CN103500361A - Micro-grid load control method based on game theory - Google Patents

Abstract

Provided is a micro-grid load control method based on the game theory. The micro-grid load control method based on the game theory comprises the following steps of (1) determining the basic idea of the game theory, and building a revenue function; (2) setting up a micro-grid model, and determining clear definitions of game elements in a micro-grid; (3) building revenue functions and solution algorithms of factories, participating in the game, in the micro-grid; (4) carrying out designing on implementation of the game theory in micro-grid load control.

Description

Based on game theoretic microgrid duty control method

Technical field

Project of the present invention relates to a kind of microgrid duty control method, particularly a kind of control method based on theory of games.

Background technology

The generation of microgrid, on the one hand, improved the utilization of regenerative resource, and to realizing energy-saving and emission-reduction, positive role plays in friendly environment society; On the other hand, improve reliability and the security of major network, further enlarged the area coverage of power network, greatly improved life and production environment.Micro-network load is controlled and is optimized the importance of distributing as embodying micro-electrical network economy, causes gradually people's concern.Along with microgrid further develops and applies, commercial power is also changing its model of purchase for electricity gradually, from traditional major network power purchase, transfer to more economic microgrid power purchase, yet, the microgrid scale is limited, micro-electric generation investment cost is excessive, and, as the main market players of profit-push income, certainly will there be the situation of competition electricity consumption in factory.Though industrial load takies the ratio that the family sum is very little, formed most power load, socioeconomic impact is seemed to particularly important.Therefore, how realizing reasonable control and the management of industrial load, is the focus that current commercial power is paid close attention to.Yet current economic optimization algorithms more used, as all there are some weak point in particle cluster algorithm, genetic algorithm and chaos algorithm etc.Chaos algorithm utilizes method of weighting that multiple goal is converted into to single goal and is optimized, and exists weight coefficient to be difficult to determine, system results is affected to excessive problem; Though particle cluster algorithm and genetic algorithm etc. can improve speed of searching optimization, still be difficult to guarantee the global convergence of gained solution.Theory of games is as a class advanced person mathematical tool, be intended to solve the relation of vying each other between different subjects, the gained Nash Equilibrium Solution is the mutual peak optimization reaction in game side, according to different degree of cooperation weight coefficients, determine the best power consumption of each factory, thereby reach the purpose that instructs the micro-power supply of microgrid reasonable arrangement to exert oneself, realize energy-saving and emission-reduction, increase social benefit.At present, the economy of microgrid is moved mainly for the capacity of micro-power supply and is exerted oneself and is optimized, and seldom relates to the reasonable control and management of power load in microgrid.

Summary of the invention

The present invention will overcome that existing microgrid finite capacity, cost of investment are excessive, factory's electricity consumption problem with keen competition in system, a kind of microgrid duty control method based on theory of games is proposed, by analyzing power consumption and the income of each factory under non-cooperative game and cooperative game, reaching, instruct microgrid to exert oneself and the purpose of factory's rational utilization of electricity.

Based on game theoretic microgrid duty control method, comprise the following steps:

Step 1, the basic thought of determining theory of games and structure revenue function;

Step 2, build the microgrid model, determine the clearly definition of each game key element in microgrid;

Step 3, set up revenue function and derivation algorithm that each factory in microgrid participates in game;

Step 4, design realize the realization of game theory in the microgrid load is controlled.

Further, determine in step (1) that game theory basic thought and structure revenue function step are as follows:

1-1), the definition of clearly game theoretic key concept and Nash Equilibrium Solution;

1-2), the key distinction of the classification of clear and definite game and different classes of game;

1-3), according to different classes of game, construct corresponding revenue function, wherein, non-cooperative game is only considered self benefits, comprise self and integral benefit in cooperative game, and, because of the degree of cooperation difference, revenue function is corresponding difference also;

Further, step (2) can be comprised of following step:

2-1), build the microgrid model, select sun power, wind energy, fuel cell as main micro-power supply, and comprise battery pack, play the function of stabilizing the system power fluctuation; Choose three kinds of different factories as the system power load;

2-2), game factor analysis: in the present invention, factory is considered as to the game participant, power consumption is game strategies, and factory account is corresponding income, and balance policy is resulting optimum power consumption after system gaming;

Further, setting up each factory in microgrid in step (2) participates in the revenue function of game and designs derivation algorithm being comprised of following step:

3-1), determine the revenue function of each factory;

Further, for step (3-1), determining step is as follows:

A1), the design factory earnings function, its income is mainly production marketing;

A2), design factory pays the function of the electricity charge, the electricity charge adopt two electricity prices processed of national regulation to be calculated;

A3), design plant maintenance function, comprise oil plant and the auxiliary expenses of repair spare part expense, labour cost, maintenance;

A4), design plant produced Master Cost.

3-2), the derivation algorithm of design problem of game.

Further, for step (3-2), the derivation algorithm step is as follows:

B1), the initial possible strategy s of random generation in each game strategies space ₀={ s ₁₀, s ₂₀... s _m0;

B2), note s _i-1strategy set for other game sides except game person i.Any i (i=1,2 ..., n) game person, with this game person's income u _ifor target, fixing s _i-1constant, belonging to the policy space S of this game side _iinside carry out single goal optimization, ask best decision optimization game person income;

B3), make strategy combination

for the result after game optimization, check s ₁feasibility, if do not met, transfer step 2 to); If meet, whether the distance (a kind of norm) of calculating between former and later two strategy combinations meets convergence criterion || s ₁-s ₀||≤ε, if meet, game finishes; If do not meet, with s ₁replace s ₀, transfer step 3) to and carry out iterative loop;

B4), draw stable Nash equilibrium solution.

Further, the performing step of step (4) is as follows:

4-1), determine the effective information of each factory, the formulation of clearly corresponding cost function;

4-2), analyze power consumption and the situation of Profit of each factory under different situations;

Further, the performing step in step (4-2) is as follows:

C1), at first analyze power consumption and the situation of Profit of each factory under non-cooperative game, draw corresponding change curve and stable Nash Equilibrium Solution;

C2), for certain, definite cooperation weight coefficient is analyzed the corresponding situation of each factory under cooperative game, and compare with the result in step (C1);

C3), analyze power consumption and the situation of Profit of each factory under different cooperation weight coefficients;

C4), for above analysis result, formulate corresponding power consumption information, for instructing factory's rational utilization of electricity.

Technical conceive of the present invention is: power consumption industrial load large at most and to social influence in microgrid, as the main market players that asks most social output, certainly will cause the game situation of competition electricity consumption.Traditional Multipurpose Optimal Method, as all there are some weak point in particle cluster algorithm, genetic algorithm and chaos algorithm etc., and in system, number one is pursued by each factory, therefore belongs to the category that single goal is optimized.The present invention utilizes the thought of theory of games to carry out single goal optimization, at first objective analysis the game situation of outwardness in micro-grid system, and the key element that participates in of clear and definite each game; For different factories, formulated corresponding revenue function; The derivation algorithm of game has been proposed simultaneously.The algorithm of carrying according to the present invention, can effectively be solved the problem of game existed in microgrid.

Advantage of the present invention is: according to the power consumption problem of game of outwardness in microgrid, propose the method for theory of games, by integrating plant information, setting up corresponding revenue function, and, in conjunction with the game derivation algorithm of carrying, obtain the optimal strategy of each factory; Institute of the present invention extracting method is simple and practical, to instructing the microgrid load, rationally controls to the thinking made new advances.

The accompanying drawing explanation

Fig. 1 is micro-grid system model of the present invention

Fig. 2 is that game of the present invention solves flow process

Fig. 3 is factory's power consumption curve under non-cooperative game of the present invention

Fig. 4 is each factory's power consumption curve under cooperative game of the present invention

Fig. 5 is the impact of different weight coefficient of the present invention on each factory's power consumption

Fig. 6 is the variation of system total electricity consumption and total revenue under different weight coefficient of the present invention

Fig. 7 is the change of each factory account rate under different weight coefficient of the present invention

Fig. 8 is each factory's user message table of the present invention

Fig. 9 is the equilibrium solution under non-cooperation of the present invention and cooperative game

Figure 10 is that each factory account rate of the present invention changes

embodiment

With reference to accompanying drawing: based on game theoretic microgrid duty control method, comprise the following steps:

1), determine basic thought and the structure revenue function of theory of games;

1-1), the definition of clearly game theoretic key concept and Nash Equilibrium Solution;

1-2), the key distinction of the classification of clear and definite game and different classes of game;

1-3), according to different classes of game, construct corresponding revenue function, wherein, non-cooperative game is only considered self benefits, comprise self and integral benefit in cooperative game, and, because of the degree of cooperation difference, revenue function is corresponding difference also, and in the present invention, revenue function is designated as

the absolute benefit of self while for game side i, taking certain action strategy,

the income of other game sides while for game side i, taking action strategy.

for weight coefficient, its value is reacted cooperation and the degree of contention between each game person, w _iivalue is large, means that the degree of cooperation is low, the competition degree is high; Work as w _ii=1 o'clock, mean between each game person only to exist competition, the cooperative game model deteriorates to the non-cooperative game model;

2), build the microgrid model as shown in Figure 1.Determine the clearly definition of each game key element in microgrid, step is as follows:

2-1), build the microgrid model, select sun power, wind energy, fuel cell as main micro-power supply, and comprise battery pack, play the function of stabilizing the system power fluctuation; Choose three kinds of different factories as the system power load;

2-2), the game factor analysis: in the present invention, factory is considered as to the game participant, is designated as i, i=1,2 ... n; Power consumption is game strategies, is designated as S _i, S={S ₁, S ₂... S _n; Factory account is corresponding income, is designated as u _i, i=1,2 ... n; Balance policy is resulting optimum power consumption after system gaming;

3), set up revenue function and the derivation algorithm of each factory's participation game in microgrid:

3-1), determine and the revenue function of each factory be designated as u _i=u _iSEL-C _iCHR-C _iMAT-C _iMAI, i=1,2 ... n;

Further, for step (3-1), determining step is as follows:

A1), the design factory earnings function, its income is mainly production marketing, is designated as u _iSEL=φ _is _i, φ _ithe degree electricity output value for different factories;

A2), design factory pays the function of the electricity charge, the electricity charge adopt two electricity prices processed of national regulation to be calculated, and are designated as C _iCHR=(ε C _iTRAN+ ρ S _i) (1+ α _i), wherein

ε is the expense that the transformer unit capacity pays per month.Base price by different electricity consumptions area regulation is collected; C _iTRANcapacity for factory's installation transformer; The unit electricity price that ρ is micro-grid system, C (S ') is the energy cost consumed in S ' time for power consumption, b _ifor the maximum electricity capacity of each factory, by factory's transformer capacity and system power factor, determined, for cost coefficient; S _imeritorious consumption for factory's electricity consumption; α _ifor the coefficient of rewards and punishment of power supply department, when the power factor of industrial load does not reach the system regulation, need punish accordingly that the award of part is arranged while reaching the regulation power factor, the standard that this coefficient is 0.9 by the power factor of national regulation is carried out value;

A3), design plant maintenance function, comprise and oil plant and the auxiliary expenses of repair spare part expense, labour cost, maintenance be designated as C _iMAI, for convenience of calculating, the maintenance cost of FU hour is reduced to definite value;

A4), design plant produced Master Cost, be designated as C _iMAT=ω _is _i, ω _ifor spending electric material consumption value.

3-2), the design problem of game derivation algorithm, as shown in Figure 2, step is as follows for algorithm flow:

B1), the initial possible strategy s of random generation in each game strategies space ₀={ s ₁₀, s ₂₀... s _m0;

B2), note s _i-1strategy set for other game sides except game person i.Any i (i=1,2 ..., n) game person, with this game person's income u _ifor target, fixing s _i-1constant, belonging to the policy space S of this game side _iinside carry out single goal optimization, ask best decision

optimization game person income;

B3), make strategy combination

for the result after game optimization, check s ₁feasibility, if do not met, transfer step 2 to); If meet, whether the distance (a kind of norm) of calculating between former and later two strategy combinations meets convergence criterion || s ₁-s ₀||≤ε, if meet, game finishes; If do not meet, with s ₁replace s ₀, transfer step 3) to and carry out iterative loop;

B4), draw stable Nash equilibrium solution.

4), design realizes the realization of game theory in the microgrid load is controlled:

4-1), determine the effective information of each factory, the formulation of clearly corresponding cost function;

4-2), analyze power consumption and the situation of Profit of each factory under different situations;

Further, the performing step in step (4-2) is as follows:

C1), at first analyze power consumption and the situation of Profit of each factory under non-cooperative game, draw corresponding change curve and stable Nash Equilibrium Solution;

C2), for certain, definite cooperation weight coefficient is analyzed the corresponding situation of each factory under cooperative game, and compare with the result in step (C1);

C3), analyze power consumption and the situation of Profit of each factory under different cooperation weight coefficients;

C4), for above analysis result, formulate corresponding power consumption information, for instructing factory's rational utilization of electricity.

1. case analysis

In microgrid, micro-source comprises sun power, wind energy, fuel cell and battery pack, and betting model comprises three different factories.Wherein, the information of each factory as shown in Figure 8, unit capacity transformer basic charge per month is taken as 30 yuan, due to microgrid internal loading electricity consumption dog-eat-dog, along with increasing of power consumption, corresponding unit electricity price also increases thereupon, in the present invention, the unit's of regulation electricity price and total electricity consumption are linear, in this analysis case, the lattice factor of fixing the price coefficient is 1/500, $\mathrm{\ρ}=\frac{1}{500}{S}^{\′},$ $S^{\′}=\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{S}_i,i=\mathrm{1,2,3}.$

A), case 1

When the system weight coefficient value is 1, mean between each factory that, fully in competitive position, each game person be take the self benefits optimum as purpose, by rational decision making, draws optimum power consumption separately.Suppose that each factory's power consumption initial value is all 100kW, after 6 take turns Optimized Iterative, the power consumption of each factory remains stable, has convergence, by institute of the present invention extracting method, draws each factory's power consumption change curve under non-cooperative game, as shown in Figure 3.Each factory takes the rear stable electricity consumption strategy of convergence, can reach the maximal value of income separately, furthermore simultaneously, be that each factory does not all have wish to change its behavioral strategy, can obtain better income, therefore, the solution that under perfect competition, non-cooperative game obtains is the Nash equilibrium solution.The power consumption of stable lower each factory is respectively 213kW, 407kW, 577kW; Income is respectively 57.28 yuan/hour separately, and 231.69 yuan/hour, 516.68 yuan/hour.

B), case 2

When the weight coefficient is not equal to 1, mean between different factories to exist cooperation, there is some cooperative constraint agreement, make each factory break noncooperative deadlock, move towards the road of cooperation.This analyzes cooperation weight coefficient w in case _iivalue is 0.7, and by the equilibrium solution under non-cooperative game, as each game person's initial policy, the algorithm of carrying according to the present invention, fix all the other game person's strategies, carries out the power consumption that single goal is optimized each factory.Relation between each factory's power consumption change curve and iteration rounds as shown in Figure 4.After 7 take turns Optimized Iterative, each factory's power consumption of system keeps stable, reaches convergence.While stablizing, the power consumption of each factory is respectively 271kW, 364kW, 430kW, and the income of each factory is respectively 152.87 yuan/hour, 292.23 yuan/hour, 451.52 yuan/hour.Fig. 9 is the equilibrium solution under non-cooperation and cooperative game, as can be known from Fig. 9 the cooperation weight coefficient be 0.7 o'clock system total electricity consumption than under non-cooperation, reducing 132kW, total revenue increases by 90.97 yuan, non-cooperative game relatively, total revenue has improved 11.3%.In fact, enterprise take that to pursue number one be criterion, known by analyzing, the financial value of industry 3 under cooperative game is less than the value under non-cooperative game, whether under cooperation, there is the infringement of number one, therefore, this paper introduces the concept of rate of profit in analysis, and the rate of profit of each factory changes as shown in figure 10.By Figure 10 analysis, drawn, under cooperative game, the rate of profit of each factory all increases under non-cooperative game relatively, and there is increase in self benefits relatively.

Analyzed from case 2, in the system cooperating situation, can not only improve factory's self benefits, also improved the whole system income simultaneously.Therefore, there is the intention of the cooperation moved towards in each factory.In case 2, analysis result is that present case is analyzed the impact of different weight coefficients on system in the situation that cooperation weight coefficient value is 0.7.Fig. 5 provides the impact of different weight coefficients on each factory's power consumption, analyze known, when the weight coefficient is less than 0.5, part factory power consumption is zero, because needs pay basic charge as per installed capacity and equipment maintenance cost etc., factory, in lossing state, therefore, is less than 0.5 o'clock each factory at the cooperation weight coefficient and can select cooperation.By the further known system total electricity consumption of Fig. 6, from total revenue, under different degree of cooperations, be different, on the whole, between each factory at non-cooperative game, be under perfect competition, the total revenue that system obtains is minimum, along with the intensification of degree of cooperation, the system total electricity consumption reduces, and total revenue increases.Analyze knownly, the cooperation weight coefficient is 0.5 o'clock, system total revenue maximum.Fig. 7 has further analyzed the variation of different factory account rates when weight coefficient changes.Identical, the factory account rate also increases along with the intensification of degree of cooperation, at weight coefficient, is 0.5 o'clock maximum.Further analysis is known, and semiworks participates in cooperation and can impel rate of profit own that large change occurs, and increases considerably its benefit output value, relies on large factory, participates in cooperation, should be the approach of following midget plant seeking development.

By above case explanation, theory of games can solve the problem of microgrid internal loading reasonable distribution.According to plant layout, capacity and corresponding expense of bearing etc., draw the optimum electricity consumption strategy under stable Nash Equilibrium meaning by derivation algorithm that the present invention carries, acquired results can be used to instruct the planned capacity of micro-power supply, improves reasonable utilization, energy-saving and emission-reduction, the increase contribution to society rate of the energy.

The described content of this instructions case study on implementation is only enumerating the way of realization of inventive concept; protection scope of the present invention should not be regarded as only limiting to the concrete form that case study on implementation is stated, protection scope of the present invention also reaches conceives the equivalent technologies means that can expect according to the present invention in those skilled in the art.

Claims (5)

1. based on game theoretic microgrid duty control method, comprise the following steps:

Step 1, the basic thought of determining theory of games and structure revenue function;

Step 2, build the microgrid model, determine the clearly definition of each game key element in microgrid;

Step 3, set up revenue function and derivation algorithm that each factory in microgrid participates in game;

Step 4, design realize the realization of game theory in the microgrid load is controlled.

2. the method for claim 1 is characterized in that: determine in step 1 that game theory basic thought and structure revenue function step are as follows:

1-1), the definition of clearly game theoretic key concept and Nash Equilibrium Solution;

1-2), the key distinction of the classification of clear and definite game and different classes of game;

1-3), according to different classes of game, construct corresponding revenue function, wherein, non-cooperative game is only considered self benefits, comprise self and integral benefit in cooperative game, and, because of the degree of cooperation difference, revenue function is corresponding difference also; Revenue function is designated as

the absolute benefit of self while for game side i, taking certain action strategy,

the income of other game sides while for game side i, taking action strategy.

for weight coefficient, its value is reacted cooperation and the degree of contention between each game person, w _iivalue is large, means that the degree of cooperation is low, the competition degree is high; Work as w _ii=1 o'clock, mean between each game person only to exist competition, the cooperative game model deteriorates to the non-cooperative game model.

3. method as claimed in claim 2, it is characterized in that: step 2 is comprised of following step:

2-1), build the microgrid model, select sun power, wind energy, fuel cell as main micro-power supply, and comprise battery pack, play the function of stabilizing the system power fluctuation; Choose three kinds of different factories as the system power load;

2-2), the game factor analysis: in the present invention, factory is considered as to the game participant, is designated as i, i=1,2 ... n; Power consumption is game strategies, is designated as S _i, S={S ₁, S ₂... S _n; Factory account is corresponding income, is designated as u _i, i=1,2 ... n; Balance policy is resulting optimum power consumption after system gaming.

4. method as claimed in claim 3 is characterized in that: setting up revenue function and the design derivation algorithm that each factory in microgrid participates in game in step 2 can be comprised of following step:

3-1), determine and the revenue function of each factory be designated as u _i=u _iSEL-C _iCHR-C _iMAT-C _iMAI, i=1,2 ... n; Specifically comprise:

A1), the design factory earnings function, its income is mainly production marketing, is designated as u _iSEL=φ _is _i, φ _ithe degree electricity output value for different factories;

A2), design factory pays the function of the electricity charge, the electricity charge adopt two electricity prices processed of national regulation to be calculated, and are designated as C _iCHR=(ε C _iTRAN+ ρ S _i) (1+ α _i), wherein 0≤S _i≤ b _i, i=1,2 ... n,

ε is the expense that the transformer unit capacity pays per month.Base price by different electricity consumptions area regulation is collected; C _iTRANcapacity for factory's installation transformer; The unit electricity price that ρ is micro-grid system, C (S ') is the energy cost consumed in S ' time for power consumption, b _ifor the maximum electricity capacity of each factory, by factory's transformer capacity and system power factor, determined,

for cost coefficient; S _imeritorious consumption for factory's electricity consumption; α _ifor the coefficient of rewards and punishment of power supply department, when the power factor of industrial load does not reach the system regulation, need punish accordingly that the award of part is arranged while reaching the regulation power factor, the standard that this coefficient is 0.9 by the power factor of national regulation is carried out value;

A3), design plant maintenance function, comprise and oil plant and the auxiliary expenses of repair spare part expense, labour cost, maintenance be designated as C _iMAI, for convenience of calculating, the maintenance cost of FU hour is reduced to definite value;

A4), design plant produced Master Cost, be designated as C _iMAT=ω _is _i, ω _ifor spending electric material consumption value;

3-2), the design problem of game derivation algorithm, the specific algorithm step is as follows:

B1), the initial possible strategy s of random generation in each game strategies space ₀={ s ₁₀, s ₂₀... s _m0;

B2), note s _i-1strategy set for other game sides except game person i.Any i (i=1,2 ..., n) game person, with this game person's income u _ifor target, fixing s _i-1constant, belonging to the policy space S of this game side _iinside carry out single goal optimization, ask best decision

optimization game person income;

B3), make strategy combination

for the result after game optimization, check s ₁feasibility, if do not met, transfer step 2 to); If meet, whether the distance (a kind of norm) of calculating between former and later two strategy combinations meets convergence criterion || s ₁-s ₀||≤ε, if meet, game finishes; If do not meet, with s ₁replace s ₀, transfer step 3) to and carry out iterative loop;

B4), draw stable Nash equilibrium solution.

5. method as claimed in claim 4 is characterized in that:, the performing step of step 4 is as follows:

4-1), determine the effective information of each factory, the formulation of clearly corresponding cost function;

4-2), analyze power consumption and the situation of Profit of each factory under different situations, specifically comprise:

C1), at first analyze power consumption and the situation of Profit of each factory under non-cooperative game, draw corresponding change curve and stable Nash Equilibrium Solution;

C2), for certain, definite cooperation weight coefficient is analyzed the corresponding situation of each factory under cooperative game, and compare with the result in step (C1);

C3), analyze power consumption and the situation of Profit of each factory under different cooperation weight coefficients;

C4), for above analysis result, formulate corresponding power consumption information, for instructing factory's rational utilization of electricity.

Images