CN106779291A - Intelligent power garden demand response strategy - Google Patents

Abstract

The invention discloses a kind of Itellectualized uptown demand response control strategy.The main body that is applicable that intelligent demand models most of at present are set up is single resident, ignores the diversity and uncertainty of actual user colony, is not directed to the garden that multiple users are polymerized.The present invention is comprised the following steps：1) to the flexible load of intelligent power garden, response mode difference is classified as desired；2) corresponding demand response Policy model is set up to different flexible loads；3) corresponding demand response Policy model is set up to whole intelligent power garden；4) effect of the garden user under different demands response policy is calculated.The present invention is with user power utilization cost minimization as target, scheduling is optimized to all demand response resources in the presence of different electricity price incentive mechanisms, reach and reduce simultaneously user power utilization cost and improve the effect of load electricity consumption behavior, intelligent power garden is effectively participated in the middle of the demand response under market and competitive mechanisms.

Description

Intelligent power garden demand response strategy

Technical field

The invention belongs to intelligent power grid technology field, particularly a kind of intelligent power garden based on novel load structure is needed Seek response policy.

Background technology

Issued after country in 2015《Some opinions on further in-depth power system reform》Afterwards, Demand Side Response Start to face new opportunities and challenges in China.For power network, demand response can improve load curve, reduce load peak, drop Low peak-valley difference, so that system operation cost is reduced, while alleviating the power grid construction investment problem brought because load increases.For Customer-side, demand response can be in the case where user power utilization satisfaction not be influenceed for user reduces electric cost.

With the development of intelligent grid, the flexible load such as intelligent building, intelligent residential district and wind-powered electricity generation, photovoltaic, energy storage, miniature combustion Gas-turbine distributed power supply starts progressively to be linked into power distribution network.Meanwhile, with the advanced measuring system for supporting various services And the application of the technology such as EMS so that the flexible load and distributed power source of user side turn into interacting for grid side May, its overall electricity consumption behavior is more intelligent, i.e. intelligent power.Under power market reform background, Load aggregation business can make For " go-between " assists each there is the sole user of demand response resource to participate in the middle of electricity market.By LA to demand The scheduling of resource response, the intelligent power garden aggregated into by industry, business, resident load under the conditions of intelligent power can be according to electricity Net side pumping signal actively, be actively engaged in the middle of demand response.Intelligent power garden is to forgive traditional load, flexibility to bear Lotus and the new garden of various distributed power sources.

Current domestic scholars there has been many research to demand response, and by the difference of method, modeling can substantially be divided into base In price elasticity of electricity demand matrix, based on consumer psychology, based on three kinds of Principle of Statistics.User power utilization pair is studied at present During the respondent behavior of tou power price, the requirement to data volume is higher, and less data amount will be unable to preferably complete data mining Work；And set up in big data excavation, different documents take the different intelligent algorithms to carry out data mining, but these are counted According to excavation be all based on it is some assume basis on, these hypothesis also result in electricity price change afterload curve simulation accuracy compared with It is low.For in the establishment of customer impact factor, most of documents use single factor analysis, such as only consider psychological factor or Person's other factors, and the main body that is applicable that big multi-model is set up is single resident, ignores the polynary of actual user colony Property and uncertainty, are not directed to the garden that multiple users are polymerized.

The content of the invention

Consider the problems such as user level is excessively single, model accuracy is relatively low for conventional requirement response model, the present invention is carried For a kind of intelligent power garden demand response strategy, it bears from Load aggregation business's angle, the flexibility to intelligent power garden Lotus as desired classified and set up corresponding demand response Policy model by response mode difference, and the traditional load of combination participates in needing Response mode and distributed power source Optimized Operation is asked to form demand response Policy model comprehensively suitable for intelligent power garden.

Therefore, the present invention is adopted the following technical scheme that：Intelligent power garden demand response strategy, comprises the following steps：

1) to the flexible load of intelligent power garden, response mode difference is classified as desired；

2) corresponding demand response Policy model is set up to different flexible loads；

3) corresponding demand response Policy model is set up to whole intelligent power garden；

4) effect of the garden user under different demands response policy is calculated.

Further, step 1) in, according to flexible load demand response characteristic, conventional load is classified as, can adjust and bear Lotus and the class of transferable load three.

Further, participate in can interrupt moderation response conventional load model constraints be：

In formula,It is the load that user j cuts off in hour t, kW；I_l,j,tFor load j hour t load cut off 0th, 1 state variable, if user j cuts off load, I in hour t_l,j,t=1, otherwise it is 0；Tune can be interrupted for user j may participate in The peak load excision power of section, kW；For user j daily participate in can interrupt moderation peak load cut off hourage, h；It is user j in the prediction load of hour t, kW.

Further, lighting load is to the response model constraints of Price Mechanisms in can adjust load：

In formula,It is load j in the power of hour t, kW；It is load j in the benchmark electric power of hour t, kW；I_a,j,tLoad Regulation 0,1 state variable for load j in hour t, the I when power network electricity price is higher than electricity price threshold value_a,j,t=1, Load power reduce toOtherwise it is 0；It is load j in the minimum electric power of hour t, kW is set by the user； ε is≤0.001 arithmetic number；ρ_tIt is the power network electricity price of hour t, It is the electricity price threshold value of user's setting,

Further, response restricted model of the air-conditioning based on Price Mechanisms is in can adjust load：

In formula, T_k,t、Respectively air conditioner load k hour t air-conditioner temperature and environment temperature, DEG C；α_kIt is radiating letter Number；It is temperature gain of the air-conditioning when opening, DEG C；Δ t is spaced for control time, 1h；C_k、R_kThe respectively thermoelectricity of air-conditioning Hold kW ˙ h/ DEG C and thermal resistance DEG C/kW, η_kIt is the operating efficiency of air-conditioning,It is air conditioner load k in the electric power of hour t, kW；Minimum per hour, the maximum electric power of respectively air-conditioning k, kW；It is the setting of hour t air conditioner load Temperature, DEG C；Air conditioner load k is higher than threshold in the former design temperature and electricity price of hour t respectively under refrigerating state ValueWhen highest setting temperature, DEG C, be set by the user；I_a,k,tFor air conditioner load k adjusts 0,1 in the design temperature of hour t State variable, when power network electricity price is higher than electricity price threshold valueWhen I_a,k,t=1, air-conditioning design temperature is increased toOtherwise it is 0。

Further, dish-washing machine, dryer are to the response model constraints of Price Mechanisms in transferable load：

In formula,It is transferable load j in the power of hour t, kW；It is the average work(of transferable load j Rate, kW；I_b,j,tIt is transferable load j in 0,1 state variable of hour t, the I when load operation_b,j,t=1, otherwise it is 0； I_b,j,(t-1)It is transferable load j in 0,1 state variable of the previous hour of hour t；For load j in hour t Run time, h；It is load j in the run time of the previous hour of hour t, h；U_b,jIt is transferable load Load operation cycle, h；τ is that user sets permissible load operation starting time, h；T terminates run time for user's assumed load, h。

Further, electric automobile load is to the response model constraints of Price Mechanisms in transferable load：

In formula,Respectively electric automobile k minimum, maximum charge power, kW hourly；For Electric automobile k hour t charge power, kW；I_b,k,tIt is 0,1 state variable that electric automobile k charges in hour t, when electronic I when automobile charges_b,k,t=1, otherwise it is 0；T_aIt is user's last time trip end time, that is, charges the time started；T_bIt is use Family regulation prepares the travel time, that is, charge the end time；It is electric automobile k in the charged state of hour t；It is electricity Charged states of the electrical automobile k in the previous hour of hour t；For electric automobile k prepares the travel time in user's regulation Charged state；Δ t is spaced for control time, 1h；It is batteries of electric automobile capacity, kWh.

Further, in the case of multiple electric automobile loads, it is considered to initial state of charge in electric automobile constraintWith User's last time trip end time T_aRandomness, obtain electric automobile initial state of chargeProbability density function For：

In formula, D is the maximum range of electric automobile, user's last time trip end time T_aThink approximate obedience Normal distribution, σ_dThe standard deviation of distance travelled, μ_dIt is the average of the distance travelled of Normal Distribution.

Further, step 3) in, the corresponding demand response Policy model set up to whole intelligent power garden is as follows：

User side is with the power-balance constraint of grid side：

In formula,It is wind power generator i in the generated output of hour t, kW；It is photovoltaic generation unit i hour t's Generated output, kW；It is gas turbine i in the generated output of hour t, kW；It is battery i in the charge and discharge power of hour t； P_g,tIt is grid side in the output power of hour t, kW；It is user j in the prediction load of hour t, kW；It is user j small When t when the load that cuts off, kW；It is lighting load j in the power of hour t, kW；It is air conditioner load k in the electricity consumption of hour t Power, kW；It is the transferable load j such as dish-washing machine, dryer in the power of hour t, kW；It is electric automobile k in hour t Charge power, kW；

Grid side power constraint is：

In formula,Respectively power network is minimum, maximum output power, and this constraint reflects power network actual power work( Rate is constrained, while in Spot Price excitation, this constraint can also prevent under the conditions of intelligent power load the electricity price relatively low period because There is new load peak in load transfer；

Gas turbine restricted model is：

In formula, F_i ^TIt is gas turbine power generation cost function；It is gas turbine i in the generated output of hour t, kW； It is the generating efficiency of gas turbine i；ρ_gIt is Gas Prices,M^TIt is operation expense,H_iIt is combustion gas The heat consumption rate of turbine i, kJ/kWh；P_i ^T,min、P_i ^T,maxThe respectively minimum of gas turbine i, maximum power generation；It is combustion gas Switching on and shutting down 0,1 state variables of the turbine i in hour t, when gas turbine is started shooting to be runOtherwise it is 0；

Battery energy storage restricted model is：

I_d,i,t+I_c,i,t≤1

E_i,0=E_i,NT

In formula,It is battery i in the charge and discharge power of hour t, kW；P_d,i,tIt is battery i in the discharge power of hour t； P_c,i,tIt is battery i in the charge power of hour t；The respectively minimum of battery i, maximum charge power, kW； I_d,i,t、I_c,i,tElectric discharge, charging 0,1 state variables of the respectively battery i in hour t, when battery i discharges in hour t, I_d,i,t =1, otherwise I_d,i,t=0, similarly, as battery i when hour t charges I_c,i,t=1, otherwise I_c,i,t=0；Respectively It is the minimum of battery i, maximum charge power, kW；Minimum, the maximum discharge power of respectively battery i, kW； E_i,tCharged states of the battery i in hour t；It is the charge and discharge efficiency of battery i；Respectively battery i is most Small, maximum capacity, kWh；NT is the total hourage of battery plan；E_i,(t-1)Charged states of the battery i in hour t；When Δ t is for control Between be spaced, 1h；E_i,0Battery initial state of charge；E_i,NTThe charged state of last hour of battery.

It is minimum as optimization aim using user power utilization cost from Load aggregation business's angle, set up Optimized model：

In formula, F_i ^TIt is gas turbine power generation cost function；It is gas turbine i in the generated output of hour t, kW；P_g,t It is grid side in the output power of hour t, kW；It is the load that user j cuts off in hour t, kW；ρ_tIt is the power network of hour t Electricity price,ρ_l,jFor the interruptible load of load j cuts off price,

Further, step 4) effect computational methods of the garden user under different demands response policy are as follows：Using CPLEX solvers are solved to it.

The invention has the advantages that：With intelligent power garden as object, detailed need are established to flexible load Seek response model, the model with user power utilization cost minimization as target, to all need in the presence of different electricity price incentive mechanisms Ask resource response to optimize scheduling, reach and reduce simultaneously user power utilization cost and improve the effect of load electricity consumption behavior, make intelligence Energy electricity consumption garden can be participated in effectively in the middle of the demand response under market and competitive mechanisms.Carried out by load power curve Power characteristic analysis can be analyzed to different electricity pricing schemes under different Price Mechanisms or Price Mechanisms of the same race, so that Optimal excitation mode is drawn, therefore the model also has Auxiliary Significance to the electricity pricing under electric system reform.

Obtained by emulation, after TOU and IL is participated in, load peak reduces 8.51%, and peak value time of occurrence is by 15h To 22h after shifting, peak-valley difference reduces 1.91%.Total electricity consumption reduces 5.79%；Load participates in load peak drop after CPP and IL Low by 4.76%, peak value time of occurrence moves forward to 14h by 15h, and total electricity consumption reduces 6.74%；Load participate in RTP and IL it Afterload peak value reduces 8.31%, and peak value time of occurrence is postponed to 23h by 15h, and peak-valley difference reduces 7.56%, total electricity consumption Reduce 8.35%.

Brief description of the drawings

Fig. 1 is principle of the invention figure.

Fig. 2 is conventional load predicted value curve map.

Fig. 3 is lighting load predicted value curve map.

Fig. 4 is day environment temperature predicted value variation diagram.

Fig. 5 is wind-powered electricity generation, solar power generation power prediction value curve map.

Fig. 6 is Spot Price (RTP) predicted value curve map.

Fig. 7 is that conventional load participates in IL response condition figures.

Fig. 8 is response condition figure under lighting load TOU.

Fig. 9 is response condition figure under air conditioner load TOU.

Figure 10 is response condition figure under dish-washing machine, dryer TOU.

Figure 11 is response condition figure under electric automobile TOU.

Figure 12 is response condition figure under gas turbine, energy-storage battery TOU.

Figure 13 is without load power curve map in the case of DR.

Figure 14 is that user participates in load power curve map during IL and TOU.

Figure 15 is that user participates in load power curve map during IL and CPP.

Figure 16 is that user participates in load power curve map during IL and RTP.

Specific embodiment

With reference to Figure of description, the invention will be further described.

(1) demand response modeling analysis are carried out to intelligent power garden conventional load

Conventional load (industrial load) can be participated in IL (can interrupt moderation) by signing a contract, cut-out load.Root According to its demand response characteristic, can obtain participating in the conventional load model constraints of IL responses.

In formula,The load (kW) cut off in hour t for user j；I_l,j,tFor load j hour t load cut off 0th, 1 state variable, if user j cuts off load, I in hour t_l,j,t=1, otherwise it is 0；The maximum of IL is may participate in for user j Load cuts off power (kW)；For the peak load that user j participates in IL daily cuts off hourage (h)；It is user j small When t prediction load (kW).

(2) demand response modeling analysis are carried out to the adjustable load in intelligent power garden

Adjustable load (illumination, air-conditioning such as in business, resident load) in flexible load can be by adjusting itself Part electric power is participated in electricity price excitation, reduces itself electricity consumption grade.An electricity price threshold can be set for type load user Value, and electric power grade is set according to itself comfort level.When power network Spot Price is higher than the threshold value, load can be reduced automatically To user preset electric power grade.

A () is based on electricity price incentive mechanism, set up lighting load model in adjustable load

By the setting of electricity price incentive mechanism and electric power grade, lighting apparatus can reduce certain intensity of illumination to setting Value, can obtain its model constraints to Price Mechanisms response

In formula,It is load j in the power (kW) of hour t；It is load j in the benchmark electric power of hour t (kW)；I_a,j,tLoad Regulation 0,1 state variable for load j in hour t, the I when power network electricity price is higher than electricity price threshold value_a,j,t= 1, load power reduce toOtherwise it is 0；It is load j in the minimum electric power (being set by the user) of hour t (kW)；ε is the arithmetic number of a very little；ρ_tIt is the power network electricity price of hour t It is the electricity price threshold value of user's setting

B () simplification obtains air-conditioning equivalence physochlaina infudibularis exponential model, its demand response restricted model is drawn based on electricity price excitation

In formula, T_k,t、Respectively air conditioner load k hour t air-conditioner temperature and environment temperature (DEG C), α_kIt is radiating letter Number,It is temperature gain (DEG C) of the air-conditioning when opening, Δ t is control time interval (1h), C_k、R_kThe respectively thermoelectricity of air-conditioning Hold (kW ˙ h/ DEG C) and thermal resistance (DEG C/kW), η_kIt is the operating efficiency of air-conditioning,It is air conditioner load k in the electric power of hour t (kW),Minimum per hour, the maximum electric power (kW) of respectively air-conditioning k,It is setting for hour t air conditioner load Constant temperature degree (DEG C).

Air-conditioning equipment can adjust temperature to preset value to reduce load under electricity price incentive mechanism, based on model above to system Air conditioner load regulation under cold state, it is constrained to：

In formula,Air conditioner load k is high in the former design temperature and electricity price of hour t respectively under refrigerating state In threshold valueWhen highest setting temperature (being set by the user) (DEG C)；I_a,k,tFor air conditioner load k is adjusted in the design temperature of hour t 0,1 state variable of section, when power network electricity price is higher than electricity price threshold valueWhen I_a,k,t=1, air-conditioning design temperature is increased toIt is no It is then 0.

(3) demand response modeling analysis are carried out to the transferable load in intelligent power garden

Power load can be transferred to low rate period and carry out electricity consumption by such flexible load according to electricity price signal.This type load Two classes, the respectively load such as dish-washing machine, dryer and electric automobile load can be divided into.

A () is based on the dish-washing machine of segmentation electricity price excitation, dryer response constraint modeling

Although different periods electric power is different in one cycle of operation of load (such as the washing of washing machine with dry two ranks Section), but fixed based on operation period total electricity consumption, can power be replaced with a constant average power per hour by the load operation period (i.e. equation 13), the period load power consumption is the product of this mean power and the total hourage of operation.User can be according to itself need Seeking regulation allows the time period of such load operation, the time that both permissible load brings into operation and must terminate.In the time period Interior, load can bring into operation in any hour, but must in stipulated time section entire run one load period (i.e. inequality 14).Such load model constraint specification is as follows：

In formula,It is transferable load j in the power (kW) of hour t；It is the mean power of transferable load j； I_b,j,tIt is transferable load j in 0,1 state variable of hour t, the I when load operation_b,j,t=1, otherwise it is 0；It is load Run times (h) of the j in hour t；U_b,jIt is load operation cycle (h) of transferable load；τ is that user sets permissible load Operation starting time (h)；T terminates run time (h) for user's assumed load.

B () is based on the electric automobile load constraint modeling of segmentation electricity price excitation；

Electric automobile is charged during being gone on a journey by second day at the end of being gone on a journey for the last time on the day of user.Due to charging Required duration is less than this period, any hour load within the period may be selected whether to charge and its charge power size (i.e. Inequality 17).Electric automobile needs fully charged (equation 19) when user goes on a journey.

In formula,Respectively electric automobile k minimum, maximum charge power (kW) hourly；It is electricity Charge powers of the electrical automobile k in hour t；I_b,k,tIt is 0,1 state variable that electric automobile k charges in hour t, when electric automobile fills I when electric_b,k,t=1, otherwise it is 0；T_aIt is user's last time trip end time (charging the time started)；T_bFor user specifies Prepare the travel time (charging the end time)；For electric automobile k hour t charged state (State of Charge, SOC)；It is batteries of electric automobile capacity (kWh).

In the case where there are multiple electric automobile loads, initial state of charge in electric automobile constraint need to be consideredWith with Family last time trip end time T_aRandomness.Electric automobile initial state of charge exercises that mileage is approximate meets line with its day Sexual intercourse.Because user's daily travel approximately obeys logarithm normal distribution, electric automobile initial state of charge can be obtained's Probability density function is：

In formula, D is the maximum range of electric automobile.User's last time trip end time T_aIt is believed that approximate clothes From normal distribution.

(4) demand response modeling analysis are integrally carried out to intelligent power garden

In intelligent power garden, in addition to electrical equipment, also new energy distributed power source, battery energy storage and combustion gas wheel Machine equipment.Wind-powered electricity generation, photovoltaic generation due to its uncontrollability and exert oneself fluctuation the characteristics of, it is necessary to have the combustion of flexible peak-shaving capability Gas-turbine and energy-storage battery carry out auxiliary adjustment in load side.User side under the conditions of equipped with wind-powered electricity generation, photovoltaic distributed power source, Preferentially can be powered using such distributed power source, remaining load is by user side gas turbine, energy-storage battery and grid side Power supply shared.

A () sets up the power-balance constraint for obtaining user side and grid side

In formulaIt is wind power generator i in the generated output (kW) of hour t；It is photovoltaic generation unit i hour t's Generated output (kW)；It is battery i in the charge and discharge power of hour t.

B () sets up and obtains grid side power constraint

In formulaRespectively power network is minimum, maximum output power.This constraint reflects power network actual power work( Rate constraint (due to the reason such as distribution system capacity-constrained or transformer constraint), while in Spot Price excitation, this constraint is also There is new load peak in the electricity price relatively low period because load is shifted in load under the conditions of intelligent power can be prevented.

C () sets up and obtains gas turbine constraint

Gas turbine power generation cost is made up of (i.e. equation gas cost and gas turbine operation maintenance cost two parts 24), Gas Turbine Output has minimax power constraint (i.e. inequality 26).

In formulaIt is the generating efficiency of gas turbine i；ρ_gIt is Gas PricesM^TIt is operation expenseH_iIt is the heat consumption rate (kJ/kWh) of gas turbine i；P_i ^T,min、P_i ^T,maxRespectively minimum, the maximum of gas turbine i Generated output；Switching on and shutting down 0,1 state variable for gas turbine i in hour t, when gas turbine is started shooting to be runIt is no It is then 0.

D () sets up and obtains battery energy storage constraint

When battery is discharged, battery is powered as power supply；When power supply is charged, battery as load (i.e. etc. 27).The battery same time can not simultaneously be charged, be discharged (i.e. inequality 28).Battery per hour receive most by charge and discharge power Greatly, minimum charge and discharge power constraint (i.e. inequality 29,30).Formula (31), (32) are battery charging state constraint, and both battery be not Electric discharge can be continued when completely depleted or continue to charge in maximum capacity.Battery planned end time charged state is filled with initially Electricity condition is identical (i.e. equation 33).

I_d,i,t+I_c,i,t≤1 (28)

E_i,0=E_i,NT (33)

In formula, P_d,i,tIt is battery i in the discharge power of hour t；P_c,i,tIt is battery i in the charge power of hour t；I_d,i,t、 I_c,i,tElectric discharge, charging 0,1 state variables of the respectively battery i in hour t, when battery i discharges in hour t, I_d,i,t=1, it is no Then I_d,i,t=0, similarly, as battery i when hour t charges I_c,i,t=1, otherwise I_c,i,t=0；Respectively battery i Minimum, maximum charge power (kW)；Minimum, the maximum discharge power (kW) of respectively battery i；E_i,tBattery i In the charged state of hour t；It is the charge and discharge efficiency of battery i；The respectively minimum of battery i, maximum capacity (kWh)；NT is the total hourage of battery plan (present invention is 24h).

E () is minimum as optimization aim using user power utilization cost from LA (Load aggregation business) angle, set up optimization mould Type；

In formulaIt is gas turbine power generation cost function；It is gas turbine i in the generated output (kW) of hour t；P_g,t It is grid side in the output power (kW) of hour t；ρ_l,jFor the interruptible load of load j cuts off price

Simulation example

This simulation example is carried out using MATLAB by taking certain intelligent power garden as an example for the uncertainty of all kinds of power wagons Monte Carlo simulation simultaneously draws the power curve after the participation of the load in the case of three kinds of electricity prices DR.It is with TOU situations first herein Example, is analyzed to single simulation result, and the response condition of interruptible load, flexible load and controlled distribution formula power supply is discussed.

The related data of interruptible load is as follows in example：Interruptible load excision price ρ_l=2.5User The daily peak load excision hourage for participating in ILUser always may participate in the peak load excision power P of IL^{l ,max}=100 (kWh).Conventional daily load prediction value is as shown in Figure 2.

Adjustable load relevant parameter is as follows：When electricity price threshold value of the electricity price higher than user's settingWhen, Minimum electric power P of the lighting load in hour t_t ^a,min(kW) it is set as the P of the moment electric power_t ^a,base80%.Illumination Predicted load is as shown in Figure 3.Radiating function alpha=0.82, thermal resistance R=2 (DEG C/kW), operating efficiency η=2.5 of air-conditioning are single Minimum per hour, the maximum electric power of platform air-conditioning is respectively 0 (kw), 3.5 (kw), former design temperature T^s,base=23 (DEG C), when Electricity price threshold value of the electricity price higher than user's settingWhen, highest setting temperature T^s,max=24 (DEG C), air-conditioning quantity It is 300.Day environment temperature T_t ^αPredicted value is as shown in Figure 4.

The relevant parameter of transferable load is as follows：

Dish-washing machine, dryer quantity are respectively 200, and its load data is as shown in the table：

Electric automobile minimum, maximum charge power hourlyRespectively 0 (kW), 3.3 (kW), it is electronic Automobile batteries capacityThe maximum range D=40 (mile) of electric automobile, user's daily travel Lognormal distribution parameter be μ_d=2.319, σ_d=0.88.User's last time trip end time T_aObeying 1 average is 17:00, variance is the normal distribution of 0.5h, is charged end time T_bIt is 24:00(h).Electric automobile quantity is 100.Wind-powered electricity generation And the pre- power scale of solar power generation is as shown in Figure 5.Gas turbine operation maintenance costGas turbine Heat consumption rate H=11613 (kJ/kWh), minimum, maximum power generationRespectively 30 (kW), 300 (kW).My god Right gas price lattice

Battery energy storage device parameter is as shown in the table：

3 kinds of electricity price types, i.e. TOU, CPP and RTP are set in this simulation example respectively.

Tou power price (TOU)：Peak period electricity priceUsually section electricity price Paddy period electricity priceCritical Peak Pricing (CPP)：The spike periodPeak period electricity price It is flat Period electricity pricePaddy period electricity price Spot Price (RTP) predicted value is as shown in Figure 6.

Conventional load participates in IL results as shown in fig. 7, respectively cutting off load in 15h and 16h in the afterload for participating in IL 100kW.Lighting load under TOU response condition as shown in figure 8, load cuts down power consumption altogether in the peak period (15-18h) 271kWh.Air conditioner load under TOU response condition as shown in figure 9, load peak is decreased to 933kW by 960kW, and peak value occurs Time is postponed to 18h by 15h.Air conditioner load total electricity consumption reduces 180kWh.Response condition is as schemed under dish-washing machine, dryer TOU Shown in 10.Due to user formulate dryer operation the period be 9-17h, in the period electricity price in dryer initial operating time point Reach minimum, therefore the load is not shifted.Dish-washing machine initial operating time point is located at electricity price usually section, therefore it runs time started (20-24h) postpones 1h to electricity price paddy period in user allows the operation period.Response condition such as Figure 11 institutes under electric automobile TOU Show.When TOU is had neither part nor lot in, electric automobile load is concentrated mainly on the electricity price peak period with usually section.It is former after participation TOU responses The electricity price peak period is transferred to electricity price paddy period, the peak load shifting effect of electric automobile load with the 95.9% of usually section power consumption Significantly.Response condition is as shown in figure 12 under gas turbine, energy-storage battery TOU.When market guidance is sent out higher than distributed gas turbine During electric cost (electricity price peak period), gas turbine is powered to reduce user side electric cost to load side.Battery energy storage in The electricity price paddy period is charged (power is negative), and is discharged in the electricity price peak period, and peak clipping is carried out to load side power curve Fill valley.

Load side integrated demand response condition under TOU, CPP, RTP and IL is discussed below analysis.

1st, as shown in figure 13, load is negative in the case where IL and TOU excitations are participated in for load curve in the case of load is not involved in DR Lotus curve is as shown in figure 14.Load peak is cut down to 1448.69kW by 1583.40kW after TOU and IL is participated in, and reduces 8.51%, while to 22h after peak value time of occurrence is moved by 15h.Peak-valley difference reduces to 773.16kW by 788.2kW, reduces 1.91%.Total electricity consumption reduces 5.79%, and the total electric cost of garden user is 1155.97 $.

2nd, load curve of the load in the case where IL and CPP excitations are participated in is as shown in figure 15.Participating in the afterload of CPP and IL Peak value is cut down to 1508.09kW, reduces 4.76%, while peak value time of occurrence moves forward to 14h by 15h.Peak-valley difference is increased to 831.76kW, increased 5.52%.Total electricity consumption reduces 6.74%, and the total electric cost of garden user is 1669.46 $.

3rd, load curve of the load in the case where IL and RTP excitations are participated in is as shown in figure 16.Participating in the afterload of RTP and IL Peak value is cut down to 1451.81kW, reduces 8.31%, while peak value time of occurrence is postponed to 23h by 15h.Peak-valley difference is reduced to 728.65kW, reduces 7.56%.Total electricity consumption reduces 8.35%, and the total electric cost of garden user is 1319.50 $.

The load responding situation that be can be seen that by above simulation result in this example under CPP excitations is least preferable.The feelings Peak value reduction degree is minimum under condition, and does not reach the effect for moving peak.Simultaneously load peak-valley difference increased, the total electricity consumption of user into This is also highest.Therefore CPP need to be developed programs and modify or use instead the electricity price energisation mode of TOU or RTP.In this example, Load peak reduction under TOU and RTP excitations is approximately the same, and two kinds of energisation modes all make load peak be moved after carrying out, but Peak-valley difference under RTP reduces situation and becomes apparent.Although user power utilization cost is less than RTP, the total use under RTP in the case of TOU Electricity decreasing value is better than TOU higher than the steady situation of load power curve under TOU, and RTP.In summary analyze, RTP is this example Middle optimal excitation mode, intelligent power garden can realize optimal response result under the excitation.

Claims (10)

1. intelligent power garden demand response strategy, it is comprised the following steps：

1) to the flexible load of intelligent power garden, response mode difference is classified as desired；

2) corresponding demand response Policy model is set up to different flexible loads；

3) corresponding demand response Policy model is set up to whole intelligent power garden；

4) effect of the garden user under different demands response policy is calculated.

2. intelligent power garden demand response strategy according to claim 1, it is characterised in that step 1) in, according to soft Property workload demand response characteristic, be classified as conventional load, adjustable load and the class of transferable load three.

3. intelligent power garden demand response strategy according to claim 2, it is characterised in that participate in can interrupt moderation ring The conventional load model constraints answered is：

$0\≤P_{j,t}^l\≤I_{l,j,t}{P}_j^{l,max}$

$\underset{t}{\Σ}{I}_{l,j,t}\≤X_{l,j}^{\max }$

$P_{j,t}^l\≤P_{j,t}^L$

In formula,It is the load that user j cuts off in hour t, kW；I_l,j,tFor 0,1 shape that load j cuts off in the load of hour t State variable, if user j cuts off load, I in hour t_l,j,t=1, otherwise it is 0；For user j may participate in can interrupt moderation Peak load cuts off power, kW；For user j daily participate in can interrupt moderation peak load cut off hourage, h； It is user j in the prediction load of hour t, kW.

4. intelligent power garden demand response strategy according to claim 2, it is characterised in that illuminated in adjustable load Load is to the response model constraints of Price Mechanisms：

$P_{j,t}^a=P_{j,t}^{a,base}(1-I_{a,j,t})+P_{j,t}^{a,\min }{I}_{a,j,t}$

$\mathrm{\&epsiv;}({\mathrm{\&rho;}}_t-{\mathrm{\&rho;}}_{a,j,t}^{\min })<I_{a,j,t}\&le;\mathrm{\&epsiv;}({\mathrm{\&rho;}}_t-{\mathrm{\&rho;}}_{a,j,t}^{\min })+1$

In formula,It is load j in the power of hour t, kW；It is load j in the benchmark electric power of hour t, kW； I_a,j,tLoad Regulation 0,1 state variable for load j in hour t, the I when power network electricity price is higher than electricity price threshold value_a,j,t=1, load Power reduce toOtherwise it is 0；It is load j in the minimum electric power of hour t, kW is set by the user；ε for≤ 0.001 arithmetic number；ρ_tIt is the power network electricity price of hour t, ￠/kWh；It is the electricity price threshold value of user's setting, ￠/kWh.

5. intelligent power garden demand response strategy according to claim 2, it is characterised in that air-conditioning in adjustable load Response restricted model based on Price Mechanisms is：

$T_{k,t+1}={\mathrm{\&alpha;}}_k{T}_{k,t}+(1-{\mathrm{\&alpha;}}_k)(T_{k,t}^{\mathrm{\&alpha;}}-T_{k,t}^g)$

${\mathrm{\&alpha;}}_k=e^{-\mathrm{\&Delta;}t/C_k{R}_k}$

$P_k^{\min }\&le;P_{k,t}^a\&le;P_k^{max}$

$T_{k,t}=T_{k,t}^s$

$T_{k,t}^s=T_{k,t}^{s,base}(1-I_{a,k,t})+T_{k,t}^{s,max}{I}_{a,k,t}$

$\mathrm{\&epsiv;}({\mathrm{\&rho;}}_t-{\mathrm{\&rho;}}_{a,k,t}^{\min })<I_{a,k,t}\&le;\mathrm{\&epsiv;}({\mathrm{\&rho;}}_t-{\mathrm{\&rho;}}_{a,k,t}^{\min })+1$

In formula, T_k,t、Respectively air conditioner load k hour t air-conditioner temperature and environment temperature, DEG C；α_kIt is radiating function；It is temperature gain of the air-conditioning when opening, DEG C；Δ t is spaced for control time, 1h；C_k、R_kThe respectively thermal capacitance kW ˙ of air-conditioning H/ DEG C and thermal resistance DEG C/kW, η_kIt is the operating efficiency of air-conditioning,It is air conditioner load k in the electric power of hour t, kW；Minimum per hour, the maximum electric power of respectively air-conditioning k, kW；It is the setting temperature of hour t air conditioner load Degree, DEG C；Air conditioner load k is higher than threshold value in the former design temperature and electricity price of hour t respectively under refrigerating stateWhen highest setting temperature, DEG C, be set by the user；I_a,k,tFor air conditioner load k adjusts 0,1 shape in the design temperature of hour t State variable, when power network electricity price is higher than electricity price threshold valueWhen I_a,k,t=1, air-conditioning design temperature is increased toOtherwise it is 0.

6. intelligent power garden demand response strategy according to claim 2, it is characterised in that washed the dishes in transferable load Machine, dryer are to the response model constraints of Price Mechanisms：

$P_{j,t}^b=P_j^{b,average}{I}_{b,j,t}$

$\&lsqb;X_{b,j,(t-1)}^{on}-U_{b,j}\&rsqb;\&lsqb;I_{b,j,(t-1)}-I_{b,j,t}\&rsqb;\&GreaterEqual;0$

$X_{b,j,t}^{on}=\underset{t=\mathrm{\&tau;}}{\overset{t}{\&Sigma;}}{I}_{b,j,t}$

$\underset{t=\mathrm{\&tau;}}{\overset{T}{\&Sigma;}}{I}_{b,j,t}=U_{b,j}$

In formula,It is transferable load j in the power of hour t, kW；It is the mean power of transferable load j, kW； I_b,j,tIt is transferable load j in 0,1 state variable of hour t, the I when load operation_b,j,t=1, otherwise it is 0；I_b,j,(t-1)For 0,1 state variables of the transferable load j in the previous hour of hour t；It is load j in the run time of hour t, h；It is load j in the run time of the previous hour of hour t, h；U_b,jIt is the load operation of transferable load Cycle, h；τ is that user sets permissible load operation starting time, h；T terminates run time, h for user's assumed load.

7. intelligent power garden demand response strategy according to claim 2, it is characterised in that electronic in transferable load Car load is to the response model constraints of Price Mechanisms：

$P_k^{b,\min }{I}_{b,k,t}\&le;P_{k,t}^b\&le;P_k^{b,\max }{I}_{b,k,t},(T_a\&le;t\&le;T_b)$

$E_{k,t}^b=E_{k,t-1}^b+P_{k,t}^b\mathrm{\&Delta;}t/E_k^{max}$

$E_{k,T_b}=1$

$E_{k,t}^b\&le;1$

In formula,Respectively electric automobile k minimum, maximum charge power, kW hourly；For electronic Automobile k hour t charge power, kW；I_b,k,tFor 0,1 state variable that electric automobile k charges in hour t, work as electric automobile I during charging_b,k,t=1, otherwise it is 0；T_aIt is user's last time trip end time, that is, charges the time started；T_bFor user advises The standby travel time is fixed, that is, is charged the end time；It is electric automobile k in the charged state of hour t；It is electronic vapour Charged states of the car k in the previous hour of hour t；Specify the charging of preparation travel time in user for electric automobile k State；Δ t is spaced for control time, 1h；It is batteries of electric automobile capacity, kWh.

8. intelligent power garden demand response strategy according to claim 7, it is characterised in that multiple electric automobile loads In the case of, it is considered to initial state of charge in electric automobile constraintWith user's last time trip end time T_aIt is random Property, obtain electric automobile initial state of chargeProbability density function be：

$f\left(E_{k,T_a}^b\right)=\frac{1}{\sqrt{2\mathrm{\&pi;}}D(1-E_{k,T_a}^b){\mathrm{\&sigma;}}_d}\&CenterDot;\exp \{-\frac{{\&lsqb;ln(1-E_{k,T_a}^b)+\ln D-{\mathrm{\&mu;}}_d\&rsqb;}^2}{2{\mathrm{\&sigma;}}_d^2}\}$

In formula, D is the maximum range of electric automobile, user's last time trip end time T_aThink approximately to obey normal state Distribution, σ_dThe standard deviation of distance travelled, μ_dIt is the average of the distance travelled of Normal Distribution.

9. intelligent power garden demand response strategy according to claim 1, it is characterised in that step 3) in, to whole The corresponding demand response Policy model that intelligent power garden is set up is as follows：

User side is with the power-balance constraint of grid side：

$\underset{i}{\&Sigma;}{P}_{i,t}^W+\underset{i}{\&Sigma;}{P}_{i,t}^V+\underset{i}{\&Sigma;}{P}_{i,t}^T+\underset{i}{\&Sigma;}{P}_{i,t}^S+P_{g,t}=\underset{j}{\&Sigma;}(P_{j,t}^L-P_{j,t}^l)+\underset{j}{\&Sigma;}{P}_{j,t}^a+\underset{k}{\&Sigma;}{P}_{k,t}^a+\underset{j}{\&Sigma;}{P}_{j,t}^b+\underset{k}{\&Sigma;}{P}_{k,t}^b$

In formula,It is wind power generator i in the generated output of hour t, kW；It is photovoltaic generation unit i in the generating of hour t Power, kW；It is gas turbine i in the generated output of hour t, kW；It is battery i in the charge and discharge power of hour t；P_g,t It is grid side in the output power of hour t, kW；It is user j in the prediction load of hour t, kW；It is user j in hour t When the load that cuts off, kW；It is lighting load j in the power of hour t, kW；For air conditioner load k uses electric work in hour t Rate, kW；It is the transferable load j such as dish-washing machine, dryer in the power of hour t, kW；It is electric automobile k hour t's Charge power, kW；

Grid side power constraint is：

$P_g^{\min }\&le;P_{g,t}\&le;P_g^{max}$

In formula,Respectively power network is minimum, maximum output power, and this constraint reflects power network actual power power about Beam, while in Spot Price excitation, this constraint can also prevent under the conditions of intelligent power load in the electricity price relatively low period because of load There is new load peak in transfer；

Gas turbine restricted model is：

$F_i^T\left(P_{i,t}^T\right)=(\frac{P_{i,t}^T}{{\mathrm{\&eta;}}_i^T}\&CenterDot;{\mathrm{\&rho;}}_g)+P_{i,t}^T\&CenterDot;M^T$

${\mathrm{\&eta;}}_i^T=\frac{1}{H_i/3600}$

$P_i^{T,min}{I}_{i,t}^T\&le;P_{i,t}^T\&le;P_i^{T,max}{I}_{i,t}^T$

In formula, F_i ^TIt is gas turbine power generation cost function；It is gas turbine i in the generated output of hour t, kW；It is combustion gas The generating efficiency of turbine i；ρ_gIt is Gas Prices, ￠/kWh；M^TIt is operation expense, ￠/kWh；H_iIt is gas turbine i's Heat consumption rate, kJ/kWh；P_i ^T,min、P_i ^T,maxThe respectively minimum of gas turbine i, maximum power generation；For gas turbine i exists Switching on and shutting down 0,1 state variable of hour t, when gas turbine is started shooting to be runOtherwise it is 0；

Battery energy storage restricted model is：

$P_{i,t}^S=P_{d,i,t}-P_{c,i,t}$

I_d,i,t+I_c,i,t≤1

$I_{c,i,t}{P}_{c,i}^{\min }\&le;P_{c,i,t}\&le;I_{c,i,t}{P}_{c,i}^{\max }$

$I_{d,i,t}{P}_{d,i}^{\min }\&le;P_{d,i,t}\&le;I_{d,i,t}{P}_{d,i}^{\max }$

$E_{i,t}=E_{i,(t-1)}-(P_{d,i,t}\&CenterDot;\frac{1}{{\mathrm{\&eta;}}_i^S}-{\mathrm{\&eta;}}_i^S{P}_{c,i,t})\mathrm{\&Delta;}t/E_i^{\max }$

$E_i^{\min }/E_i^{\max }\&le;E_{i,t}\&le;1$

E_i,0=E_i,NT

In formula,It is battery i in the charge and discharge power of hour t, kW；P_d,i,tIt is battery i in the discharge power of hour t, kW； P_c,i,tIt is battery i in the charge power of hour t, kW；The respectively minimum of battery i, maximum charge power, kW；I_d,i,t、I_c,i,tElectric discharge, charging 0,1 state variables of the respectively battery i in hour t, when battery i discharges in hour t, I_d,i,t=1, otherwise I_d,i,t=0, similarly, as battery i when hour t charges I_c,i,t=1, otherwise I_c,i,t=0； The respectively minimum of battery i, maximum charge power, kW；Minimum, the maximum discharge power of respectively battery i, kW；E_i,tCharged states of the battery i in hour t；It is the charge and discharge efficiency of battery i； Respectively battery i's Minimum, maximum capacity, kWh；NT is the total hourage of battery plan；E_i,(t-1)Charged states of the battery i in hour t；Δ t is control Time interval, 1h；E_i,0Battery initial state of charge；E_i,NTThe charged state of last hour of battery；

It is minimum as optimization aim using user power utilization cost from Load aggregation business's angle, set up Optimized model：

$\begin{array}{cc}Min& \underset{t}{\&Sigma;}\underset{i}{\&Sigma;}{F}_i^T\left(P_{i,t}^T\right)+\underset{t}{\&Sigma;}{\mathrm{\&rho;}}_t{P}_{g,t}-\underset{t}{\&Sigma;}\underset{j}{\&Sigma;}{\mathrm{\&rho;}}_{l,j}{P}_{j,t}^l\end{array}$

$\begin{array}{cc}s.t.& \left\{\begin{array}{c}\left(1\right)-\left(20\right)\\ \left(22\right)-\left(33\right)\end{array}\right.\end{array}$

In formula, F_i ^TIt is gas turbine power generation cost function；It is gas turbine i in the generated output of hour t, kW；P_g,tIt is electricity Net side hour t output power, kW；It is the load that user j cuts off in hour t, kW；ρ_tIt is the power network electricity of hour t Valency, ￠/kWh；ρ_l,jFor the interruptible load of load j cuts off price, ￠/kWh.

10. intelligent power garden demand response strategy according to claim 1, it is characterised in that step 4) garden user Effect computational methods under different demands response policy are as follows：It is solved using CPLEX solvers.