CN104537435A - Distributed power source optimizing configuration method based on user-side economic indexes - Google Patents

Abstract

The invention discloses a distributed power source optimizing configuration method based on user-side economic indexes and relates to the technical field of distributed power generation in electric systems. The method includes the steps that firstly, an optimization problem mathematical model is established with the minimum user cost as a target and in combination with operation constraint conditions of all units, and then benefits and costs of a user side and a power grid side are obtained; then, capacities and power of all distributed power sources are changed and then are led into the optimization model to be calculated, corresponding benefits and costs are obtained, and the benefits and costs are used as inputs and outputs to train a neural network, and therefore the optimum distributed power source capacity and power can be obtained. The method has the advantages that economic analysis and estimation methods for a distributed power source grid-connected system based on the user-side economic indexes are provided, and the optimum configuration from which users can obtain the greatest benefits is obtained; based on the method, economic estimation is also performed on a power supply side, and the optimizing problem of the system is simplified to be a linear programming problem; accordingly, computation is simplified, a relatively comprehensive and systematic economic analysis method is provided, so the method has engineering promotional value.

Description

Based on the distributed power source Optimal Configuration Method of user side economic index

Technical field

The invention belongs to electric system distributed generation technology field, be specifically related to a kind of Optimal Configuration Method of the distributed power source based on user side economic index.

Background technology

Along with becoming increasingly conspicuous of global environment and energy problem, environmental friendliness and the distributed power generation of economy are paid close attention to more.The system combined by distributed power source and energy storage device, while distributed power source generating, due to adding of energy storage device, the energy flow of system is controlled more flexible, the type of flow of energy can be controlled according to the output electric energy of distributed power source and user's electricity price situation used, such system can reduce terminal user's expense, improve the quality of power supply, safeguard the stable operation of electrical network, there is environmental friendliness simultaneously, the advantage of regeneration sustainable use is Future New Energy Source development and the important directions promoted.Because different distributed power sources has different qualities, its scope that is suitable for also different, user according to the region at place and environmental quality, need select the distributed power source access electrical network be applicable to.In Future Ten year, the small-scale distributed power source of the most applicable use in the intelligent residential district in city and small business is photovoltaic generation, miniature gas turbine.

In user side, for distributed power source access electrical network, whether interests can be brought to be problems of first concern for user, so need set of system and the Optimal Configuration Method of the convenient distributed power source calculated helps economy and the reliability of evaluation and analysis distributed power source.The domestic and international research to distributed power source aspect at present mainly concentrates on the economic load dispatching of system and the addressing constant volume aspect of power supply, lacks a set of comparison system and complete economic analysis and evaluation scheme; In optimized algorithm, mostly by being distributed power source, energy storage device founding mathematical models, minimum for objective function with the gross investment of system, with the reliability service of electrical network for constraint condition, determine dispatching method and constant volume, but the more advance focusing on algorithm of these researchs, cause calculating loaded down with trivial details and the confidence level of result is not high, engineering practice is not strong, and promotional value is not high.

Therefore, set up a set of fairly perfect, and calculate simple, be convenient to the Optimal Configuration Method of the distributed power source promoted in engineering, be very urgent for future world greatly develops new forms of energy.

For the mathematical model of photovoltaic and miniature gas turbine, can determine according to existing method.

1) photovoltaic array:

P _PV＝η*S*R*[1-0.005(T-25)]

S＝P _PVm/P _PVm1

Wherein, P _pVfor the output power of photovoltaic array, η is the specified conversion efficiency of photovoltaic, and S is the area (m of photovoltaic array ²), R is the solar irradiation intensity (W/m on photovoltaic module dip plane ²), T is photovoltaic module temperature (degree Celsius); P _pVm1for the peak power output (see manufacturer's technical parameter) of every square metre of photovoltaic; P _pVmfor the peak power output of the photovoltaic array that user adopts.

Model is visible thus, as long as the specified efficiency eta of input photovoltaic, the area S of photovoltaic array, the solar irradiation intensity R on photovoltaic module dip plane, photovoltaic module temperature T, just can obtain the output power of photovoltaic array.For the photovoltaic module of given model, its performance parameter-photovoltaic peak power P _pVmbe directly proportional to area S.(YonaA，SenjyuT，Funabashi T.Applicationofrecurrentneural networktoshort-term-aheadgeneratingpowerforecastingforphotovoltaicsystem[C].IEEEPowerEngineering SocietyGeneralMeeting，2007)。

2) miniature gas turbine

For the miniature gas turbine of given model, efficiency, specific heat capacity, loss coefficient, regenerator effectiveness, the coefficients such as die swell ratio, can obtain from the technical parameter of gas turbine, or obtain from exemplary reference value engineering, be convenient to calculate to simplify gas turbine model, compressor inlet air pressure P ₁, entering air temperature T ₁, inlet air flow rate G _k, the delivery temperature T of regenerator _4aexemplary reference value can be got.So entry of combustion chamber fuel flow rate G only need be inputted _f, just can obtain the output power P of this gas turbine _i(t) and the heat Q produced _i(t).In other words, given gas turbine model, so its generating efficiency also just determines, and drops into a certain amount of fuel to firing chamber, and the output of electricity and heat is determined, and can be represented by following linear relationship:

P _i(t)＝m×G _f(t)

$Q_i\left(t\right)=n\×G_f\left(t\right)=\frac{n}{m}{P}_i\left(t\right)$

Wherein G _ft () is entry of combustion chamber fuel flow rate; P _it output power that () is gas turbine, Q _it output heat energy that () is gas turbine, m, n are constant (being determined by miniature gas turbine model).(Ma Wanling, the research [D] of miniature gas turbine cogeneration of heat and power in distributed energy resource system, HeFei University of Technology, 2008.3 (16-26))

Neural network is a kind of imitation animal nerve network behavior feature, carries out the algorithm mathematics model of distributed parallel information processing.This network relies on the complexity of system, by adjusting interconnective relation between inner great deal of nodes, thus reaches the object of process information.Neural network training is carried out with sample, namely the structure of network is set, each neuronic tax function alive, and parameter such as training function and learning rate etc., after initialization network, neural network is input to sample input value, obtain corresponding output rear compared with the output valve of sample, roll-off network are carried out according to error, through certain number of times and the training of a period of time, namely error reaches trains complete in preset range, the neural network obtained with this can relation very accurately in analog sample between input and output.

Genetic algorithm is the mathematical model according to nature biotechnology chromosome evolution, and it imitates that chromosomal gene carries out selecting in the process of natural evolution, crossover and mutation generates population of future generation.The step of searching the operation of neural network optimal value with genetic algorithm is: dimension and the objective function of determining input variable, and the span of given input variable, determines population scale, maximum genetic algebra, optimization aim precision, and selects the parameters such as crossover operator.So just can export the optimal value of objective function and input corresponding to optimal value by the calculating of software.

Summary of the invention

The invention provides a kind of distributed power source Optimal Configuration Method based on user side economic index, the method is first minimum for target with user cost, in conjunction with each unit operation constraint condition, set up the mathematical model of optimization problem, obtain income and the cost of user side and grid side, change the capacity of each distributed power source and power again to bring in Optimized model and calculate, obtain corresponding income and cost, and in this, as input and output neural network training, thus obtain optimum distributed power source capacity and power.Advantage is: construct system optimization scheduling model, give the economic analysis based on the Grid-connected Distributed Generation Power System of user side economic index and appraisal procedure, and obtain making user to make a profit maximum allocation optimum, simultaneously on this basis also for supply side has carried out economic evaluation.Wherein the optimization problem of system is reduced to linear programming problem, enormously simplify calculating, and give relatively comprehensively and the Economic Analysis Method of system, tool engineering promotional value.

In order to achieve the above object, the present invention adopts following technical scheme:

1, based on a distributed power source Optimal Configuration Method for user side economic index, it is characterized in that, the method comprises the following step:

A, determine user, determined specification and the model of used photovoltaic, miniature gas cogeneration units, accumulator and heat storage water tank by user, the index listed by table 1 is the value determined,

Table 1 input quantity

Determine cogeneration of heat and power operational mode again, photovoltaic peak power P _pVm, miniature gas turbine installs number N, accumulator capacity with heat storage water tank capacity H ^total;

B, set up the mathematical model being minimised as target with user cost and be:

Objective function minf=f _gas+ f _grid+ f _a-Ql+ f _a-Qs+ f _a-Ps+ f _p-loss+ f _q-loss

Wherein: f represents user's operating cost of 24 hours, f _gasrepresent the gas cost that miniature gas turbine consumes for 24 hours, f _gridrepresent user's electricity charge paid to electrical network of 24 hours, f _a-Qlrepresent the punishment cost abandoning thermal load for 24 hours, f _a-Qsrepresent the punishment cost that the heat production of 24 hours is unnecessary, f _a-Psrepresent the punishment cost that the electrogenesis of 24 hours is unnecessary, f _p-lossthe punishment cost of accumulator discharge and recharge simultaneously, f _q-lossit is the punishment cost of heat storage water tank heat supply simultaneously and heat accumulation;

$f_{\mathrm{gas}}=\underset{k=1}{\overset{24/\mathrm{\ΔT}}{\mathrm{\Σ}}}{C}_n*N*G_f\left(k\right)*\mathrm{\ΔT}$

Wherein C _nrepresent the unit price of rock gas; N represents the gas turbine quantity of installation, G _fk () represents the air inflow of a kth sampling instant gas turbine, Δ T is the sampling time;

$f_{\mathrm{grid}}=\underset{k=1}{\overset{24/\mathrm{\ΔT}}{\mathrm{\Σ}}}{C}_{\mathrm{grid}}\left(k\right)*P_{\mathrm{grid}}\left(k\right)*\mathrm{\ΔT}$

Wherein C _gridk () represents the price of a kth sampling instant electrical network to user's sale of electricity, P _gridk () represents a kth electric energy that sampling instant electrical network provides to user, Δ T is the sampling time;

$f_{a-\mathrm{Ql}}=\underset{k=1}{\overset{24/\mathrm{\ΔT}}{\mathrm{\Σ}}}{C}_{a-\mathrm{Ql}}*Q_{\mathrm{abanl}}\left(k\right)*\mathrm{\ΔT}$

Wherein C _a-Qlrepresent and abandon thermal load penalty factor, Q _abanlk () represents that thermal load amount is abandoned in a kth sampling instant, Δ T is the sampling time;

$f_{a-\mathrm{Qs}}=\underset{k=1}{\overset{24/\mathrm{\ΔT}}{\mathrm{\Σ}}}{C}_{a-\mathrm{Qs}}*Q_{\mathrm{abans}}\left(k\right)*\mathrm{\ΔT}$

Wherein C _a-Qsrepresent the unnecessary penalty factor of heat production, Q _abansk () represents a kth unnecessary amount of sampling instant heat production, Δ T is the sampling time;

$f_{a-\mathrm{Ps}}=\underset{k=1}{\overset{24/\mathrm{\ΔT}}{\mathrm{\Σ}}}{C}_{a-\mathrm{Ps}}*Q_{\mathrm{abans}}\left(k\right)*\mathrm{\ΔT}$

Wherein C _a-Psrepresent the unnecessary penalty factor of electrogenesis, P _abansk () represents a kth unnecessary amount of sampling instant electrogenesis, Δ T is the sampling time;

$f_{P-\mathrm{loss}}=\underset{k=1}{\overset{24/\mathrm{\ΔT}}{\mathrm{\Σ}}}\{C_{P-\mathrm{loss}}*[(\frac{1}{{\mathrm{\η}}_2}-1)*P_{\mathrm{ba}-\mathrm{dis}}\left(k\right)+(1-{\mathrm{\η}}_1)*P_{\mathrm{ba}-\mathrm{ch}}\left(k\right)]*\mathrm{\ΔT}\}$

Wherein C _p-lossrepresent the penalty factor of the artificial accumulator arranged discharge and recharge simultaneously, P _ba-disk () represents the discharge power of a kth moment accumulator, P _ba-chk () represents the charge power of a kth moment accumulator, η ₁represent the charge efficiency of accumulator, η ₂represent the discharging efficiency of accumulator, Δ T is the sampling time;

Wherein C _p-loss=2*C _a-Ps, due to P _ba-dis(k), P _ba-chk (), as battery discharging power and charge power, must meet P in the moment _ba-dis(k) * P _ba-chk ()=0, directly calculates if this constraint condition is put into model, then this is planned to a nonlinear programming, and calculating can be very complicated, and the solution obtained also does not have reliability, but sets up if this problem be converted into

C _p-loss=2*C _a-Ps, then the unnecessary amount of electrogenesis can directly bleed off, and can not be consumed, also just ensure that P by the mode of accumulator discharge and recharge simultaneously off-energy _ba-dis(k) * P _ba-ch(k)=0.

$f_{Q-\mathrm{loss}}=\underset{k=1}{\overset{24/\mathrm{\ΔT}}{\mathrm{\Σ}}}\{C_{Q-\mathrm{loss}}*[(\frac{1}{{\mathrm{\η}}_4}-1)*Q_{\tan k-\sup }\left(k\right)+(1-{\mathrm{\η}}_3)*Q_{\tan k-\mathrm{sto}}\left(k\right)]*\mathrm{\ΔT}\}$

Wherein C _q-lossrepresent the penalty factor of the artificial heat storage water tank arranged heat accumulation and heat supply simultaneously, and C _q-loss=2*C _a-Qs;

Q _tank-supk () represents the heating power of a kth moment heat storage water tank, Q _tank-stok () represents the heat accumulation power of a kth moment heat storage water tank, η ₃represent the heat accumulation efficiency of heat storage water tank, η ₄represent the heating efficiency of heat storage water tank, Δ T is the sampling time;

Wherein C _q-loss=2*C _a-Qs.Due to Q _tank-sup(k) and Q _tank-stok (), as the heating power of heat storage water tank and heat accumulation power, must meet Q in the moment _tank-sup(k) * Q _tank-stok ()=0, directly calculates if this constraint condition is put into model, then this is planned to a nonlinear programming, and calculating can be very complicated, and the solution obtained also does not have reliability, but sets up C if this problem be converted into _q-loss=2*C _a-Qs, then the unnecessary amount of heat production can directly bleed off, and can not be consumed, just can ensure Q by the mode of heat accumulation while of heat storage water tank and this off-energy of heat supply _tank-sup(k) * Q _tank-sto(k)=0.

Constraint condition one

Electric energy balance constraint P _pV(k)+P _grid(k)+P _ba-dis(k)-P _ba-ch(k)+N*P _i(k)=P _l(k)+P _abansk () be P wherein _pVk () represents the output electric energy of a kth moment photovoltaic array prediction, P _gridk () represents a kth electric energy that moment electrical network is supplied to user, P _ba-disk () represents the discharge power of a kth moment accumulator, P _ba-chk () represents the charge power of a kth moment accumulator, N represents the gas turbine quantity of user installation, P _ik () represents the output electric energy of a kth moment single gas turbine, P _lk () represents the prediction power load of a kth moment user, P _abansk () represents a kth unnecessary amount of moment electrogenesis, N represents the gas turbine quantity of installation;

Wherein P _i(k)=m*G _f(k), G _fk () is gas turbine inlet fuel flow rate, m is the constant determined by miniature gas turbine model;

Constraint condition two

Thermal energy balance constraint N*Q _i(k)-Q _abans(k)+Q _tank-sup(k)-Q _tank-sto(k)=Q _l(k)-Q _abanl(k)

Wherein Q _ik () represents the output heat energy of a kth moment single gas turbine, Q _abansk () represents a kth unnecessary amount of moment heat production, Q _tank-supk () represents the heating power of a kth moment heat storage water tank, Q _tank-stok () represents the heat accumulation power of a kth moment heat storage water tank, Q _lk () represents the prediction thermal load of a kth moment user, Q _abanlk () represents that a kth moment abandons thermal load amount, N represents the gas turbine quantity of installation;

Wherein p _ik () represents the output electric energy of a kth moment single gas turbine; M, n are the constants determined by miniature gas turbine model;

Constraint condition three

Miniature gas turbine runs constraint

$P_i^{\min }\≤P_i\left(k\right)\≤P_i^{\max }$

$R_i^{\mathrm{down}}*\mathrm{\ΔT}\≤P_i\left(k\right)-P_i(k-1)\≤R_i^{\mathrm{up}}*\mathrm{\ΔT}$

Wherein with represent the minimum of gas turbine operation and peak power restriction respectively, P _ik () represents the output power in a gas turbine kth moment, represent the upper and lower creep speed restriction of gas turbine respectively;

Constraint condition four

Accumulator runs constraint

E _s(k)＝E _s(k-1)+P _ba-ch(k)*η ₁*ΔT-P _ba-dis(k)/η ₂*ΔT

$0\≤P_{\mathrm{ba}-\mathrm{ch}}\left(k\right)\≤P_s^{\max }$

$0\≤P_{\mathrm{ba}-\mathrm{dis}}\left(k\right)\≤P_s^{\max }$

$E_s^{\mathrm{total}}*{\mathrm{SOC}}_{\min }\≤E_s\left(k\right)\≤E_s^{\mathrm{total}}*{\mathrm{SOC}}_{\max }$

$E_s^0={\mathrm{SOC}}_{\min }*E_s^{\mathrm{total}}$

Wherein E _sk () represents the dump energy of a kth moment accumulator, P _ba-disk () represents the discharge power of a kth moment accumulator, P _ba-chk () represents the charge power of a kth moment accumulator, represent the peak power restriction that accumulator runs, η ₁represent the charge efficiency of accumulator, η ₂represent the discharging efficiency of accumulator, represent the maximum energy storage of accumulator, SOC _minand SOC _maxrepresent the minimum and maximum energy storage state of accumulator respectively, represent the initial energy storage state of accumulator;

Constraint condition five

Heat storage water tank runs constraint

H(k)＝H(k-1)+Q _tank-sto(k)*η ₃*ΔT-Q _tank-sup(k)/η ₄*ΔT

$Q_{\tan k}^{\min }\≤Q_{\tan k-\mathrm{sto}}\left(k\right)\≤Q_{\tan k}^{\max }$

$Q_{\tan k}^{\min }\≤Q_{\tan k-\sup }\left(k\right)\≤Q_{\tan k}^{\max }$

H ^total*SOT _min≤H(k)≤H ^total*SOT _max

H ⁰＝SOT _min*H ^total

Wherein H (k) represents the dump energy of a kth moment heat storage water tank, Q _tank-supk () represents the heating power of a kth moment heat storage water tank, Q _tank-stok () represents the heat accumulation power of a kth moment heat storage water tank, with represent the minimum and maximum operate power of heat storage water tank respectively, H ^totalrepresent the maximum stored energy capacitance of heat storage water tank, SOT _minand SOT _maxrepresent the minimum and maximum energy storage state of heat storage water tank respectively, H ⁰represent the initial energy storage state of heat storage water tank;

Constraint condition six

The constraint of electrical network supply electric energy

Wherein P _gridk () represents a kth electric energy that moment electrical network is supplied to user, represent the peak power restriction of electrical network supply;

Constraint condition seven

Miniature gas turbine is grid-connected two kinds of operational modes: electricity determining by heat pattern and do not limit thermoelectricity pattern.Electricity determining by heat pattern wherein for grid-connected cogeneration of heat and power has Q _abanl(k)=0

Wherein Q _abanlk () represents that a kth moment abandons thermal load amount;

Solve the objective function meeting above-mentioned 7 constraint conditions, the Optimized Operation scheme of system every day can be obtained;

Determine cogeneration of heat and power operational mode, photovoltaic peak power P _pVm, miniature gas turbine installs number N, accumulator capacity with heat storage water tank capacity H ^total, solve the objective function meeting above-mentioned 7 constraint conditions, the Optimized Operation scheme of system every day can be obtained;

The income f of c, user side ₁computing formula as follows:

$f_1=\underset{k=1}{\overset{876/\mathrm{\ΔT}}{\mathrm{\Σ}}}(P_l\left(k\right)*C_{\mathrm{grid}}\left(k\right)+P_{\mathrm{pv}}\left(k\right)*C_{\mathrm{pvsub}})*\mathrm{\ΔT}-(f_{\mathrm{gas}}+f_{\mathrm{grid}}+f_{a-\mathrm{Ql}}+f_{a-\mathrm{Qs}}+f_{a-\mathrm{Ps}})$

Wherein f ₁represent user's income, P _lk () represents the prediction power load of a kth sampling instant user, C _pvsubrepresent the subsidy of photovoltaic generation, P _pvk () represents the generated output of photovoltaic, C _gridk () represents the price of a kth sampling instant electrical network to user's sale of electricity; f _gasrepresent the gas cost that miniature gas turbine consumes; f _gridrepresent the electricity charge that user pays to electrical network; f _a-Qlrepresent the punishment cost abandoning thermal load; f _a-Qsthe punishment cost that the heat production represented is unnecessary; f _a-Psrepresent the punishment cost that electrogenesis is unnecessary; Δ T is the sampling time;

D, the investment of user side and cost f ₂computing formula as follows:

f ₂＝f _PV/n _PV+f _gas-turbine/n _i+f _tank/n _tank+f _battery/n _ba

Wherein f ₂represent the customer investment of amounting to into a year; f _pv=C _pv* P _pvmrepresent the installation price of photovoltaic array; C _pvrepresent the price of photovoltaic, P _pvmrepresent the dressing amount of photovoltaic, n _pVrepresent the tenure of use of photovoltaic, f _gas-turbine=C _i* N represents the installation price of miniature gas turbine, C _irepresent the price of gas turbine, N represents the number of units of miniature gas turbine; n _irepresent the tenure of use of miniature gas turbine, f _tank=C _tank* H ^totalrepresent the installation price of heat storage water tank, C _tankrepresent the unit price of heat storage water tank, H ^totalrepresent the capacity of heat storage water tank, n _tankrepresent the tenure of use of heat storage water tank, represent the installation price of accumulator, C _barepresent the unit price of accumulator, represent the capacity of accumulator, n _barepresent the tenure of use of accumulator;

E, grid side income f ₃computing formula as follows:

$\mathrm{\Δ}{P}_l^{\max }=P_l^{\max }-P_{\mathrm{grid}}^{\max }$

$f_3=\mathrm{\Δ}{P}_1^{\max }*C_{\mathrm{tran}}$

Wherein f ₃represent electrical network income, represent the difference of the peak load of distributed power source grid-connected front and back electrical network, represent the grid-connected front peak load of distributed power source, peak load after expression distributed power source is grid-connected, C _tranthe investment of indication transformer every kilovolt-ampere;

F, grid side income reduction and cost f ₄computing formula as follows:

$f_4=\underset{k=1}{\overset{8760/\mathrm{\ΔT}}{\mathrm{\Σ}}}[C_{\mathrm{grid}}\left(k\right)-C_{\mathrm{plant}-\mathrm{grid}}\left(k\right)]*[P_l\left(k\right)-P_{\mathrm{grid}}\left(k\right)]$

Wherein f ₄represent electrical network income reduction, C _plant-gridk () represents thermal power plant's rate for incorporation into the power network, C _gridk () represents the electricity price of electrical network to user's sale of electricity, P _lk () represents user's electric load, P _gridk () represents the load that the grid-connected rear electrical network of distributed power source is supplied to user, k is sampling instant, and Δ T is the sampling time;

When distributed power source combination: photovoltaic peak power P _pVm, miniature gas turbine installs number N, accumulator capacity heat storage water tank capacity H ^totalafter determining, the income of available user side and grid side and cost f1, f2, f3, f4;

According to the following step, adopt capacity and the power combination of different distributed power sources:

$0\≤P_{\mathrm{PVm}}\≤P_{\mathrm{PVm}}^{\max }$

Wherein P _pVmthe upper limit for the maximal value of user power utilization load;

0≤N≤N ^max

The wherein upper limit of N for the peak power output of miniature gas turbine, [] expression rounds the numerical value inside bracket;

Wherein the upper limit $E_s^{\mathrm{total}\max }=0.3*\frac{1}{365}\underset{k=1}{\overset{8760/\mathrm{\ΔT}}{\mathrm{\Σ}}}{P}_l\left(k\right)*\mathrm{\ΔT},$ P _lk power load value that () is each sampling instant, Δ T is the sampling time;

0≤H ^total≤H ^totalmax

The upper limit wherein $H^{\mathrm{total}\max }=0.3*\frac{1}{365}\underset{k=1}{\overset{8760/\mathrm{\ΔT}}{\mathrm{\Σ}}}{Q}_l\left(k\right)*\mathrm{\ΔT},$ Q _lk () is the thermic load value of each sampling instant, Δ T is the sampling time;

P is changed according to the following step _pVm, N, h ^totalvalue, obtain income and the cost f1 of corresponding user side and grid side, f2, f3, f4;

Step1.k＝1；

Step2.i1＝1；

Step3.i2＝1；

Step4.i3＝1；

Step5.i4＝1；

Step6. $P_{\mathrm{PVm}}\left(k\right)=(i1-1)*P_{\mathrm{PVm}}^{\max }/20;$

N(k)＝i2；

$E_s^{\mathrm{total}}\left(k\right)=(i3-1)*E_s^{\mathrm{total}\max }/20;$

H ^total(k)＝(i4-1)*H ^totalmax/20；

K＝k+1；

By P _pVm(k), N (k), h ^totalk () substitutes in step b-f, obtain corresponding f1 (k), f2 (k), f3 (k), the value of f4 (k);

Step7.i4＝i4+1；

If i4≤21, then return step6;

Step8.i3＝i3+1；

If i3≤21, then return step5;

Step9.i2＝i2+1；

If i2≤N ^max, then step4 is returned;

Step10.i 1＝i 1+1；

If i1≤21, then return step3;

Obtain above 21 ⁴* N ^maxgroup (P _pVm, N, h ^total, f1, f2, f3, f4) and vector carrys out neural network training as the training sample of neural network, and the neural network obtained can represent input value (P _pVm, N, h ^total) and the relation of output valve (f1, f2, f3, f4);

Searched the globally optimal solution of neural network by genetic algorithm, thus obtain optimum distributed power source capacity and power.

This method is first minimum for target with user cost, in conjunction with each unit operation constraint condition, set up the mathematical model of optimization problem, obtain income and the cost of user side and grid side, change the capacity of each distributed power source and power again to bring in Optimized model and calculate, obtain corresponding income and cost, and in this, as input and output neural network training, thus obtain optimum distributed power source capacity and power.

Technique effect of the present invention is: construct system optimization scheduling model, give the economic analysis based on the Grid-connected Distributed Generation Power System of user side economic index and appraisal procedure, and obtain making user to make a profit maximum allocation optimum, simultaneously on this basis also for supply side has carried out economic evaluation.Wherein the optimization problem of system is reduced to linear programming problem, enormously simplify calculating, and give relatively comprehensively and the Economic Analysis Method of system, have engineering promotional value.

Accompanying drawing explanation

Fig. 1 is the schematic diagram of present system structure.

Fig. 2 is the schematic diagram of algorithm flow of the present invention.

Embodiment

In order to describe the present invention more specifically, below in conjunction with accompanying drawing, for the Grid-connected Distributed Generation Power System of certain industry and commerce user, grid-connected distributed power supply system structured flowchart as shown in Figure 1, carry out analyzing to its economy and assess, and obtaining allocation optimum, concrete steps are as follows:.

1) mathematical model and the resource data of distributed power source is determined

For photovoltaic generation, obtain this area photometric data R of a year and temperature data T; Photovoltaic cost: C _pv=10yuan/w, conversion ratio η=10%, every square meter photovoltaic peak power output P _pVm1=0.07 life-span n _pvabout 20 years.

For miniature gas turbine, according to actual conditions determination parameter, obtain air inflow and the relation between generated output and heat production power, i.e. P _i(t)=m*G _f(t) and Q _i(t)=n*G _f(t), wherein P _it () represents the generated output of gas turbine, Q _it () represents the heating power of waste heat boiler, G _ft () represents the air inflow of gas turbine, m, n are constant.To Capstone C65 miniature gas turbine, can obtain according to its model: P _i(t)=12057.38986*G _f(t), Q _i(t)=20718.71506*G _f(t), and the bound of MT operate power $P_i^{\min }=14\mathrm{KW},P_i^{\max }=65\mathrm{KW},$ Upper and lower creep speed restriction meets $R_i^{\mathrm{down}}=5\mathrm{kw}/\min ,R_i^{\mathrm{up}}=10\mathrm{kw}/\min .$ Price Ci is 500,000 yuan/platform, life-span ni about 20 years.

Charge in batteries efficiency eta ₁=0.9, discharging efficiency η ₂=0.8, charge-discharge electric power limits energy storage state bound SOC _min=0.2, SOC _max=0.8, initial energy storage state price C _babe 500 yuan/kw, life-span n _baabout 2 years.

Heat storage water tank heat accumulation efficiency eta ₃=0.9, heating efficiency η ₄=0.8, heat accumulation and heating power limit water tank energy storage state bound SOT _min=0.2, SOT _max=0.8, initial energy storage state H ⁰=SOT _min* H ^total.Price C _tankfor 21.5/kw, life-span n _tankabout 10 years.

For grid side: thermal power plant rate for incorporation into the power network C _plant-grid=0.3 yuan, transformer capacity invests in maintenance cost C _tran=600yuan/kVA.

Obtain user 1 year prediction electric load P _lwith prediction thermal load data Q _l.

2) certainty annuity parameter and cogeneration of heat and power pattern

For certain concrete system, its user side parameter: sampling time Δ T=1h, Gas Prices Cn=2.718yuan/kg, tou power price C _gridas shown in table 1, abandon thermal load penalty factor _a-Ql=0.46, the unnecessary penalty factor of heat production _a-Qs=0.25, the unnecessary penalty factor of electrogenesis _a-ps=0.5, the penalty factor of heat storage water tank heat accumulation and heat supply simultaneously _p-loss=2*C _a-ps=1.The penalty factor of energy storage device discharge and recharge simultaneously _q-loss=2*C _a-Qs=0.5.

Cogeneration of heat and power is not to limit thermoelectricity mode operation.

Table 1 tou power price

Period (hour)	01-08	09-10	11-12	13-18	19-22
						Electricity price (unit)	0.2616	0.5450	0.981	0.5450	0.981

3) Optimal Operation Model of every day is:

Objective function:

$\min C=\underset{k=1}{\overset{24}{\mathrm{\Σ}}}\left[\begin{array}{c}{C}_n*N*G_f\left(k\right)*\mathrm{\ΔT}+C_{\mathrm{grid}}\left(k\right)*P_{\mathrm{grid}}\left(k\right)*\mathrm{\ΔT}+C_{a-\mathrm{Ql}}*Q_{\mathrm{abanl}}\left(k\right)*\mathrm{\ΔT}\\ +C_{a-\mathrm{Qs}}*Q_{\mathrm{abans}}\left(k\right)*\mathrm{\ΔT}+C_{a-\mathrm{ps}}*P_{\mathrm{abans}}\left(k\right)*\mathrm{\ΔT}\\ +C_{P-\mathrm{loss}}*[(\frac{1}{{\mathrm{\η}}_2}-1)*P_{\mathrm{ba}-\mathrm{dis}}\left(k\right)+(1-{\mathrm{\η}}_1)*P_{\mathrm{ba}-\mathrm{ch}}\left(k\right)]*\mathrm{\ΔT}\\ +C_{Q-\mathrm{loss}}*[(\frac{1}{{\mathrm{\η}}_4}-1)*Q_{\tan k-\sup }\left(k\right)+(1-{\mathrm{\η}}_3)*Q_{\tan k-\mathrm{sto}}\left(k\right)]\end{array}\right]$

Constraint condition:

P _PV(k)+P _grid(k)+P _ba-dis(k)-P _ba-ch(k)+N*P _i(k)＝P _l(k)+P _abans(k)

N*Q _i(k)-Q _abans(k)+Q _tank-sup(k)-Q _tank-sto(k)＝Q _l(k)-Q _abanl(k)

P _i(t)＝m×G _f(t)

$Q_i\left(t\right)=n\×G_f\left(t\right)=\frac{n}{m}{P}_i\left(t\right)$

E _s(k)＝E _S(k-1)+P _ba-ch(k)*η ₁*ΔT-P _ba-dis(k)/η ₂*ΔT

H(k)＝H(k-1)+Q _tank-sto(k)*η ₃*ΔT-Q _tank-sup(k)/η ₄*ΔT

$R_i^{\mathrm{down}}*\mathrm{\ΔT}\≤P_i\left(k\right)-P_i(k-1)\≤R_i^{\mathrm{up}}*\mathrm{\ΔT}$

$0\≤P_{\mathrm{grid}}\left(k\right)\≤P_{\mathrm{grid}}^{\max }$

$P_i^{\min }\≤P_i\left(k\right)\≤P_i^{\max }$

$0\≤P_{\mathrm{ba}-\mathrm{ch}}\left(k\right)\≤P_S^{\max }$

$0\≤P_{\mathrm{ba}-\mathrm{dis}}\left(k\right)\≤P_S^{\max }$

$E_S^{\mathrm{total}}*{\mathrm{SOC}}_{\min }\≤E_S\left(k\right)\≤E_S^{\mathrm{total}}*{\mathrm{SOC}}_{\max }$

$Q_{\tan k}^{\min }\≤Q_{\tan k-\mathrm{sto}}\left(k\right)\≤Q_{\tan k}^{\max }$

$Q_{\tan k}^{\min }\≤Q_{\tan k-\sup }\left(k\right)\≤Q_{\tan k}^{\max }$

H ^total*SOT _min≤H(k)≤H ^total*SOT _max

3) user's income and the investment of 1 year is calculated by the scheduling result of every day, and the income of grid side and investment:

User's annual earnings:

$f_1=\underset{k=1}{\overset{8760/\mathrm{\ΔT}}{\mathrm{\Σ}}}(P_l\left(k\right)*C_{\mathrm{grid}}\left(k\right)+P_{\mathrm{pv}}\left(k\right)*C_{\mathrm{pvsub}})*\mathrm{\ΔT}-(f_{\mathrm{gas}}+f_{\mathrm{grid}}+f_{a-\mathrm{Ql}}+f_{a-\mathrm{Qs}}+f_{a-\mathrm{Ps}})$

User's year invests:

f ₂＝f _PV/n _PV+f _gas-turbine/n _i+f _tank/n _tank+f _battery/n _ba

Grid side annual earnings:

$f_3=\mathrm{\Δ}{P}_l^{\max }*C_{\mathrm{tran}}$

Grid side year invests:

$f_4=\underset{k=1}{\overset{8760/\mathrm{\ΔT}}{\mathrm{\Σ}}}[C_{\mathrm{grid}}\left(k\right)-C_{\mathrm{plant}-\mathrm{grid}}\left(k\right)]*[P_l\left(k\right)-P_{\mathrm{grid}}\left(k\right)]$

4) with different distributed power source combination (capacity, power): photovoltaic peak power P _pVm, miniature gas turbine installs number N, energy-storage battery capacity heat storage water tank capacity H ^totalas the input of model in step 1, with the income of the corresponding user side that obtains and grid side and cost f1, f2, f3, f4 are that output carrys out neural network training, then search globally optimal solution, i.e. optimum power configuration by the neural network trained.Algorithm flow chart as shown in Figure 2.

4) above-mentioned parameter is brought in model calculate, obtain the optimum combination of distributed power source, and user's income of correspondence and investment, electrical network income and investment.

Calculating this allocation optimum is:

Photovoltaic peak power P _pVm=352kw;

Miniature gas turbine number N=43;

Energy-storage battery capacity E _s=870kwh;

Heat storage water tank capacity H=2552kwh.

Corresponding output is (1 year):

User income f ₁=253.43 ten thousand yuan

Customer investment f ₂=131.56 ten thousand yuan

Electrical network income f ₃=229.96 ten thousand yuan

Electric grid investment f ₄=139.42 ten thousand yuan

User net proceeds f1-f2=121.87 ten thousand yuan

5) economy of allocation optimum combination

Change the configuration of distributed installed capacity, calculate the net proceeds of user, compare with the user's net proceeds under optimum installed capacity

As photovoltaic peak power P _pVm=500kw;

Miniature gas turbine number N=50;

Energy-storage battery capacity E _s=870kwh;

Heat storage water tank capacity H=2552kwh.

Corresponding output is (1 year):

User income f ₁=260.21 ten thousand yuan

Customer investment f ₂=149.80 ten thousand yuan

Now user net proceeds f1-f2=116.41 ten thousand yuan

As photovoltaic peak power P _pVm=352kw;

Miniature gas turbine number N=50;

Energy-storage battery capacity E _s=1000kwh;

Heat storage water tank capacity H=2552kwh.

Corresponding output is (1 year):

User income f ₁=262.87 ten thousand yuan

Customer investment f ₂=152.31 ten thousand yuan

Now user net proceeds f1-f2=110.56 ten thousand yuan

As photovoltaic peak power P _pVm=200kw;

Miniature gas turbine number N=50;

Energy-storage battery capacity E _s=1000kwh;

Heat storage water tank capacity H=2000kwh.

Corresponding output is (1 year):

User income f ₁=264.87 ten thousand yuan

Customer investment f ₂=160.44 ten thousand yuan

Now user net proceeds f1-f2=104.43 ten thousand yuan

Can see, after changing the installed capacity of distributed power source, the net proceeds of user is all little than the user's net proceeds under optimum installed capacity, thus demonstrates its economy.

Claims (1)

1., based on a distributed power source Optimal Configuration Method for user side economic index, it is characterized in that, the method comprises the following step:

A, determine user, determined specification and the model of used photovoltaic, miniature gas cogeneration units, accumulator and heat storage water tank by user, then the index listed by table 1 is the value determined,

Table 1 input quantity

Determine cogeneration of heat and power operational mode again, photovoltaic peak power P _pVm, miniature gas turbine installs number N, accumulator capacity with heat storage water tank capacity H ^total;

B, set up the mathematical model being minimised as target with user cost and be:

Objective function minf=f _gas+ f _grid+ f _a-Ql+ f _a-Qs+ f _a-Ps+ f _p-loss+ f _q-loss

Wherein: f represents user's operating cost of 24 hours, f _gasrepresent the gas cost that miniature gas turbine consumes for 24 hours, f _gridrepresent user's electricity charge paid to electrical network of 24 hours, f _a-Qlrepresent the punishment cost abandoning thermal load for 24 hours, f _a-Qsrepresent the punishment cost that the heat production of 24 hours is unnecessary, f _a-Psrepresent the punishment cost that the electrogenesis of 24 hours is unnecessary, f _p-lossthe punishment cost of accumulator discharge and recharge simultaneously, f _q-lossit is the punishment cost of heat storage water tank heat supply simultaneously and heat accumulation;

$f_{\mathrm{gas}}=\underset{k=1}{\overset{24/\mathrm{\ΔT}}{\mathrm{\Σ}}}{C}_n*N*G_f\left(k\right)*\mathrm{\ΔT}$

Wherein C _nrepresent the unit price of rock gas; N represents the gas turbine quantity of installation, G _fk () represents the air inflow of a kth sampling instant gas turbine, Δ T is the sampling time;

$f_{\mathrm{grid}}=\underset{k=1}{\overset{24/\mathrm{\ΔT}}{\mathrm{\Σ}}}{C}_{\mathrm{grid}}\left(k\right)*P_{\mathrm{grid}}\left(k\right)*\mathrm{\ΔT}$

Wherein C _gridk () represents the price of a kth sampling instant electrical network to user's sale of electricity, P _gridk () represents a kth electric energy that sampling instant electrical network provides to user, Δ T is the sampling time;

$f_{a-\mathrm{Ql}}=\underset{k=1}{\overset{24/\mathrm{\ΔT}}{\mathrm{\Σ}}}{C}_{a-\mathrm{Ql}}*Q_{\mathrm{abanl}}\left(k\right)*\mathrm{\ΔT}$

Wherein C _a-Qlrepresent and abandon thermal load penalty factor, Q _abanlk () represents that thermal load amount is abandoned in a kth sampling instant, Δ T is the sampling time;

$f_{a-\mathrm{Qs}}=\underset{k=1}{\overset{24/\mathrm{\ΔT}}{\mathrm{\Σ}}}{C}_{a-\mathrm{Qs}}*Q_{\mathrm{abans}}\left(k\right)*\mathrm{\ΔT}$

Wherein C _a-Qsrepresent the unnecessary penalty factor of heat production, Q _abansk () represents a kth unnecessary amount of sampling instant heat production, Δ T is the sampling time;

$f_{a-\mathrm{Ps}}=\underset{k=1}{\overset{24/\mathrm{\ΔT}}{\mathrm{\Σ}}}{C}_{a-\mathrm{Ps}}*P_{\mathrm{abans}}\left(k\right)*\mathrm{\ΔT}$

Wherein C _a-Psrepresent the unnecessary penalty factor of electrogenesis, P _abansk () represents a kth unnecessary amount of sampling instant electrogenesis, Δ T is the sampling time;

$f_{P-\mathrm{loss}}=\underset{k=1}{\overset{24/\mathrm{\ΔT}}{\mathrm{\Σ}}}\{C_{P-\mathrm{loss}}*[(\frac{1}{{\mathrm{\η}}_2}-1)*P_{\mathrm{ba}-\mathrm{dis}}\left(k\right)+(1-{\mathrm{\η}}_1)*P_{\mathrm{ba}-\mathrm{ch}}\left(k\right)]*\mathrm{\ΔT}\}$

Wherein C _p-lossrepresent the penalty factor of the artificial accumulator arranged discharge and recharge simultaneously, and C _p-loss=2*C _a-Ps; P _ba-disk () represents the discharge power of a kth moment accumulator, P _ba-chk () represents the charge power of a kth moment accumulator, η ₁represent the charge efficiency of accumulator, η ₂represent the discharging efficiency of accumulator, Δ T is the sampling time;

$f_{Q-\mathrm{loss}}=\underset{k=1}{\overset{24/\mathrm{\ΔT}}{\mathrm{\Σ}}}\{C_{Q-\mathrm{loss}}*[(\frac{1}{{\mathrm{\η}}_4}-1)*Q_{\tan k-\sup }\left(k\right)+(1-{\mathrm{\η}}_3)*Q_{\tan k-\mathrm{sto}}\left(k\right)]*\mathrm{\ΔT}\}$

Wherein C _q-lossrepresent the penalty factor of the artificial heat storage water tank arranged heat accumulation and heat supply simultaneously, and C _q-loss=2*C _a-Qs; Q _tank-supk () represents the heating power of a kth moment heat storage water tank, Q _tank-stok () represents the heat accumulation power of a kth moment heat storage water tank, η ₃represent the heat accumulation efficiency of heat storage water tank, η ₄represent the heating efficiency of heat storage water tank, Δ T is the sampling time;

Constraint condition one

Electric energy balance constraint P _pV(k)+P _grid(k)+P _ba-dis(k)-P _ba-ch(k)+N*P _i(k)=P _l(k)+P _abansk () be P wherein _pVk () represents the output electric energy of a kth moment photovoltaic array prediction, P _gridk () represents a kth electric energy that moment electrical network is supplied to user, P _ba-disk () represents the discharge power of a kth moment accumulator, P _ba-chk () represents the charge power of a kth moment accumulator, N represents the gas turbine quantity of user installation, P _ik () represents the output electric energy of a kth moment single gas turbine, P _lk () represents the prediction power load of a kth moment user, P _abansk () represents a kth unnecessary amount of moment electrogenesis, N represents the gas turbine quantity of installation;

Wherein P _i(k)=m*G _f(k), G _fk () is gas turbine inlet fuel flow rate, m is the constant determined by miniature gas turbine model;

Constraint condition two

Thermal energy balance constraint N*Q _i(k)-Q _abans(k)+Q _tank-sup(k)-Q _tank-sto(k)=Q _l(k)-Q _abanlk () be Q wherein _ik () represents the output heat energy of a kth moment single gas turbine, Q _abansk () represents a kth unnecessary amount of moment heat production, Q _tank-supk () represents the heating power of a kth moment heat storage water tank, Q _tank-stok () represents the heat accumulation power of a kth moment heat storage water tank, Q _lk () represents the prediction thermal load of a kth moment user, Q _abanlk () represents that a kth moment abandons thermal load amount, N represents the gas turbine quantity of installation;

Wherein p _ik () represents the output electric energy of a kth moment single gas turbine; M, n are the constants determined by miniature gas turbine model;

Constraint condition three

Miniature gas turbine runs constraint

$P_i^{\min }\≤P_i\left(k\right)\≤P_i^{\max }$

$-R_i^{\mathrm{down}}*\mathrm{\ΔT}\≤P_i\left(k\right)-P_i(k-1)\≤R_i^{\mathrm{up}}*\mathrm{\ΔT}$

Wherein with represent the minimum of gas turbine operation and peak power restriction respectively, P _ik () represents the output power in a gas turbine kth moment, represent the upper and lower creep speed restriction of gas turbine respectively;

Constraint condition four

Accumulator runs constraint

E _s(k)＝E _s(k-1)+P _ba-ch(k)*η ₁*ΔT-P _ba-dis(k)/η ₂*ΔT

$0\≤P_{\mathrm{ba}-\mathrm{ch}}\left(k\right)\≤P_s^{\max }$

$0\≤P_{\mathrm{ba}-\mathrm{dis}}\left(k\right)\≤P_s^{\max }$

$E_s^{\mathrm{total}}*{\mathrm{SOC}}_{\min }\≤E_s\left(k\right)\≤E_s^{\mathrm{total}}*{\mathrm{SOC}}_{\max }$

$E_s^0={\mathrm{SOC}}_{\min }*E_s^{\mathrm{total}}$

Wherein E _sk () represents the dump energy of a kth moment accumulator, P _ba-disk () represents the discharge power of a kth moment accumulator, P _ba-chk () represents the charge power of a kth moment accumulator, represent the peak power restriction that accumulator runs, η ₁represent the charge efficiency of accumulator, η ₂represent the discharging efficiency of accumulator, represent the maximum energy storage of accumulator, SOC _minand SOC _maxrepresent the minimum and maximum energy storage state of accumulator respectively, represent the initial energy storage state of accumulator;

Constraint condition five

Heat storage water tank runs constraint

H(k)＝H(k-1)+Q _tank-sto(k)*η ₃*ΔT-Q _tank-sup(k)/η ₄*ΔT

$Q_{\tan k}^{\min }\≤Q_{\tan k-\mathrm{sto}}\left(k\right)\≤Q_{\tan k}^{\max }$

$Q_{\tan k}^{\min }\≤Q_{\tan k-\sup }\left(k\right)\≤Q_{\tan k}^{\max }$

H ^total*SOT _min≤H(k)≤H ^total*SOT _max

H ⁰＝SOT _min*H ^total

Wherein H (k) represents the dump energy of a kth moment heat storage water tank, Q _tank-supk () represents the heating power of a kth moment heat storage water tank, Q _tank-stok () represents the heat accumulation power of a kth moment heat storage water tank, with represent the minimum and maximum operate power of heat storage water tank respectively, H ^totalrepresent the maximum stored energy capacitance of heat storage water tank, SOT _minand SOT _maxrepresent the minimum and maximum energy storage state of heat storage water tank respectively, H ⁰represent the initial energy storage state of heat storage water tank;

Constraint condition six

The constraint of electrical network supply electric energy $0\≤P_{\mathrm{grid}}\left(k\right)\≤P_{\mathrm{grid}}^{\max }$

Wherein P _gridk () represents a kth electric energy that moment electrical network is supplied to user, represent the peak power restriction of electrical network supply;

Constraint condition seven

Electricity determining by heat pattern for grid-connected cogeneration of heat and power has Q _abanl(k)=0

Wherein Q _abanlk () represents that a kth moment abandons thermal load amount;

Solve the objective function meeting above-mentioned 7 constraint conditions, the Optimized Operation scheme of system every day can be obtained;

The income f of c, user side ₁computing formula as follows:

$f_1=\underset{k=1}{\overset{8760/\mathrm{\ΔT}}{\mathrm{\Σ}}}(P_1\left(k\right)*C_{\mathrm{grid}}\left(k\right)+P_{\mathrm{pv}}\left(k\right)*C_{\mathrm{pvsub}})*\mathrm{\ΔT}-(f_{\mathrm{gas}}+f_{\mathrm{grid}}+f_{a-\mathrm{Ql}}+f_{a-\mathrm{Qs}}+f_{a-\mathrm{Ps}})$

Wherein f ₁represent user's income, P _lk () represents the prediction power load of a kth sampling instant user, C _pvsubrepresent the subsidy of photovoltaic generation, P _pvk () represents the generated output of photovoltaic, C _gridk () represents the price of a kth sampling instant electrical network to user's sale of electricity; f _gasrepresent the gas cost that miniature gas turbine consumes; f _gridrepresent the electricity charge that user pays to electrical network; f _a-Qlrepresent the punishment cost abandoning thermal load; f _a-Qsthe punishment cost that the heat production represented is unnecessary; f _a-Psrepresent the punishment cost that electrogenesis is unnecessary; Δ T is the sampling time;

D, the investment of user side and cost f ₂computing formula as follows:

f ₂＝f _PV/n _PV+f _gas-turbine/n _i+f _tank/n _tank+f _battery/n _ba

Wherein f ₂represent the customer investment of amounting to into a year; f _pv=C _pv* P _pvmrepresent the installation price of photovoltaic array; C _pvrepresent the price of photovoltaic, P _pvmrepresent the dressing amount of photovoltaic, n _pVrepresent the tenure of use of photovoltaic, f _gas-turbine=C _i* N represents the installation price of miniature gas turbine, C _irepresent the price of gas turbine, N represents the number of units of miniature gas turbine; n _irepresent the tenure of use of miniature gas turbine, f _tank=C _tank* H ^totalrepresent the installation price of heat storage water tank, C _tankrepresent the unit price of heat storage water tank, H ^totalrepresent the capacity of heat storage water tank, n _tankrepresent the tenure of use of heat storage water tank, represent the installation price of accumulator, C _barepresent the unit price of accumulator, represent the capacity of accumulator, n _barepresent the tenure of use of accumulator;

E, grid side income f ₃computing formula as follows:

$\mathrm{\Δ}{P}_l^{\max }=P_l^{\max }-P_{\mathrm{grid}}^{\max }$

$f_3={\mathrm{\ΔP}}_l^{\max }*C_{\mathrm{tran}}$

Wherein f ₃represent electrical network income, represent the difference of the peak load of distributed power source grid-connected front and back electrical network, represent the grid-connected front peak load of distributed power source, peak load after expression distributed power source is grid-connected, C _tranthe investment of indication transformer every kilovolt-ampere;

F, grid side income reduction and cost f ₄computing formula as follows:

$f_4=\underset{k=1}{\overset{8760/\mathrm{\ΔT}}{\mathrm{\Σ}}}[C_{\mathrm{grid}}\left(k\right)-C_{\mathrm{plant}-\mathrm{grid}}\left(k\right)]*[P_l\left(k\right)-P_{\mathrm{grid}}\left(k\right)]$

Wherein f ₄represent electrical network income reduction, C _plant-gridk () represents thermal power plant's rate for incorporation into the power network, C _gridk () represents the electricity price of electrical network to user's sale of electricity, P _lk () represents user's electric load, P _gridk () represents the load that the grid-connected rear electrical network of distributed power source is supplied to user, k is sampling instant, and Δ T is the sampling time;

When distributed power source combination: photovoltaic peak power P _pVm, miniature gas turbine installs number N, accumulator capacity heat storage water tank capacity H ^totalafter determining, the income of available user side and grid side and cost f1, f2, f3, f4;

According to the following step, adopt capacity and the power combination of different distributed power sources:

$0\≤P_{\mathrm{PVm}}\≤P_{\mathrm{PVm}}^{\max }$

Wherein P _pVmthe upper limit for the maximal value of user power utilization load;

0≤N≤N ^max

The wherein upper limit of N for the peak power output of miniature gas turbine, [] expression rounds the numerical value inside bracket;

$0\≤E_s^{\mathrm{total}}\≤E_s^{\mathrm{total}\max }$

Wherein the upper limit $E_s^{\mathrm{total}\max }=0.3*\frac{1}{365}\underset{k=1}{\overset{8760/\mathrm{\ΔT}}{\mathrm{\Σ}}}{P}_l\left(k\right)*\mathrm{\ΔT},$ P _lk power load value that () is each sampling instant, Δ T is the sampling time;

0≤H ^total≤H ^totalmax

The upper limit wherein $H^{\mathrm{total}\max }=0.3*\frac{1}{365}\underset{k=1}{\overset{8760/\mathrm{\ΔT}}{\mathrm{\Σ}}}{Q}_l\left(k\right)*\mathrm{\ΔT},$ Q _lk () is the thermic load value of each sampling instant, Δ T is the sampling time;

P is changed according to the following step _pVm, N, h ^totalvalue, obtain income and the cost f1 of corresponding user side and grid side, f2, f3, f4;

Step1.k＝1；

Step2.i 1＝1；

Step3.i2＝1；

Step4.i3＝1；

Step5.i4＝1；

$\mathrm{Step}6.P_{\mathrm{PVm}}\left(k\right)=(i1-1)*P_{\mathrm{PVm}}^{\max }/20;$

N(k)＝i2；

$E_s^{\mathrm{total}}\left(k\right)=(i3-1)*E_s^{\mathrm{total}\max }/20;$

H ^total(k)＝(i4-1)*H ^totalmax/20；

K＝k+1；

By P _pVm(k), N (k), h ^totalk () substitutes in step b-f, obtain corresponding f1 (k), f2 (k), f3 (k), the value of f4 (k);

Step7.i4＝i4+1；

If i4≤21, then return step6;

Step8.i3＝i3+1；

If i3≤21, then return step5;

Step9.i2＝i2+1；

If i2≤N ^max, then step4 is returned;

Step10.i1＝i1+1；

If i1≤21, then return step3;

Obtain above 21 ⁴* N ^maxgroup (P _pVm, N, h ^total, f1, f2, f3, f4) and vector carrys out neural network training as the training sample of neural network, and the neural network obtained can represent input value (P _pVm, N, h ^total) and the relation of output valve (f1, f2, f3, f4);

Searched the globally optimal solution of neural network by genetic algorithm, thus obtain optimum distributed power source capacity and power.