CN108898282A - Data center resource Optimization Scheduling and computer storage medium - Google Patents
Data center resource Optimization Scheduling and computer storage medium Download PDFInfo
- Publication number
- CN108898282A CN108898282A CN201810573669.8A CN201810573669A CN108898282A CN 108898282 A CN108898282 A CN 108898282A CN 201810573669 A CN201810573669 A CN 201810573669A CN 108898282 A CN108898282 A CN 108898282A
- Authority
- CN
- China
- Prior art keywords
- data center
- time period
- power
- carbon emission
- conventional rack
- 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.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
- G06Q10/06312—Adjustment or analysis of established resource schedule, e.g. resource or task levelling, or dynamic rescheduling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/06—Electricity, gas or water supply
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Abstract
The present invention provides a kind of data center resource Optimization Scheduling and computer storage medium, wherein method includes:Different types of load capacity is predicted according to the history workload data of data center;Obtain the first electric energy supply of renewable energy, the second electric energy supply of energy-storage system, Day-ahead electricity market electric rate and power grid carbon emission amount;The load distribution of each period and the power supply composition of the data center are determined according to the load capacity, the first electric energy supply, the second electric energy supply, the electric rate and the power grid carbon emission amount.The present invention can be realized the joint optimal operation of economic operation cost and carbon emission amount, and provide theoretical foundation and guidance for the operation of data center's actual operation.
Description
Technical field
The present invention relates to the load of energy source optimization configuration technology field more particularly to data center and power consumption management method,
It is specifically exactly a kind of data center resource Optimization Scheduling and computer storage medium.
Background technique
With the continuous development of Internet technology, the quantity of global data center and scale are in that explosion type increases in recent years
It is long.The data center that size is stood in great numbers also consumes a large amount of electric energy and generates the huge electricity charge while providing service for people,
It improves the efficiency of data center and reduces electric cost expenditure and have become the problem of data center operator is most paid close attention to.2011 years
The 1.5% of global electric quantity consumption has been accounted for according to center electricity consumption, according to measuring and calculating, the whole world will be accounted for the year two thousand twenty data center electricity consumption
The 8% of electric quantity consumption.The data center in China is also growing rapidly, and total amount is more than 400,000 at present.2016, China
Analyzing Total Electricity Consumption is 5,919,800,000,000 degree, and wherein up to 88,800,000,000 degree, the ratio for accounting for social total electricity consumption is more than data center's electricity consumption
1.5%, the electricity charge have become the maximum expenditure of data center's operation.
On the other hand, carbon emission problem caused by data center's electricity consumption also can not be ignored.Data center's carbon emission at present
Amount has accounted for the 2% of global carbon emission, is equivalent to the carbon emission amount that entire aircraft industry generates.According to measuring and calculating, to the year two thousand thirty whole world
Data center will generate 340 tons of carbon emission amounts every year.For this purpose, some large-scale data center suppliers, such as Google, apple, face
Book etc. has closely begun to use renewable energy for data center's power supply.Wherein, Google is maximum renewable energy purchase in the world
Enterprise is bought, and realized the supply of 100% renewable energy of data center in 2017.However wind-powered electricity generation, photovoltaic electric energy etc. are renewable
Energy resource supply often has the characteristics that unstability, intermittence and time variation, but data center needs reliable, stable electric energy
Supply, this demand side for making renewable energy power generation whether can satisfy data center are faced with lot of challenges.
Currently, having there are many those skilled in the art to propose various optimizations in terms of reducing data center's operating cost
Method but is seldom related to above-mentioned including adjusting the switch state of server, using energy storage device or renewable energy etc.
The combination of a variety of methods.And goal in research before is mostly that cost is minimum or profit is maximum, ignores data center's electricity consumption
The environmental problem of generation.Traditional data center's method for scheduling task is divided into two classes:For the tune of interactive workload
Degree or the scheduling for batch processing workload, but the data center in reality will often provide a variety of services, place parallel
A variety of workloads are managed, and the load treating capacity of data center is directly related to power consumption.
Therefore, those skilled in the art need to research and develop it is a kind of handle how many loads in different moments, with which kind of power supply to number
According to the priority scheduling of resource method of center feed, on the basis of reducing data center's electric cost, carbon emission amount is reduced.
Summary of the invention
In view of this, the technical problem to be solved in the present invention is that provide a kind of data center resource Optimization Scheduling and
Computer storage medium, solve in the prior art not and meanwhile consider data center a variety of power supplies and a variety of work
Load, leads to the problem that data center's total operating cost is excessively high or carbon emission amount is excessive.
In order to solve the above-mentioned technical problem, a specific embodiment of the invention provides a kind of data center resource Optimized Operation
Method, including:Different types of load capacity and power grid carbon emission amount are predicted according to the history workload data of data center;
Obtain the first electric energy supply of renewable energy, the second electric energy supply of energy-storage system, Day-ahead electricity market electricity charge valence
Lattice;According to the load capacity, the first electric energy supply, the second electric energy supply, the electric rate and the electricity
Net carbon emission amount determines the load distribution of each period and the power supply composition of the data center.
A specific embodiment of the invention also provides a kind of computer storage medium comprising computer executed instructions, described
When computer executed instructions are handled via data processing equipment, which executes data center resource Optimized Operation
Method.
Above-mentioned specific embodiment according to the present invention is it is found that data center resource Optimization Scheduling and computer storage
Medium at least has the advantages that:Pass through the history workload data prediction directly processing load (interaction of data center
Formula load) and can postpone load (batch processing load) load capacity, finally using profit maximum or/and environmental impact minimization as mesh
Mark arranges the processing of batch processing load, so that data center reaches while scheduling power supply handles Interactive workload in real time
The solution of minimum or both the tradeoff of, environmental pollution minimum to economic cost;Realize economic operation cost and carbon emission amount
Joint optimal operation, may be implemented that its economic cost is minimum or the smallest operation target of environmental pollution, and can flexible modulation
Two-part weight has very high perspective, innovative;By adjusting from the traditional machine of the purchase of electricity of power grid, conventional power source
Optimization aim is realized in group combination, energy storage charge and discharge, the distribution flexibly loaded, considers that content is comprehensive, compatibility is strong;It and is number
Theoretical foundation and guidance are provided according to the operation of center actual operation.
It is to be understood that above-mentioned general description and following specific embodiments are merely illustrative and illustrative, not
The range of the invention to be advocated can be limited.
Detailed description of the invention
Following appended attached drawing is part of specification of the invention, depicts example embodiments of the present invention, institute
Attached drawing is used to illustrate the principle of the present invention together with the description of specification.
Fig. 1 is a kind of embodiment one for data center resource Optimization Debugging method that the specific embodiment of the invention provides
Flow chart.
Fig. 2 is a kind of embodiment two for data center resource Optimization Debugging method that the specific embodiment of the invention provides
Flow chart.
Fig. 3 is a kind of embodiment three for data center resource Optimization Debugging method that the specific embodiment of the invention provides
Flow chart.
Fig. 4 be the specific embodiment of the invention provide a kind of powered by multiple power sources and handled in the data of a variety of loads
Hearty cord composition.
The charge and discharge of Fig. 5 is a kind of carbon emission weight γ that the specific embodiment of the invention provides when being 0.0025 energy-storage system
Electrical schematic.
Batch processing load point when being 0.0025 that Fig. 6 is a kind of carbon emission weight γ that the specific embodiment of the invention provides
Match.
The power generation for conventional rack that Fig. 7 is a kind of carbon emission weight γ that the specific embodiment of the invention provides when being 0.0025
Plan.
Fig. 8 is a kind of carbon emission weight γ that the specific embodiment of the invention provides when being 0.0025, and data center totally provides
Source optimization schematic diagram.
Fig. 9 is a kind of carbon emission weight γ that the specific embodiment of the invention provides when being 80, and the charge and discharge of energy-storage system is given instructions by telegraph
It is intended to.
Batch processing load distribution when being 80 that Figure 10 is a kind of carbon emission weight γ that the specific embodiment of the invention provides.
The power generation meter for conventional rack that Figure 11 is a kind of carbon emission weight γ that the specific embodiment of the invention provides when being 80
It draws.
Figure 12 is a kind of carbon emission weight γ that the specific embodiment of the invention provides data center's aggregate resource when being 80
Optimize schematic diagram.
Specific embodiment
Understand in order to make the object, technical scheme and advantages of the embodiment of the invention clearer, below will with attached drawing and in detail
Narration clearly illustrates the spirit of disclosed content, and any skilled artisan is understanding the content of present invention
Embodiment after, when the technology that can be taught by the content of present invention, be changed and modify, without departing from the content of present invention
Spirit and scope.
The illustrative embodiments of the present invention and their descriptions are used to explain the present invention, but not as a limitation of the invention.
In addition, in the drawings and embodiments the use of element/component of same or like label is for representing same or like portion
Point.
About " first " used herein, " second " ... etc., not especially censure the meaning of order or cis-position,
It is non-to limit the present invention, only for distinguish with same technique term description element or operation.
About direction term used herein, such as:Upper and lower, left and right, front or rear etc. are only with reference to attached drawing
Direction.Therefore, the direction term used is intended to be illustrative and not intended to limit this creation.
It is open term, i.e., about "comprising" used herein, " comprising ", " having ", " containing " etc.
Mean including but not limited to.
About it is used herein " and/or ", including any of the things or all combination.
It include " two " and " two or more " about " multiple " herein;It include " two groups " about " multiple groups " herein
And " more than two ".
About term used herein " substantially ", " about " etc., to modify it is any can be with the quantity or mistake of microvariations
Difference, but this slight variations or error can't change its essence.In general, microvariations or error that such term is modified
Range in some embodiments can be 20%, in some embodiments can be 10%, in some embodiments can for 5% or
It is other numerical value.It will be understood by those skilled in the art that the aforementioned numerical value referred to can be adjusted according to actual demand, not with this
It is limited.
It is certain to describe the word of the application by lower or discuss in the other places of this specification, to provide art technology
Personnel's guidance additional in relation to the description of the present application.
Fig. 1 is a kind of embodiment one for data center resource Optimization Debugging method that the specific embodiment of the invention provides
Flow chart, as shown in Figure 1, according in the certain time period of history workload data prediction data center in the near future
Power consumption is loaded, the renewable energy that data center can obtain, the energy and Day-ahead electricity market of energy-storage system storage are obtained
Electric rate, finally determine the power supply composition (i.e. power supply mode) and each period of data center based on these data
Load distribution.
In the specific embodiment shown in the drawings, data center resource Optimization Debugging method includes:
Step 101:Different types of load capacity is predicted according to the history workload data of data center.Reality of the invention
Apply in example, load type include data center in different time period power load type (including Interactive workload and batch at
Reason load), different types of load capacity refers specifically to Interactive workload and batch processing loads data volume to be treated.History work
Make load data i.e. historic load, is to predict soon the important references of (such as second day) variety classes load capacity in the future.
Step 102:Obtain the first electric energy supply of renewable energy, the second electric energy supply of energy-storage system, a few days ago
The electric rate and power grid carbon emission amount of electricity market.In the embodiment of the present invention, renewable energy includes wind energy, photovoltaic energy
With the electric energy of the generations such as water potential energy;Day-ahead electricity market can refer to the electric rate in same day different time sections.Energy-storage system exists
Extensive application in data center, frequently as uninterruptible power supply (UPS) after main power supply interruption, before backup power source starting
10 seconds~20 seconds use, the Capacity design of UPS generally can for data center run 5-30 minutes.Power grid carbon emission amount refers to
The carbon generated when the formation of the carbon emission amount that data center generates from power grid power purchase, i.e. power grid electric energy.Therefore ensuring power supply safety
Under the premise of, it can use electricity price and the fluctuation of power grid carbon emission amount at any time be electrically operated to energy-storage system progress charge and discharge, from
And the electricity charge are optimized.Specifically, in the case where mainly considering electricity price, energy-storage system should charge when electricity price is low,
It powers when electricity price is high for data center;In the case where mainly considering carbon emission, energy-storage system should be when power grid carbon emission amount be low
Charging is load supplying when carbon emission amount is high.
Step 103:According to the load capacity, the first electric energy supply, the second electric energy supply, the electricity charge
Price and the power grid carbon emission amount determine the load distribution of each period and the power supply composition of the data center.Examine
The power supply mode for considering loadtype and data center determines in the near future as the power supply of data center's power supply.
Referring to Fig. 1, based on different types of load capacity by adjusting from the conventional rack of the purchase of electricity of power grid, conventional power source
Data center resource optimization aim is realized in combination, energy storage charge and discharge, the distribution flexibly loaded, realizes economic operation cost and carbon
The joint optimal operation of discharge amount considers that content is comprehensive, and compatibility is strong.
Fig. 2 is a kind of embodiment two for data center resource Optimization Debugging method that the specific embodiment of the invention provides
Flow chart, as shown in Figure 2, it is contemplated that the power supply of data center is varied, and the workload of data center is also more
Kind multiplicity needs the operation of workload, energy-storage system, the operation of conventional rack and power balance to data center to carry out
Constraint.
In the specific embodiment shown in the drawings, before step 101, data center resource Optimization Debugging method is also wrapped
It includes:
Step 100:Configure the constraint condition of the data center.In the embodiment of the present invention, the constraint condition includes
Workload constraint, energy-storage system operation constraint, conventional rack operation constraint and power balance constraint.
Referring to fig. 2, the constraint condition of data center is set, power supply caused by preventing data center's power resource from optimizing
Risk ensures the power supply safety of data center, while guaranteeing power resource optimization, does not influence the normal operation of data center.
Specifically workload constraint, energy-storage system operation constraint, conventional rack operation constraint and power balance constraint:
(1) workload constrains
The workload of data center can be according to there are many its service content, can be by it from the angle of Optimized Operation
It can be mainly divided into two classes:Interactive workload (Interactive workload) and batch processing workload (batch processing load).
Interactive workload is also known as delay-sensitive load, i.e. the workload of the type once reaches data center, must
Must be handled immediately, the reaction time of permission generally within a few seconds, such as user's Real time request, network service.It should
Class load does not have controllability, it is therefore necessary to meet electricity consumption required for this fractional load.
Batch processing load is also known as delay-tolerant load, it is characterized in that permitting after the type workload reaches data center
Perhaps the processing time is more flexible, as long as completing before deadline, even can postpone a few hours sometimes.It is general next
It says, batch processing load is usually made of computation-intensive task, such as data mining, financial analysis, image procossing work.This
One characteristic provides possibility for that can be transferred into the period processing that power grid electricity price is lower or carbon emission amount is less, is also number
Greatly optimization space is provided according to the load side management at center.
Assuming that data center's interactive mode workload has i kind, batch processing load has j kind, and every kind of load respectively corresponds specific
User demand.WithThe quantity of two kinds of workloads is respectively represented, then load capacity w total in the t periodt,sFor:
Accessible load capacity no more than data center IT service ability, i.e.,:
Wherein, BITFor data center's load capacity upper limit.
(2) energy-storage system operation constraint
Energy-storage system extensive application in the data center, frequently as UPS after main power supply interruption, backup power source opens
It uses within 10 seconds~20 seconds before dynamic, the Capacity design of UPS can generally be run 5 minutes~30 minutes for data center.Therefore exist
Under the premise of ensuring Supply Security, the fluctuation of the fluctuation and data central loading of electricity price at any time can use to energy storage
System progress charge and discharge is electrically operated, to optimize to the electricity charge.
In the case where mainly considering electricity price, energy storage should charge when electricity price is low, power when electricity price is high for data center;
It in the case where mainly considering carbon emission, should charge when power grid carbon emission amount is low, be load supplying when carbon emission amount is high.
Following thermal energy storage process formula for energy-storage system:
The formula illustrates that the thermal energy storage process of energy-storage system, the energy state of next period (t+1) are equal to upper a period of time
Between section (t-1) energy state plus this period (t) charge volume, subtract the discharge capacity of this period (t).Wherein, EStTable
Show energy state of the energy-storage system within the t period, due to having kwh loss, η in energy-storage system charge and discharge processcAnd ηdPoint
Not Wei charge efficiency and discharging efficiency, Pt charAnd Pt discharFor the charge volume and discharge capacity in the period, ESt+1For energy storage system
Energy state of the system in subsequent time period.
Wherein, ESmaxThe maximum value of electricity, ES are stored for energy-storage systemminThe minimum value of electricity is stored for energy-storage system.
Due to energy-storage system charge and discharge process cannot carry out simultaneously, herein introduce two 0-1 variablesWith When being 1, illustrate that current slot energy-storage system is charging,When being 0, illustrate current slot energy storage system
System is not charging;SimilarlyWhen being 1, show that current slot energy-storage system is discharging,When being 0, when current
Between section energy-storage system do not discharging.
In the charge power P of current slot energy-storage systemt charIt is certain to be more than or equal to 0, and it is less than the charge power upper limitThe discharge power P of current slot energy-storage systemt discharIt is also at 0 and the discharge power upper limitBetween.
(3) conventional electric power generation unit operation constraint
It runs without interruption within 24 hours to meet data center, guarantee service quality, there is pole in data center to power supply reliability
High requirement, therefore must be equipped with can satisfy the stable generating set of contributing of its power demand, such as coal-fired, combustion in microgrid
The conventional racks such as gas, diesel-driven generator.The operation constraint of conventional rack has:
Wherein, formula shows the generated energy P of all conventional racks in current slott unitEqual to each unit generation amountThe sum of.
Wherein,For conventional rack l time period t the lower limit of the power,For conventional rack l time period t function
The rate upper limit, ot,lIndicate whether conventional rack l is in operating status in time period t.
-ot-1,l+ot,l-ok,l≤0
Above formula indicates that conventional rack most short opens the machine time what Qi Jishi should be kept, wherein ot,lAnd ok,lIt respectively indicates
Operating status of the unit l at t the and k moment (when value is 1, shows that conventional rack is currently running, when value is 0, shows tradition
Unit is not in operating status), MUlMost short for conventional rack l opens the machine time.
-ot-1,l-ot,l+ok,l≤1
Above formula indicates the Minimum Idle Time that conventional rack needs, wherein ot,lAnd ok,lRespectively indicate conventional rack l when
Between section t and period k operating status (value be 1 when, show that conventional rack is currently running, value be 0 when, show traditional machine
Group is not in operating status), MDlFor the Minimum Idle Time of conventional rack l.
Above formula illustrates the condition that conventional rack l need to meet in time period t booting, wherein ut,lIndicate that conventional rack l exists
Whether time period t is in open state.If in the upper period t-1, o of time period tt-1,lIt is 0, current slot t, ot,lFor
1, then ut,lIt is 1, as open state.
Above formula illustrates that conventional rack l shuts down the condition that need to meet in time period t, wherein vt,lIndicate that conventional rack l exists
Whether time period t is in shutdown status.If in the upper period t-1, o of time period tt-1,lIt is 1, current slot t, ot,lFor
0, then vt,lIt is 1, as shutdown status.
Above formula illustrates power constraint of conventional rack l during climbing.Wherein, RUlFor the climbing function of conventional rack l
Rate.
Above formula illustrates power constraint of conventional rack l during descending.Wherein, RDlFor the descending function of conventional rack l
Rate.
(4) power balance constrains
Data center usually runs in the environment of micro-capacitance sensor, and conventional power source (fire coal, combustion gas, bavin are generally comprised in microgrid
Oil machine group etc.), energy storage device and renewable energy (such as wind-powered electricity generation, photovoltaic).As shown in figure 4, the operation of data center's micro-capacitance sensor
Quotient by electricity market purchase of electricity, local power supply generated energy and data center's batch processing load allocation plan into
Row combined optimization can make reliable, economic and environmental-friendly resource allocation proposal.
For entire data center, following power balance condition should be met:
Power supply is made of several parts:Electricity P is bought from power gridt grid, the generated energy P of conventional rackt unit, Wind turbines
Generated energy Pt wind, the generated energy P of photovoltaict PVAnd the charge volume P of energy storaget dischar, electricity needs part is first is that in data
Heart electricity consumption Qt, second is that the charge volume P of energy storaget char.It must assure that Real-time Balancing between power generation and electricity consumption.
Wherein, the generated energy P of wind-powered electricity generation and photovoltaict windAnd Pt PVFor the numerical value obtained according to short-term processing prediction.Data center
Electricity consumption QtCalculation method it is as follows:
Wherein, M is the total server number of data center, wtFor total load amount at this time.
In formula, ε, θ is calculated by following formula:
ε=Pidle+(PUE-1)Ppeak
θ=Ppeak-Pidle
Wherein, PpeakFor the power consumption of individual server full load, PidlePower consumption when for the individual server free time, PUE
(power usage effectiveness) is data center's energy consumption efficiency, by the total energy consumption and information technoloy equipment energy consumption of data center
Ratio obtain;ε and θ is the auxiliary variable in calculating process.
Fig. 3 is a kind of embodiment three for data center resource Optimization Debugging method that the specific embodiment of the invention provides
Flow chart, as shown in figure 3, total operating cost and total carbon emissions amount are calculated based on the power supply for being determined as data center's power supply, so that
Data center operator can be realized that economic cost is minimum, minimum or both the solution taken into account of environmental pollution.
In the specific embodiment shown in the drawings, after step 103, data center resource Optimization Debugging method is also wrapped
It includes:
Step 104:The total operating cost and total carbon emissions amount of the data center are calculated according to the power supply.This hair
In bright embodiment, total operating cost includes cost and conventional rack (conventional electric power generation unit) of the data center from power grid power purchase
Operating cost;Renewable energy and the operating cost of energy-storage system are very small, can be ignored.Total carbon emissions amount includes
The carbon emission amount that the conventional rack operation for the carbon emission amount and data center that data center generates from power grid power purchase generates;It can be again
The raw energy and energy-storage system will not consume fossil energy, belong to clean energy resource, carbon emission amount 0.
Referring to Fig. 3, the economic cost that data center operator can be realized data center is minimum or environmental pollution is minimum
Or both the solution taken into account.
Further, the expression Total of the total operating cost of the data center and total carbon emissions amount is:
Wherein, t is the period;T is the set of time period t;For total operating cost;For
Total carbon emissions amount;γFor the weighted value of total carbon emissions amount;It is data center in time period t from the cost of power grid power purchase;
For the operation cost of the conventional rack of data center in time period t;It is produced for data center in time period t from power grid power purchase
Raw carbon emission amount;The carbon emission amount generated in conventional rack battalion row for data center in time period t.According to γ
The weighted value of preset carbon emission amount, if only considering, economic cost is minimum, is indifferent to carbon emission, then γ can take 0, more
Concern for the environment pollution, γ value are bigger.
Further, cost of the data center from power grid power purchaseSpecific formula be:
Wherein, t is the period;Pt gridThe electricity bought in time period t from power grid for data center;For in the time
The electricity price of electricity market in section t.Since data center's electricity consumption is huge, neighbouring power supply facilities can not meet number steadily in the long term
According to the power demand at center, therefore, data center needs to buy big volume and electricity from power grid.But electricity price is with Supply and demand trend and line
Road congestion situations real-time fluctuations, therefore the cost of power purchase also has uncertainty from power grid.
Further, the operation cost of the conventional rack of the data centerSpecific formula be:
Wherein, t is the period;T is the set of time period t;L is conventional rack number;L is conventional rack quantity;SUlFor
Conventional rack l opens machine cost time period t;SDlIndicate conventional rack l in the shutdown cost of time period t;OlIndicate traditional machine
Idle running cost of the group l in time period t under no-load condition;ClFor operation of the conventional rack l in time period t at
This; ut,lWhether respectively conventional rack l is in time period t in 0,1 variable that is switched on;vt,lIt is conventional rack l in time period t
0,1 variable whether shut down;ot,lFor conventional rack l whether in 0,1 unloaded variable in time period t;For when
Between in section t conventional rack generated energy.Conventional rack l only can be under a kind of state in time period t.The operation of conventional rack
Cost mainly includes start-up and shut-down costs, operating cost and unloaded cost (idle running cost).Since conventional rack is from shutting down to opening
It is dynamic, or from run to shut down during will consume a large amount of fuel, especially fired power generating unit open machine, shutdown time generally compared with
Long, start-up and shut-down costs are very important.Operating cost refers to the incremental cost generated with actual power, and fired power generating unit is come
Say mainly fuel cost.In addition if in the case that generating set is in no-load running, it can also generate certain cost and disappear
Consumption, which is known as unloaded cost.
Further, carbon emission amount (i.e. marginal carbon emission amount, the every hair 1,000 that the data center generates from power grid power purchase
The carbon emission amount that watt-hour electricity generates)Specific formula be:
Wherein, t is the period;K is fuel type sequence;N is fuel type total amount;Pt gridIt is data center in the period
The electricity bought in t from power grid;ekTo indicate carbon emission rate of the kth kind fuel in power generation;gk,tTo indicate kth kind fuel
Accounting of the marginal unit in time period t.
Further, the carbon emission amount generated in the conventional rack operation of the data centerSpecific formula be:
Wherein, t is the period;L is conventional rack number;L is conventional rack quantity;elIt is being sent out for l platform conventional rack
Carbon emission rate when electric;For the generated energy of l platform conventional rack.
A specific embodiment of the invention provides a kind of computer storage medium comprising computer executed instructions, the meter
Calculation machine executes instruction when handling via data processing equipment, which executes data center resource Optimization Debugging side
Method.Method includes the following steps:
Step 101:Different types of load capacity is predicted according to the history workload data of data center.
Step 102:Obtain the first electric energy supply of renewable energy, the second electric energy supply of energy-storage system, a few days ago
The electric rate and power grid carbon emission amount of electricity market.
Step 103:According to the load capacity, the first electric energy supply, the second electric energy supply, the electricity charge
Price and the power grid carbon emission amount determine the load distribution of each period and the power supply composition of the data center.
Another embodiment of the present invention provides a kind of computer storage medium comprising computer executed instructions, institute
When stating computer executed instructions and handling via data processing equipment, which executes data center resource optimization and adjusts
Method for testing.Method includes the following steps:
Step 100:Configure the constraint condition of the data center.
Step 101:Different types of load capacity is predicted according to the history workload data of data center.
Step 102:Obtain the first electric energy supply of renewable energy, the second electric energy supply of energy-storage system and a few days ago
The electric rate and power grid carbon emission amount of electricity market.
Step 103:According to the load capacity, the first electric energy supply, the second electric energy supply, the electricity charge
Price and the power grid carbon emission amount determine the load distribution of each period and the power supply composition of the data center.
Another embodiment of the present invention provides a kind of computer storage medium comprising computer executed instructions, institute
When stating computer executed instructions and handling via data processing equipment, which executes data center resource optimization and adjusts
Method for testing.Method includes the following steps:
Step 101:Different types of load capacity is predicted according to the history workload data of data center.
Step 102:Obtain the first electric energy supply of renewable energy, the second electric energy supply of energy-storage system, a few days ago
The electric rate and power grid carbon emission amount of electricity market.
Step 103:According to the load capacity, the first electric energy supply, the second electric energy supply, the electricity charge
Price and the power grid carbon emission amount determine the load distribution of each period and the power supply composition of the data center.
Step 104:The total operating cost and total carbon emissions amount of the data center are calculated according to the power supply.
In implementation embodiment of the invention, it is assumed that there are three conventional rack, (No. 1 and No. 2 units are to fire altogether for data center
Mechanism of qi group, No. 3 are coal unit), but also it is distributed formula wind-powered electricity generation and photovoltaic unit and energy-storage system.Following table 1 is number
According to the parameter at center.
Table 1
Data center's related data is as shown in table 1, and the maximum electric power of data center's power consumption is 50MW, number of servers M
It is 2.5 × 105, power P that when individual server free time consumesidleFor 100MW, the power of individual server full load consumption
PpeakValue for 200MW, data center's energy consumption efficiency (PUE) is set as 1.2.
Following table 2 is the parameter of energy-storage system.
Table 2
The design parameter of energy-storage system is according to set by document, as shown in table 2.Wherein energy-storage system maximum energy storage electricity is
30MWh, minimum energy storage electricity are 5MWh, and initial quantity of electricity and final electricity are 5MWh, and maximum charge and discharge power is 5MW/h,
Efficiency for charge-discharge is 0.8.
Following Table 3 is the parameter of conventional rack.
Table 3
The relevant parameter of conventional electric power generation unit is listed in table 3, wherein conventional rack 1,2 is jet dynamic control, is passed
System unit 3 is Thermal generation unit, and the constraint condition limitation of each conventional rack and cost are all from real data.
Following table 4 is the carbon emission rate of all kinds of fuel units.
Table 4
Fuel type | It is coal-fired | Combustion gas | Wind-powered electricity generation | Photovoltaic | Nuclear power |
Carbon emission rate (kg/MWh) | 979 | 418 | 0 | 0 | 0 |
Conventional rack and the carbon emission data of distributed generation resource are as shown in table 4.Renewable energy power generation unit was being run
There is no carbon emission in journey, therefore is 0.
Following table 5 is batch processing load parameter.
Table 5
Load class | Arrival time | Deadline limitation |
Load 1 | 0:00 | 24:00 |
Load 2 | 0:00 | 12:00 |
Load 3 | 13:00 | 24:00 |
For Interactive workload data source in document, this method considers 3 kinds of batch processing loads.Interactive workload in total load
Half ratio is respectively accounted for batch processing load.Batch processing load 1 and batch processing load 2 evening, 12 points of arrival data centers, load 3
It is reached for noon, and various loads have the respective deadline to limit.As shown in table 5.
Following table 6 is the power prediction value of wind-power electricity generation.
Table 6
It can be seen that the fluctuation of wind-power electricity generation is larger, 7:Wind power highest when 00, value 6.968MW,
15:When 00, wind power is minimum, and value 2.399MW differs by more than one times.
Following table 7 is the power prediction value of photovoltaic power generation.
Table 7
It can be seen that photovoltaic power generation only just can be carried out when there is illumination, 1:00-7:00 and 22:00-24:00 function
Rate is 0.
Following table 8 is the intraday electricity price of electricity market.
Table 8
Moment | 1:00 | 2:00 | 3:00 | 4:00 | 5:00 | 6:00 | 7:00 | 8:00 | 9:00 | 10:00 | 11:00 | 12:00 |
Market guidance (member/ MWh) | 19.76 | 18.77 | 18.09 | 17.54 | 18.06 | 20.08 | 21.82 | 22.48 | 22.38 | 23.42 | 24.31 | 19.76 |
Moment | 13:00 | 14:00 | 15:00 | 16:00 | 17:00 | 18:00 | 19:00 | 20:00 | 21:00 | 22:00 | 23:00 | 24:00 |
Market guidance (member/ MWh) | 22.49 | 21.63 | 21.62 | 21.59 | 21.83 | 23.18 | 25.11 | 22.73 | 18.63 | 17.02 | 15.64 | 14.14 |
The intraday electricity price of electricity market can be obtained by the actual motion value of U.S.'s PJM electricity market.
Corresponding objective function and constraint condition are formed with data according to above-mentioned parameter.This is a MIXED INTEGER linear gauge
(mixed-integer linear programming, the MILP) problem of drawing, is solved using the Cplex module in Matlab
This MILP problem, all model programmings can be realized on the computer of AMD [email protected], 8GB memory.
Specific embodiment 1:
Assuming that carbon emission weight γ=0.0025, i.e. operator are primarily upon economic cost minimum, obtained scheduling knot
Fruit is:
Main to consider the smallest operation plan a few days ago of economic cost as shown in Fig. 5~Fig. 8, Fig. 5 is present invention specific implementation
The charge and discharge electrical schematic of energy-storage system when a kind of carbon emission weight γ that mode provides is 0.0025;Fig. 6 is that the present invention is specific real
Batch processing load distribution when a kind of carbon emission weight γ that the mode of applying provides is 0.0025;Fig. 7 is the specific embodiment of the invention
The generation schedule of conventional rack when a kind of carbon emission weight γ provided is 0.0025;Fig. 8 mentions for the specific embodiment of the invention
Data center's aggregate resource optimizes schematic diagram when a kind of carbon emission weight γ supplied is 0.0025.It is to reduce electricity consumption because for the purpose of
Cost is added to spot market electricity price (ahead market electricity price) as reference in Fig. 6 and Fig. 8.
As shown in figure 5, energy storage is charged in 1-6,8,16-17,23-24 period, main cause is these periods
Electricity price is lower, and the Optimum cost to run system can choose to charge during this period of time.In 7,9-15,18-20
Energy storage system discharges in period are powered for data center, because electricity price is higher at this time, by energy storage system discharges come in data
The heart, which provides electric power, can reduce the electricity bought from power grid, to reduce purchases strategies.
Need processed as shown in fig. 6, Interactive workload reaches, quantity is interior more in 9-15, this and practical feelings
Condition is consistent.Batch processing 1 arrival time of load is 0:00, deadline 24:00, so the time that it can be distributed is more clever
It is living, according to result it can be seen that when it is assigned to 2,3 when and when 4 and when 23 and when 24 in the electricity price lower period;At batch
The arrival time of reason load 2 is 0:00, deadline 12:00, so it can only be dispensed on 12:Electricity price before 00 is lower
Time, according to result it can be seen that being handled when it is assigned to 2 and when 4;The arrival time of batch processing load 3 is 13:00,
Deadline is 24:00, therefore it must be distributed 13:00 later electricity price lower period was handled, and can be seen according to result
Electricity price can be reduced when it is ultimately assigned to 22 out and when 23 to handle.
In the Unit Combination result of Fig. 7, No. 3 unit generation amounts are noticeably greater than No. 1 and No. 2, and No. 2 units are 3:It is opened when 00
Begin to shut down and be continued until that the same day terminates.This is because the marginal cost of No. 3 coal units is minimum, thus reduce at
Need to have more power under this target, and No. 1 unit and No. 2 units experienced primary shutdown in the electricity price lower period, wherein No. 2
The bigger reason of Unit Commitment cost.
As can be seen from Figure 8, when market guidance is lower (2:00-5:00,21:00-24:00), data center has greatly
Ratio electricity consumption comes from electricity market, when electricity price is relatively high (6:00-20:00), the electricity consumption of data center is mainly by generator
Group and energy storage electric discharge provide.Whole day has a certain proportion of electric power from renewable energy, their generated energy is by data center
It makes full use of.
Specific embodiment 2:
Assuming that carbon emission weight γ=80, i.e. data center operator are primarily upon environmental pollution minimum, Fig. 9 is this hair
The charge and discharge electrical schematic of energy-storage system when a kind of carbon emission weight γ that bright specific embodiment provides is 80;Figure 10 is this hair
Batch processing load distribution when a kind of carbon emission weight γ that bright specific embodiment provides is 80;Figure 11 is that the present invention is specific real
The generation schedule of conventional rack when a kind of carbon emission weight γ that the mode of applying provides is 80;Figure 12 is specific embodiment party of the present invention
Data center's aggregate resource optimizes schematic diagram when a kind of carbon emission weight γ that formula provides is 80, wherein Fig. 8 and Fig. 9 acceptance of the bid
The carbon emission rate of bulk power grid has been infused as reference.
As shown in figure 9, energy-storage system charges in the period of 2-6,9,16,18,22,24, main cause is these periods
Power grid carbon emission amount is lower.In 7,10-15,17,19-21 period energy storage system discharges, because of the carbon emission amount of power grid at this time
It is higher, the electricity bought from power grid can be reduced to provide electric power for data center by energy storage system discharges, to reduce entirety
Carbon emission.
As shown in Figure 10, according to result it can be seen that carbon emission amount when batch processing load 1 is assigned to 2,4-6 is lower
Period in;It is handled when batch processing load 2 is assigned to 2,3;Locate when batch processing load 3 is ultimately assigned to 18,22
Reason, can reduce carbon emission amount to the maximum extent.
As shown in figure 11,1, No. 2 conventional rack generated energy is noticeably greater than No. 3, and No. 3 conventional racks are 2:Start to close when 00
Machine and be continued until the same day terminate.This is because the carbon emission amount of No. 3 coal units is higher, therefore reducing system carbon
Need to contribute less as far as possible under the target of discharge amount, although and No. 1 unit it is identical as No. 2 unit carbon emission rates, No. 2 unit sides
Border cost is lower, therefore than No. 1 unit output of No. 2 units is more.
In figure 12 it can be seen that the maximum power on daytime is supplied mainly from more environmentally friendly Gas Generator Set, when power grid
Carbon emission is lower than when the carbon emission of Gas Generator Set, the electric power of data center is mainly from electricity market in microgrid;When power grid discharges
Gao Shi, electric power are mainly provided by Gas Generator Set and energy storage system discharges.Whole day has a certain proportion of electric power from renewable
The energy, their generated energy are made full use of by data center.
The embodiment of the present invention provides a kind of data center resource Optimization Scheduling and computer storage medium, passes through data
The history workload data prediction directly processing at center loads (Interactive workload) and can postpone to load (batch processing load)
Load capacity, finally maximum using profit or/and environmental impact minimization is target, scheduling power supply handle interactive bear in real time
While load, arrange batch processing load processing so that data center reaches, economic cost is minimum, environmental pollution is minimum or
The solution of the two tradeoff;The joint optimal operation for realizing economic operation cost and carbon emission amount, may be implemented its it is economical at
This minimum or the smallest operation target of environmental pollution, and can the two-part weight of flexible modulation, have it is very high it is perspective,
It is innovative;By adjusting from the combination of the conventional rack of the purchase of electricity of power grid, conventional power source, energy storage charge and discharge, flexibly load point
With optimization aim is realized, consider that content is comprehensive, compatibility is strong;And for the operation of data center's actual operation provide it is theoretical according to
According to and guidance.
The above-mentioned embodiment of the present invention can be implemented in various hardware, Software Coding or both combination.For example, this hair
Bright embodiment can also be the execution above method in data signal processor (Digital Signal Processor, DSP)
Program code.The present invention can also refer to computer processor, digital signal processor, microprocessor or field programmable gate
The multiple functions that array (Field Programmable Gate Array, FPGA) executes.It can configure according to the present invention above-mentioned
Processor executes particular task, by execute define the machine-readable software code of ad hoc approach that the present invention discloses or
Firmware code is completed.Software code or firmware code can be developed as different program languages and different formats or form.
It can also be different target platform composing software codes.However, executing the software code and other types of task according to the present invention
Different code pattern, type and the language of configuration code do not depart from spirit and scope of the invention.
The foregoing is merely the schematical specific embodiments of the present invention, before not departing from conceptions and principles of the invention
It puts, the equivalent changes and modifications that any those skilled in the art is made should belong to the scope of protection of the invention.
Claims (10)
1. a kind of data center resource Optimization Debugging method, which is characterized in that this method includes:
Different types of load capacity is predicted according to the history workload data of data center;
Obtain the first electric energy supply of renewable energy, the second electric energy supply of energy-storage system, electricity market electricity charge valence
Lattice and power grid carbon emission amount;And
According to the load capacity, the first electric energy supply, the second electric energy supply, the electric rate and the electricity
Net carbon emission amount determines the load distribution of each period and the power supply composition of the data center.
2. data center resource Optimization Debugging method as described in claim 1, which is characterized in that according to the history of data center
Before workload data predicts the step of different types of load capacity, this method further includes:
Configure the constraint condition of the data center.
3. data center resource Optimization Debugging method as claimed in claim 2, which is characterized in that the constraint condition includes work
Make load restraint, energy-storage system operation constraint, conventional rack operation constraint and power balance constraint.
4. data center resource Optimization Debugging method as described in claim 1, which is characterized in that according to the load capacity, institute
It states the first electric energy supply, the second electric energy supply, the electric rate and the power grid carbon emission amount and determines the number
After the step of the load distribution of each period at center and power supply composition, this method further includes:
The total operating cost and total carbon emissions amount of the data center are calculated according to the power supply.
5. data center resource Optimization Debugging method as claimed in claim 4, which is characterized in that total fortune of the data center
Battalion's cost and the expression Total of total carbon emissions amount are:
Wherein, t is the period;T is the set of time period t;For total operating cost;For total carbon
Discharge amount;γ is the weighted value of total carbon emissions amount;It is data center in time period t from the cost of power grid power purchase;For
The operation cost of the conventional rack of data center in time period t;It is generated for data center in time period t from power grid power purchase
Carbon emission amount;The carbon emission amount generated in conventional rack operation for data center in time period t.
6. data center resource Optimization Debugging method as claimed in claim 5, which is characterized in that the data center is from power grid
The cost of power purchaseSpecific formula be:
Wherein, t is the period;Pt gridThe electricity bought in time period t from power grid for data center;For in time period t
The electricity price of interior electricity market.
7. data center resource Optimization Debugging method as claimed in claim 6, which is characterized in that the tradition of the data center
The operation cost of unitSpecific formula be:
Wherein, t is the period;T is the set of time period t;L is conventional rack number;L is conventional rack quantity;SUlFor tradition
Unit l opens machine cost time period t;SDlIndicate conventional rack l in the shutdown cost of time period t;OlIndicate that conventional rack l exists
Idle running cost in time period t under no-load condition;ClFor operating cost of the conventional rack l in time period t;ut,l
For conventional rack l whether in 0,1 variable that is switched in time period t;vt,lFor conventional rack l whether shut down in time period t 0,
1 variable;ot,lFor conventional rack l whether in 0,1 unloaded variable in time period t;For the traditional machine in time period t
The generated energy of group.
8. data center resource Optimization Debugging method as claimed in claim 7, which is characterized in that the data center is from power grid
The carbon emission amount that power purchase generatesSpecific formula be:
Wherein, t is the period;K is fuel type sequence;N is fuel type total amount;Pt gridIt is data center in time period t
The electricity bought from power grid;ekTo indicate carbon emission rate of the kth kind fuel in power generation;gk,tFor the limit for indicating kth kind fuel
Accounting of the unit in time period t.
9. data center resource Optimization Debugging method as claimed in claim 8, which is characterized in that the tradition of the data center
The carbon emission amount generated in unit operationSpecific formula be:
Wherein, t is the period;L is conventional rack number;L is conventional rack quantity;elFor l platform conventional rack in power generation
Carbon emission rate;For the generated energy of l platform conventional rack.
10. a kind of computer storage medium comprising computer executed instructions, the computer executed instructions are via data processing
When equipment processing, which requires 1~9 any data center resource Optimization Debugging method.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810573669.8A CN108898282A (en) | 2018-06-06 | 2018-06-06 | Data center resource Optimization Scheduling and computer storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810573669.8A CN108898282A (en) | 2018-06-06 | 2018-06-06 | Data center resource Optimization Scheduling and computer storage medium |
Publications (1)
Publication Number | Publication Date |
---|---|
CN108898282A true CN108898282A (en) | 2018-11-27 |
Family
ID=64343934
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810573669.8A Pending CN108898282A (en) | 2018-06-06 | 2018-06-06 | Data center resource Optimization Scheduling and computer storage medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108898282A (en) |
Cited By (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110084444A (en) * | 2019-05-27 | 2019-08-02 | 华北电力大学 | A kind of cloud data center power load dispatching method considering natural resources randomness |
CN111275285A (en) * | 2019-12-30 | 2020-06-12 | 昆明电力交易中心有限责任公司 | Power consumption regulation and control method and system considering interruptible load capacity |
CN112801331A (en) * | 2019-11-14 | 2021-05-14 | 谷歌有限责任公司 | Shaping of computational loads using real-time scheduling of virtual capacity and preferred location |
CN112926785A (en) * | 2021-03-10 | 2021-06-08 | 国网吉林省电力有限公司信息通信公司 | Cloud data center load optimization scheduling method supporting electric power big data |
CN112966857A (en) * | 2021-02-10 | 2021-06-15 | 合肥工业大学 | Multi-energy collaborative optimization method and system for data center |
CN113872180A (en) * | 2020-06-30 | 2021-12-31 | 华为技术有限公司 | Method and system for supplying power to equipment and related equipment |
CN114936810A (en) * | 2022-07-25 | 2022-08-23 | 东南大学溧阳研究院 | Day-ahead scheduling method based on data center space-time transfer characteristics |
CN115149572A (en) * | 2022-07-13 | 2022-10-04 | 北京百度网讯科技有限公司 | Parallel operation control method and device of generator and computer program product |
CN115443592A (en) * | 2020-02-24 | 2022-12-06 | 领先电力解决方案有限公司 | Techniques for dynamically scheduling generator power |
CN116207844A (en) * | 2023-04-27 | 2023-06-02 | 南方电网调峰调频发电有限公司信息通信分公司 | Fault prevention method and device based on power communication equipment |
CN117196683A (en) * | 2023-09-26 | 2023-12-08 | 国网宁夏电力有限公司 | LSTM-based thermal power enterprise intra-month rolling matching transaction bidding system |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103426122A (en) * | 2013-05-17 | 2013-12-04 | 中国能源建设集团广东省电力设计研究院 | Comprehensive evaluation method of micro-grid |
US8645150B2 (en) * | 2010-04-12 | 2014-02-04 | International Business Machines Corporation | Source aware data center power profiles |
CN107025610A (en) * | 2017-04-24 | 2017-08-08 | 东华大学 | A kind of data center's Cost Optimization Approach |
-
2018
- 2018-06-06 CN CN201810573669.8A patent/CN108898282A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US8645150B2 (en) * | 2010-04-12 | 2014-02-04 | International Business Machines Corporation | Source aware data center power profiles |
CN103426122A (en) * | 2013-05-17 | 2013-12-04 | 中国能源建设集团广东省电力设计研究院 | Comprehensive evaluation method of micro-grid |
CN107025610A (en) * | 2017-04-24 | 2017-08-08 | 东华大学 | A kind of data center's Cost Optimization Approach |
Non-Patent Citations (2)
Title |
---|
CHUANGANG REN ET AL.: "Carbon-Aware Energy Capacity Planning for Datacenters", 《2012 IEEE 20TH INTERNATIONAL SYMPOSIUM ON MODELING, ANALYSIS AND SIMULATION OF COMPUTER AND TELECOMMUNICATION SYSTEMS》 * |
邓维 等: "云计算数据中心的新能源应用:研究现状与趋势", 《计算机学报》 * |
Cited By (19)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110084444B (en) * | 2019-05-27 | 2021-05-14 | 华北电力大学 | Cloud data center power load scheduling method considering randomness of natural resources |
CN110084444A (en) * | 2019-05-27 | 2019-08-02 | 华北电力大学 | A kind of cloud data center power load dispatching method considering natural resources randomness |
CN112801331A (en) * | 2019-11-14 | 2021-05-14 | 谷歌有限责任公司 | Shaping of computational loads using real-time scheduling of virtual capacity and preferred location |
US11960255B2 (en) | 2019-11-14 | 2024-04-16 | Google Llc | Compute load shaping using virtual capacity and preferential location real time scheduling |
CN112801331B (en) * | 2019-11-14 | 2024-02-06 | 谷歌有限责任公司 | Shaping of computational loads with virtual capacity and preferred location real-time scheduling |
CN111275285A (en) * | 2019-12-30 | 2020-06-12 | 昆明电力交易中心有限责任公司 | Power consumption regulation and control method and system considering interruptible load capacity |
CN111275285B (en) * | 2019-12-30 | 2024-02-20 | 昆明电力交易中心有限责任公司 | Power consumption regulation and control method and system considering interruptible load capacity |
CN115443592A (en) * | 2020-02-24 | 2022-12-06 | 领先电力解决方案有限公司 | Techniques for dynamically scheduling generator power |
CN113872180A (en) * | 2020-06-30 | 2021-12-31 | 华为技术有限公司 | Method and system for supplying power to equipment and related equipment |
WO2022001304A1 (en) * | 2020-06-30 | 2022-01-06 | 华为技术有限公司 | Device power supply method, system, and related device |
CN112966857A (en) * | 2021-02-10 | 2021-06-15 | 合肥工业大学 | Multi-energy collaborative optimization method and system for data center |
CN112966857B (en) * | 2021-02-10 | 2022-09-30 | 合肥工业大学 | Data center multifunctional collaborative optimization method and system |
CN112926785A (en) * | 2021-03-10 | 2021-06-08 | 国网吉林省电力有限公司信息通信公司 | Cloud data center load optimization scheduling method supporting electric power big data |
CN115149572B (en) * | 2022-07-13 | 2024-02-02 | 北京百度网讯科技有限公司 | Parallel operation control method and device of generator and computer program product |
CN115149572A (en) * | 2022-07-13 | 2022-10-04 | 北京百度网讯科技有限公司 | Parallel operation control method and device of generator and computer program product |
CN114936810B (en) * | 2022-07-25 | 2022-10-18 | 东南大学溧阳研究院 | Day-ahead scheduling method based on data center space-time transfer characteristics |
CN114936810A (en) * | 2022-07-25 | 2022-08-23 | 东南大学溧阳研究院 | Day-ahead scheduling method based on data center space-time transfer characteristics |
CN116207844A (en) * | 2023-04-27 | 2023-06-02 | 南方电网调峰调频发电有限公司信息通信分公司 | Fault prevention method and device based on power communication equipment |
CN117196683A (en) * | 2023-09-26 | 2023-12-08 | 国网宁夏电力有限公司 | LSTM-based thermal power enterprise intra-month rolling matching transaction bidding system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108898282A (en) | Data center resource Optimization Scheduling and computer storage medium | |
Luna et al. | Online energy management systems for microgrids: Experimental validation and assessment framework | |
Ding et al. | Emission-aware stochastic resource planning scheme for data center microgrid considering batch workload scheduling and risk management | |
Wu et al. | GENCO's risk-based maintenance outage scheduling | |
Mazidi et al. | A model for flexibility analysis of RESS with electric energy storage and reserve | |
Jawad et al. | A robust optimization technique for energy cost minimization of cloud data centers | |
CN105006844B (en) | Electric power system day-ahead robust scheduling system under intermittent power generation grid-connected condition | |
US20160218505A1 (en) | System and Method for Energy Asset Sizing and Optimal Dispatch | |
Xiao et al. | Joint chance-constrained multi-objective optimal function of multi-energy microgrid containing energy storages and carbon recycling system | |
Olsen et al. | Profitable emissions-reducing energy storage | |
CN110232583B (en) | Electric power market marginal price planning method considering carbon emission right | |
CN113205273B (en) | Low-carbonization power supply planning method and system considering off-site energy transaction | |
Daneshvar et al. | A transactive energy management framework for regional network of microgrids | |
KR20210100699A (en) | hybrid power plant | |
Awasthi et al. | Operation of datacenter as virtual power plant | |
Tang et al. | Generation scheduling under a CO 2 emission reduction policy in the deregulated market | |
CN115189420A (en) | Power grid dispatching operation optimization method and system considering uncertainty of supply and demand sides | |
Cui et al. | Optimal electricity allocation model under China's planning-market double-track mechanism considering bidding game of generation companies | |
Ye et al. | Joint workload scheduling in geo-distributed data centers considering UPS power losses | |
Ryu | Operation planning of energy storage system considering multiperiod energy supplies and demands | |
CN111210079B (en) | Operation optimization method and system for distributed energy virtual power plant | |
CN112308411A (en) | Comprehensive energy station random planning method and system based on dynamic carbon transaction model | |
Luna et al. | Generation and demand scheduling for a grid-connected hybrid microgrid considering price-based incentives | |
Wu et al. | Improved light robust optimization strategy for virtual power plant operations with fluctuating demand | |
Tian et al. | Coordinated RES and ESS Planning Framework Considering Financial Incentives Within Centralized Electricity Market |
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 |