day-ahead scheduling to obtain the output plan of each device therein. The present invention can reduce the operation cost of the data center and reduce the risks brought by system uncertainty.
Description
Technical Field
The invention belongs to the field of data center microgrid optimization, and particularly relates to a data center microgrid operation optimization method based on demand side response.
Background
Statistics show that the data scale in China reaches 64ZB in 2020, and still expands rapidly with a 50% annual average growth rate. Data centers as an important infrastructure can provide physical support for mass data computing, and with the rapid growth of their industry, industry power consumption is also growing at a rate of more than 10% per year, which makes the data centers incur high electricity costs per year.
As industrial-grade large-power users, the data center has strong flexible regulation capability on energy consumption, so that interaction with the power market can be performed through participation in demand response, and the running cost of the data center is effectively reduced. At present, in the aspect of interaction between a data center and an electric power market, the existing research shows that the interaction between the data center and the electric power market based on a demand side response (DemandResponse, DR) mechanism plays an important role in reducing the running cost of the data center and improving the economic stability of a power grid system.
However, there are still some difficulties in the interaction of data centers with the power market that need to be addressed. On the one hand, the problem of load uncertainty is generated while the flexible regulation capability of the self energy consumption of the data center is utilized in the design level of the power market clearing mechanism, but the risk avoidance problem in an uncertain environment is not considered in the existing power market clearing model, and on the other hand, the problem of how to combine the self load-reducible condition as a data center of a load provider to submit an optimal quotation strategy to an electric power system operator and how to schedule and optimize the operation of a micro grid of the data center under the market clearing frame is still to be solved in the interactive level of the data center and the power market.
Disclosure of Invention
Aiming at the defects and improvement demands of the prior art, the invention provides a data center micro-grid operation optimization method based on demand side response, which aims to reduce the operation cost of a data center and effectively reduce the risk caused by system uncertainty.
To achieve the above object, according to one aspect of the present invention, there is provided a data center microgrid operation optimization method based on a demand side response, comprising:
The method comprises the steps of respectively obtaining a day-ahead quotation on a load side and a power generation side, wherein the day-ahead quotation on the load side comprises total load demands of all time periods of all load providers and load electric quantity and corresponding prices which can be reduced, and the day-ahead quotation on the power generation side comprises power generation plans of all time periods of all units and corresponding prices;
Establishing a two-stage power market clearing model based on CVaR and solving decision variables, wherein in the two-stage power market clearing model, the optimization objective of the first stage is to realize the maximization of social benefits based on daily quotations of a load side and a power generation side, the decision variables comprise power generation plans of each period of a power generation side after clearing and load reduction plans of each period of the load side after clearing;
and aiming at minimizing the running cost of the DCMG, carrying out daily scheduling optimization on the micro-grid of the data center according to the power price and the load reduction plan of each period of the DCMG after the power price is cleared, and obtaining the output plan of each device in the DCMG.
Further, in the two-stage power market clearing model, the objective function of the first stage is:
Wherein x represents a set of decision variables T represents the total number of scheduling periods, I represents the index of the generator set on the generator side, i=1, 2..i, I represents the number of generator sets, j represents the index of price-electricity-quantity on the generator side to the bid section; the power generation plan of each unit in each period of time on the power generation side after cleaning is shown, Respectively representing the power, standby and frequency-modulated planned output results of the jth bidding section of the generator set i in the period t,Respectively represents the electricity, standby and frequency modulation prices of the j-th bidding sections of the generator set i,The total number of the bidding sections of the power, standby and frequency modulation of the generator set i is respectively represented; for a 0/1 variable, 0 means that genset i does not provide a frequency modulation during period t, 1 means that genset i provides a frequency modulation during period t, n means the index of the load provider, K means the index of the load side bid section, and K n means the total number of bid sections of the load provider n; Showing the load shedding schedule for each period on the load side after clearing, Shows the planned load non-reducible amount of load on the load side in the post-cleaning period t, P L shows the price of the load non-reducible amount of load on the load side,The amount of load that can be reduced in the kth bid section of the load provider n during the post-clearing period t is shown, and LP n,k shows the price that can be reduced in the kth bid section of the load provider n during the post-clearing period t.
Further, the constraints of the first stage include:
r1-1, the total amount of the power, standby and frequency modulation plans of each generator set does not exceed the power generation capacity;
r1-2, the power, standby and frequency modulation plan output of each unit in each period do not exceed the upper limit and the lower limit of bidding electric quantity;
r1-3, the generated energy of the power generation side is balanced with the load of the load side;
r1-4, wherein the load capacity of the load side does not exceed the upper limit and the lower limit of the bidding electric quantity;
R1-5, frequency modulation constraint, the expression is as follows:
r1-6, standby constraint, the expression is as follows:
wherein, the Represents a lower limit value that is defined in advance,Represents a predefined upper limit value, lambda 1 represents a preset risk adjustment coefficient, and lambda 2 represents a preset scaling coefficient.
Further, in the two-stage power market clearing model, the objective function of the second stage is:
Wherein S' represents the total number of scenes, S represents the scene index, ρ represents the risk avoidance factor, ρ E [0,1], η represents the threshold of the social benefit expected value, α represents the confidence level, ζ s represents the auxiliary variable in the S-th scene, p s represents the occurrence probability of the scene, and the decision variable set Representing the second stage adjustment of the first stage decision variable in the s-th scenario, Respectively, for the first stage decision variables in the s-th scene Is used for adjusting the adjustment amount of the (a).
Further, in the two-stage power market clearing model, the constraint conditions of the second stage further include:
r2-1, after adjustment, the total amount of the power, standby and frequency modulation plans of each generator set does not exceed the power generation capacity;
R2-2, after adjustment, the power, standby and frequency modulation plan output of each unit in each period do not exceed the upper limit and the lower limit of bidding electric quantity;
R2-3, after adjustment, the generated energy of the power generation side is balanced with the load of the load side;
r2-4, after being adjusted, the load capacity of the load side does not exceed the upper limit and the lower limit of the bidding electric quantity;
r2-5, frequency modulation constraint, the expression is as follows:
R2-6, standby constraint, the expression is as follows:
further, before solving the two-stage power market finding model, the method further comprises:
linearizing the frequency modulation constraint in two stages.
Further, for any load provider on the load side, the load capacity can be reduced in any period t in the daily front quotationAnd priceThe method comprises the following steps of:
Where U represents the index of the legacy units, u=1, 2,..u, U represents the total number of legacy units managed by the load provider; For a conservative pre-estimate of the period t payload, Representing the output of the conventional unit u during period t-1,Representing the upper power output limit of conventional unit u, RU u represents the ramp up rate of conventional unit u, and β u represents the fuel cost parameter of conventional unit u.
Further, the objective function of the data center microgrid for day-ahead scheduling optimization is as follows:
Wherein, alpha u represents the no-load cost of the traditional unit u; SU u represents the starting cost of the conventional unit u, SD u represents the shutdown cost of the conventional unit u, τ u,t and τ u,t-1 are both 0/1 variables, 1 represents that the conventional unit u is started in the corresponding period, and 0 represents that the conventional unit u is not started in the corresponding period; The power outlet price of the period t is represented, and P t grid represents the electricity purchase quantity in the period t.
Further, constraints of the data center microgrid for day-ahead scheduling optimization include:
Power balance constraint:
Force upper and lower limits and climbing rate constraint:
ESS operation constraints for electrical energy storage systems:
the data center is constrained by the capacity of the transmission line when the local area power grid is accessed to purchase power to the power grid:
Wherein P t disc represents the discharge capacity in the period t; For a 0/1 variable, 1 means discharge in a period t, 0 means no discharge in the period t, ER t means new energy generation in the period t, P t DC means power load of the data center in the period t, and P t char means charge quantity in the period t; A variable of 0/1, a1 indicating charging during period t, and a0 indicating no charging during period t; And Respectively represent the lower limit and the upper limit of the electric power output of the conventional unit u, RD u and RU u represent the descending and ascending climbing powers of the conventional unit u,ES t and ES t+1 respectively represent the charging states of the electric energy storage system in the time period t and t+1, ES min and ES max respectively represent the minimum and maximum charging states of the electric energy storage system, and eta char and eta disc respectively represent the charging efficiency and the discharging efficiency of the electric energy storage system; And Respectively representing the maximum charging power and the maximum discharging power of the electric energy storage system; Representing the transmission capacity of the line.
According to another aspect of the invention, a computer readable storage medium is provided, which comprises a stored computer program, and when the computer program is executed by a processor, the computer readable storage medium is controlled to execute the data center micro-grid operation optimization method based on the demand side response.
In general, through the above technical solutions conceived by the present invention, the following beneficial effects can be obtained:
(1) According to the invention, a data center micro-grid is used as a load side, a two-stage power market clearing model is built based on CVaR, when the model is built, a power generation plan of each period of each unit on the power generation side and the reducible load electric quantity and corresponding price of each period on the load side are considered at the same time, so that the response of a demand side can be considered, the flexible regulation capability of the self energy consumption of the data center micro-grid is fully utilized, meanwhile, in the two-stage power market clearing model, after a decision variable which maximizes social benefit is obtained by solving in the first stage, the decision variable regulation quantity in each scene is optimized and solved in the second stage on the basis of the decision variable in the first stage, the decision variable in each scene is regulated, finally, the final decision variable in each scene is obtained by integrating, and the current price and the load reduction plan in clearing are provided for the data center micro-grid, so that the data center micro-grid is optimized in operation, and the risk caused by uncertainty can be effectively reduced. In general, the invention reduces the running cost of the data center and effectively reduces the risk brought by the uncertainty of the system.
(2) According to the invention, a two-stage power market clearing model is established based on CVaR, and the risk avoidance degree can be flexibly controlled through the risk avoidance factor rho so as to adapt to different application requirements.
(3) In a preferred embodiment of the invention, the and 0/1 variable is compared prior to solving the two-stage power market modelAnd the related frequency modulation constraint is subjected to linearization processing, so that the efficiency and the precision of model solving can be improved.
(4) In the preferred scheme of the invention, in the day-ahead quotation provided by the load side, the maximum power generation capacity of the traditional unit in the data center micro-grid in each period is determined as the load-reducible electric quantity in the corresponding period, the linear marginal cost of the traditional unit (namely the fuel cost parameter of the traditional unit) is determined as the load-reducible price in the corresponding period, the quotation mechanism is consistent with the load condition in the data center micro-grid, the power market clearing model established based on the quotation information is effectively ensured to fully consider the response of the demand side, and the flexible regulation capability of the energy consumption of the data center micro-grid is fully utilized.
Drawings
FIG. 1 is a schematic diagram of a data center microgrid operation optimization method based on a demand side response according to an embodiment of the present invention;
FIG. 2 is a predicted power generation of a wind farm and a photovoltaic power plant provided by an embodiment of the present invention;
FIG. 3 shows a demand side response based power market price for electricity, including three resources including power, frequency modulation, and standby, according to an embodiment of the present invention;
FIG. 4 illustrates a reducible load shedding amount of a data center micro-grid after market clearing according to an embodiment of the present invention;
FIG. 5 is a graph showing social benefits and conditional risk value (CVaR) under different risk avoidance factors ρ provided by an embodiment of the present invention;
FIG. 6 shows scheduling results of an electric Energy Storage System (ESS) in different power rate contexts, wherein (a) is an ESS scheduling result in a fixed power rate context, and (b) is an ESS scheduling result in a time-of-use power rate context;
Fig. 7 shows scheduling results of each unit in a data center micro-grid (DCMG) under different power rate backgrounds, where (a) is the scheduling result of each unit in the DCMG under a fixed power rate background, and (b) is the scheduling result of each unit in the DCMG under a time-of-use power rate background.
Detailed Description
The present invention will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present invention more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention. In addition, the technical features of the embodiments of the present invention described below may be combined with each other as long as they do not collide with each other.
In the present invention, the terms "first," "second," and the like in the description and in the drawings, if any, are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order.
In order to reduce the running cost of the data center and effectively reduce the risk brought by the uncertainty of the system, the invention provides a data center micro-grid running optimization method based on the response of a demand side, which has the following overall thought: when the data center micro-grid is used as the load side of the power market, when the power market clearing model is built, the daily quote provided to the power market by the data center micro-grid based on the condition that the load can be reduced in the data center micro-grid is considered, so that the data center micro-grid actively participates in or even influences the clearing process of the power market, and meanwhile, a two-stage risk avoidance power market clearing model is built based on CVaR, so that the risk brought by uncertainty is reduced while the running cost of the data center micro-grid is optimized.
In the present invention, the main english abbreviations involved are as follows:
DR is DemandResponse, demand side response;
DATA CENTER Micro Grid, data center Micro Grid;
Energy Storage System, an electrical energy storage system;
CVaR: conditional value at risk, conditional risk value;
PSO Power System Operator, electric power system operator.
The following are examples.
A data center micro-grid operation optimization method based on demand side response is shown in fig. 1, and comprises the steps of respectively obtaining daily front quotations of a load side and a power generation side on a PSO side, wherein the daily front quotations of the load side comprise total load demands of all load providers in all time periods, load electric quantity and corresponding prices can be reduced, the daily front quotations of the power generation side comprise power generation plans of all units in all time periods and corresponding prices, and the load providers of the load side comprise DCMG.
In this embodiment, the daily price quotation provided by the load provider implements a DR price quotation mechanism based on load reduction, taking a data center micro-grid as an example, where the reducible load electric quantity and the corresponding price in each period are determined in the following manner:
firstly, taking the output condition of a new energy unit into consideration, in the embodiment, the new energy unit of the micro-grid of the data center mainly comprises a wind turbine unit and a photovoltaic unit, and the conservative estimation of the DCMG for the net load of each period The method comprises the following steps:
wherein, the Representing the total power load of the data center in a period t under a scene s; And In this embodiment, optionally, scene sampling is performed by a Monte Carlo sampling method, and multiple scenes and the probability of occurrence of each scene can be obtained by the scene sampling, so as to reduce the calculation burden (namely dimension disaster) in exponential growth caused by the increase of the number of the scenes, and scene subtraction is adopted to extract the scenes.
DCMG determines load shedding amountMaximum power generation capacity of self-contained generator set in t period, i.e
Wherein, the Representing a conservative estimate of the payload over a period t,The output of the traditional unit u in a period t-1 is represented; The upper power output limit of the conventional unit u is represented, and RU u represents the ascending ramp rate of the conventional unit u.
DCMG reporting corresponds toPrice of (2)The formula of the linear marginal cost of the unit u is as follows:
Where β u represents the fuel cost parameter of the conventional unit u.
Through the calculation, the daily quotation provided by the data center is a segmented load-reducible electric quantity-price pair, namelyThe load-side load-reducible amount and the corresponding price of each period of the load side are determined in the mode, and the load-side load-reducible amount accords with the load condition in the data center micro-grid and reflects the response of the demand side.
In consideration of the uncertainty of the load caused by the load-side load-reducible plan, the method further comprises the steps of establishing a two-stage power market clearing model based on CVaR and solving decision variables after acquiring the day-ahead quotation information of the power generation side and the load side, wherein the corresponding model is as follows:
Wherein F (x) is social benefit optimized in the first stage of decision x, namely the cost paid by load side electricity purchase, and the difference of income obtained by a power supply party, deltax represents decision adjustment quantity related to a random scene set xi, F (Deltax|xi) is decision social benefit desire under the scene set xi, ρ is risk avoidance factor in the range of [0,1 ], xine s represents the s-th scene in the scene set xi, p s represents occurrence probability of scene xi s, eta represents threshold value of social benefit desire value, Representing a real set, CVaR parameter measures a lower social benefit expectation in an uncertain environment.
The method comprises the steps of carrying out scene sampling in a second stage, solving the adjustment quantity of a first-stage decision variable in each scene, adjusting the first-stage decision variable to obtain a decision variable corresponding to each scene, multiplying the decision variable by the corresponding scene probability, and accumulating the decision variable to obtain a second-stage decision variable.
In this embodiment, in the two-stage power market clearing model, the objective function of the first stage is:
Wherein x represents a set of decision variables The method comprises the following steps of (1) setting a power generation side unit, wherein T represents the total number of scheduling time periods, I represents the index of the power generation side unit, i=1, 2, I and I represent the number of the power generation unit, j represents the index of price-electricity quantity of the power generation side unit to bidding sections, the bidding sections represent price measuring sections which can be bid by the power generation side unit in one time period, and each price measuring section comprises available electricity quantity and corresponding price; the power generation plan of each unit in each period of time on the power generation side after cleaning is shown, Respectively representing the power, standby and frequency-modulated planned output results of the jth bidding section of the generator set i in the period t,Respectively represents the electricity, standby and frequency modulation prices of the j-th bidding sections of the generator set i,The total number of the bidding sections of the power, standby and frequency modulation of the generator set i is respectively represented; For a 0/1 variable, 0 means that genset i does not provide a frequency modulation during period t, 1 means that genset i provides a frequency modulation during period t, n means the index of the load side load quotient, K means the index of the load side bid section, and K n means the total number of bid sections of load quotient n; Showing the load shedding schedule for each period on the load side after clearing, Shows the planned load non-reducible amount of load on the load side in the post-cleaning period t, P L shows the price of the load non-reducible amount of load on the load side,The load-reducible plan amount in the kth bid section of the load quotient n corresponding to the post-clearing period t is shown, the LP n,k shows the load-reducible price in the kth bid section of the load quotient n corresponding to the post-clearing period t, and in the first-stage objective function,Indicating the total price of the PSO wholesale to the utility, g t indicating the total cost of the PSO purchasing power, backup, and frequency modulation to the power generation side,I.e. represents social welfare;
The constraints of the first stage include:
R1-1, the total amount of the power, standby and frequency modulation plans of each generator set does not exceed the power generation capacity, and the related expression is:
Wherein O i represents the power generation capacity of the i-th generator set on the power generation side;
r1-2, the power, standby and frequency modulation plan output of each unit in each period do not exceed the upper limit and the lower limit of bidding electric quantity, and the related expression is:
wherein, the Respectively representing upper limits of the power, standby and frequency modulation power of the j-th bidding section of the generator set i;
r1-3, the generated energy of the power generation side is balanced with the load amount of the load side, and the related expression is:
r1-4, the load capacity of the load side does not exceed the upper limit and the lower limit of the bidding electric quantity, and the related expression is:
wherein, the The upper limit of the planned load reducible amount in the kth bid section of the load quotient n corresponding to the period t is represented, and SNC t represents the upper limit of the planned load non-reducible amount on the load side in the period t;
r1-5, frequency modulation constraint, specifically comprising (i) when each generator set provides frequency modulation, the power generation plan amount is more than the frequency modulation plan amount, and the difference value is not less than the predefined lower limit value (Ii) The sum of the power generation schedule amount and the frequency modulation schedule amount of each power generation set is smaller than a predefined upper limit valueThe expression is as follows:
r1-6, standby constraint, the expression is as follows:
Wherein lambda 1 represents a preset risk adjustment coefficient, lambda 2 represents a preset proportionality coefficient, the standby plan amount is enough to compensate the power loss caused by the faults of some generators unlike the power generation and frequency modulation constraint, and the standby plan amount can be further adjusted by the risk adjustment coefficient lambda 1 (set by PSO) to ensure a safer power supply, otherwise, the standby provided by each unit cannot exceed lambda 2 of the power generation capacity of each unit.
After the first stage is solved, decision variables which maximize social benefit under the condition of no uncertainty are obtained, wherein the decision variables comprise power generation plans of each unit in each period of the power generation side after the power generation side is clearedLoad shedding plan for each period on load side after clearing
The second-stage optimization decision is adjusted based on the scheduling decision of the first stage after the load uncertainty is realized based on the probability scene set, and the expected social benefit is maximized under the scene set xi. Introducing an auxiliary variable ζ s for each scene ζ s, the second stage risk avoidance market-clearing optimization objective can be converted into the following resolvable form:
Wherein S' represents the total number of scenes, S represents the scene index, ρ represents the risk avoidance factor, ρ E [0,1], η represents the threshold of the social benefit expected value, α represents the confidence level, ζ s represents the auxiliary variable in the S-th scene, p s represents the occurrence probability of the scene, and the decision variable set Representing the second stage adjustment of the first stage decision variable in the s-th scenario, Respectively, for the first stage decision variables in the s-th scene An adjustment amount of (2); For the second stage of adjusting the amount of the electric wholesale price adjustment caused, Representing the power generation cost bias due to the second stage decision adjustment.
The constraints of the second stage further include:
R2-1, after adjustment, the total amount of the power, standby and frequency modulation plans of each generator set does not exceed the power generation capacity, and the related expression is:
r2-2, after adjustment, the power, standby and frequency modulation plan output of each unit in each period do not exceed the upper limit and the lower limit of bidding electric quantity, and the related expression is:
R2-3, after adjustment, the generated energy of the power generation side is balanced with the load of the load side, and the related expression is:
r2-4, after being adjusted, the load capacity of the load side does not exceed the upper limit and the lower limit of the bidding electric quantity, and the related expression is as follows:
r2-5, frequency modulation constraint, specifically comprising (i) when each generator set provides frequency modulation after adjustment, the power generation plan amount is more than the frequency modulation plan amount, and the difference value of the two is not less than a predefined lower limit value (Ii) After adjustment, the sum of the power generation planned quantity and the frequency modulation planned quantity of each generator set is smaller than a predefined upper limit valueThe expression is as follows:
R2-6, standby constraint, the expression is as follows:
in the above constraint, and In order to improve the efficiency and precision of model solving, the embodiment linearizes the related nonlinear constraint before the model solving, and optionally, in the embodiment, a BigM method is specifically used for linearizing, namely relaxing the constraint, and introducing a maximum auxiliary variable Inf, wherein the relaxation of the frequency modulation constraint in the first stage is as follows:
the second phase of frequency modulation constraint relaxation is:
After the first-stage decision variables are adjusted through the model solving result of the second stage, the obtained second-stage decision variables can effectively reduce economic risks caused by load uncertainty.
Extracting load reduction plans of each period of the DCMG after clearing from the second-stage decision variables, and calculating the clearing electricity price corresponding to the second-stage decision variables;
aiming at minimizing the operation cost of the DCMG, carrying out daily scheduling optimization on the micro-grid of the data center according to the power price and the load reduction plan of each period of the DCMG after the power price is cleared, and obtaining the output plan of each device in the DCMG;
the objective function of the data center micro-grid for day-ahead scheduling optimization is as follows:
Wherein, alpha u represents the no-load cost of the traditional unit u; SU u represents the starting cost of the conventional unit u, SD u represents the shutdown cost of the conventional unit u, τ u,t and τ u,t-1 are both 0/1 variables, 1 represents that the conventional unit u is started in the corresponding period, and 0 represents that the conventional unit u is not started in the corresponding period; The clear electricity price of the period t is represented, and P t grid represents the electricity purchase quantity in the period t; the starting and stopping cost and the fuel cost of the traditional generator set are represented; namely, the electricity purchasing cost is represented;
The constraint conditions of the data center micro-grid for day-ahead scheduling optimization include:
Power balance constraint:
Force upper and lower limits and climbing rate constraint:
ESS operation constraints for electrical energy storage systems:
the data center is constrained by the capacity of the transmission line when the local area power grid is accessed to purchase power to the power grid:
Wherein P t disc represents the discharge capacity in the period t; For 0/1 variable, 1 means discharge in a period t, 0 means no discharge in the period t, ER t means new energy power generation in the period t, in this embodiment, ER t＝Pt wind+Pt PV,Pt wind and P t PV respectively mean wind turbine generator output and photovoltaic turbine generator output in the period t, P t DC means power load of a data center in the period t, and P t char means charging electric quantity in the period t; A variable of 0/1, a1 indicating charging during period t, and a0 indicating no charging during period t; And Respectively represent the lower limit and the upper limit of the electric power output of the conventional unit u, RD u and RU u represent the descending and ascending climbing powers of the conventional unit u,ES t and ES t+1 respectively represent the charging states of the electric energy storage system in the time period t and t+1, ES min and ES max respectively represent the minimum and maximum charging states of the electric energy storage system, and eta char and eta disc respectively represent the charging efficiency and the discharging efficiency of the electric energy storage system; And Respectively representing the maximum charging power and the maximum discharging power of the electric energy storage system; Representing the transmission capacity of the line;
The method and the device can obtain the output plans of the devices which minimize the running cost of the data center micro-grid after solving the running optimization model, and can effectively reduce the economic risk caused by load uncertainty while reducing the running cost of the data center micro-grid because the load-reducible plan and the power-off price according to the running optimization of the data center micro-grid are obtained by solving the CvaR-based two-stage power market-off model.
Example 2:
A computer readable storage medium includes a stored computer program, which when executed by a processor, controls a device in which the computer readable storage medium is located to execute the data center micro grid operation optimization method based on the demand side response provided in the foregoing embodiment 1.
The advantages achieved by the invention will be further explained below in connection with a specific application example.
It is assumed that the micro-grid of the data center is respectively composed of a generator set, a wind farm and a photovoltaic power station, and is simultaneously incorporated into the main grid. Technical parameters of each device in the data center power system are listed in table 1, economic parameters of each device in the data center power system are listed in table 2, IT loads related to the data center power loads are shown in table 3, wind speed data and photovoltaic power generation data are obtained by selecting meteorological data of the Hebei Zhangkou region from Renewables, ninja websites, a wind power plant is assumed to consist of 15 fans of 1.6MW, the installed capacity of the photovoltaic power plant is 40MW, and predicted power generation amounts of the wind power plant and the photovoltaic power plant are shown in fig. 2. The power market participants included 10 power generation companies on the power generation side, 3 load providers on the load side, and 1 data center operator. The power generation company quotation information is shown in table 4. PSO defines λ 1 =1.5 and λ 2 =0.6, the confidence α is 0.9, and the initial value of the risk avoidance factor ρ is 0.1. The total load of the system is predicted to generate a system load scene according to the load published by the EMA official website.
TABLE 1 list of technical parameters for devices in DCMG
TABLE 2 list of economic parameters for devices in DCMG
Table 3 data set of IT load
Table 4 electric market generation side bid information
Table 5 load reduction amount of electric power market load provider report information
Table 6 electric market load provider (DCMG) offer load cut price
Based on the quotation mechanism in the above embodiment 1, the daily quotation information provided by the load provider and the DCMG operator is shown in tables 5 and 6.
After acquiring daily quotes on the power generation side and the load side, the PSO performs market clearing based on CVaR risk avoidance two-stage DR electric power market clearing models. The price of the power market clearing is shown in figure 3, the price of the power and the price of the clearing power of the standby and frequency modulation auxiliary services determined by the market clearing model provided by the invention is higher in the 'supply-shortage' stage of the load peak period of 9-15 hours and 18-21 hours, and lower in the 'supply-shortage' stage of the load valley period of 2-7 hours, and the time-division price accords with the time-sequence load characteristic.
The planned load reduction of the DCMG after market has been cleared is shown in fig. 4, in which the upper column on the abscissa represents the load reduction before the DCMG day, and the lower column on the abscissa represents the planned load reduction of the DCMG after market has been cleared. It can be seen that, in the time period with lower clearing price, such as 1-7 hours and 24 hours, the planned load reduction amount of the DCMG after clearing is smaller than the reducible load electric quantity in the prior quotation, that is, the DCMG tends to purchase electricity in the electric power market, and in the time period with higher clearing price, the planned load reduction amount of the DCMG is more, that is, the DCMG avoids purchasing electricity in the electric power market as much as possible, so that the effectiveness of the response mechanism at the demand side is reflected.
TABLE 7 comparison of objective functions under different models
Table 7 shows the objective function comparisons for different models (not CVaR-based model and CVaR-based model), compared to the non-CVaR-based optimization model, the conditional risk value (CVaR) in CVaR-based two-stage model optimization results was 428.8$ higher, i.e., social Functions in the partial probability scenario with lower objective function values were optimized, while risk optimization resulted in a slight decrease in the overall social benefit expectations, i.e., a decrease in social benefit expectations by 8.0$. Therefore, by introducing CVaR into the two-stage market clearing model, risks can be effectively inhibited, less social benefits are sacrificed, more robust decisions are expected to be made, and social benefits in a poor probability scene are improved.
As shown in fig. 5, as the risk avoidance factor ρ increases, CVaR, the social benefit in a worse probability scene gradually increases, while the overall social benefit expectation on the ordinate decreases. The risk avoidance market clearing model based on CVaR provided by the invention is shown, the economic risk brought by system uncertainty can be effectively reduced by adopting CVaR factors related to social welfare, and the degree of risk avoidance can be controlled by the magnitude of a risk avoidance factor rho.
After the DR-based power market is cleared to obtain the time-sharing price, the DCMG day-ahead operation is optimally scheduled based on the prediction information, and a fixed electricity price (the price is the time-sharing electricity price average value) is adopted as a comparison case to carry out a simulation experiment. The scheduling results of ESS in the fixed and time-of-use power rate contexts are shown in (a) and (b) in fig. 6, and the scheduling results of each unit in the DCMG in the fixed and time-of-use power rate contexts are shown in (a) and (b) in fig. 7, respectively, and the running cost of the DCMG in the fixed and time-of-use power rate contexts is 24.71k$, and the running cost of the DCMG in the DR power market context is 24.15k$, respectively, which means that the participating power market is beneficial to the optimization of the running cost of the DCMG.
It will be readily appreciated by those skilled in the art that the foregoing description is merely a preferred embodiment of the invention and is not intended to limit the invention, but any modifications, equivalents, improvements or alternatives falling within the spirit and principles of the invention are intended to be included within the scope of the invention.
Claims (9)
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211436226.7A CN115700655B (en) | 2022-11-16 | 2022-11-16 | Data center micro-grid operation optimization method based on demand side response |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211436226.7A CN115700655B (en) | 2022-11-16 | 2022-11-16 | Data center micro-grid operation optimization method based on demand side response |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115700655A CN115700655A (en) | 2023-02-07 |
| CN115700655B true CN115700655B (en) | 2025-07-18 |
Family
ID=85121169
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211436226.7A Active CN115700655B (en) | 2022-11-16 | 2022-11-16 | Data center micro-grid operation optimization method based on demand side response |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115700655B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN119850247B (en) * | 2024-12-24 | 2025-10-17 | 华中科技大学 | Electric power transaction method, device and system based on Markov optimization |
| CN119624077B (en) * | 2025-02-17 | 2025-05-13 | 合肥工业大学 | Electrical carbon coupling market multi-stage clearing method based on user carbon emission intensity |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2019196375A1 (en) * | 2018-04-13 | 2019-10-17 | 华南理工大学 | Demand side response-based microgrid optimal unit and time-of-use electricity price optimization method |
| CN112529622A (en) * | 2020-12-08 | 2021-03-19 | 国网河南省电力公司经济技术研究院 | Virtual power plant-based method for clearing multiple micro-micro main bodies participating in spot market |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2015525057A (en) * | 2012-08-16 | 2015-08-27 | ロベルト・ボッシュ・ゲゼルシャフト・ミト・ベシュレンクテル・ハフツングＲｏｂｅｒｔ Ｂｏｓｃｈ Ｇｍｂｈ | DC building system with energy storage and control system |
| CN114565153A (en) * | 2022-02-24 | 2022-05-31 | 中冶赛迪工程技术股份有限公司 | Regional wind, light and vehicle combination optimization method based on dispatchable capacity of electric vehicle |
-
2022
- 2022-11-16 CN CN202211436226.7A patent/CN115700655B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2019196375A1 (en) * | 2018-04-13 | 2019-10-17 | 华南理工大学 | Demand side response-based microgrid optimal unit and time-of-use electricity price optimization method |
| CN112529622A (en) * | 2020-12-08 | 2021-03-19 | 国网河南省电力公司经济技术研究院 | Virtual power plant-based method for clearing multiple micro-micro main bodies participating in spot market |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115700655A (en) | 2023-02-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Babatunde et al. | Power system flexibility: A review | |
| Nojavan et al. | Optimal bidding and offering strategies of merchant compressed air energy storage in deregulated electricity market using robust optimization approach | |
| O’Dwyer et al. | Efficient large-scale energy storage dispatch: challenges in future high renewable systems | |
| Fares et al. | A dynamic model-based estimate of the value of a vanadium redox flow battery for frequency regulation in Texas | |
| CN115688970B (en) | Micro-grid two-stage self-adaptive robust optimal scheduling method based on interval probability uncertainty set | |
| CN111200281B (en) | Optimization method for energy storage configuration expansion of internet microgrid | |
| CN115700655B (en) | Data center micro-grid operation optimization method based on demand side response | |
| CN112950098A (en) | Energy planning method and device based on comprehensive energy system and terminal equipment | |
| Tian et al. | Coordinated RES and ESS planning framework considering financial incentives within centralized electricity market | |
| Jin et al. | An overview of virtual power plant development from the perspective of market participation | |
| Lu et al. | Day‐Ahead Scheduling for Renewable Energy Generation Systems considering Concentrating Solar Power Plants | |
| Wang et al. | A two-stage optimization strategy for VPP trading in multi-market considering setting method and marginal revenue and expenditure of standby capacity | |
| Pruckner et al. | A study on the impact of packet loss and latency on real-time demand response in smart grid | |
| Shrestha et al. | Impact of market-driven energy storage system operation on the operational adequacy of wind integrated power systems | |
| CN120579781A (en) | Planning method for integrated electricity-hydrogen energy system considering multi-timescale uncertainties | |
| CN113780742A (en) | Computing method for flexibility improvement economy of power generation unit in auxiliary service market environment | |
| Huang et al. | Analysis of optimal configuration of energy storage in wind-solar micro-grid based on improved gray wolf optimization | |
| CN109728609B (en) | A dispatch method and system for abandoned wind power | |
| CN111160767A (en) | A comprehensive energy service benefit assessment method | |
| Guo et al. | Robust based optimal operation model of virtual power plant in electricity market | |
| CN116681294A (en) | Adjustable Risk Preference Integrated Energy System Capacity Control Method Using Improved WCVaR | |
| Xuan et al. | Unit commitment in microgrid considering customer satisfaction in incentives-based demand response program: a fuzzy logic model | |
| Wang et al. | Study on the economic dispatch of regional integrated energy system based on master-slave game | |
| Li et al. | Scheduling model of power system with renewable energy and transaction mode of direct electricity purchase by large consumers considering network security constraints | |
| Zhang et al. | Extended LMP under high-penetration wind power |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |