CN106485339A

CN106485339A - A kind of part throttle characteristics of power system determines method and system

Info

Publication number: CN106485339A
Application number: CN201510548903.8A
Authority: CN
Inventors: 沈良雄; 徐从谦; 耿辉
Original assignee: CRRC Dalian R&D Co Ltd
Current assignee: CRRC Dalian R&D Co Ltd
Priority date: 2015-08-31
Filing date: 2015-08-31
Publication date: 2017-03-08
Also published as: WO2017035964A1

Abstract

The present invention provides a kind of part throttle characteristics of power system to determine method and system, and the method includes：Carry out the load model parameters identification of power system using S type inertia weight particle cluster algorithm, then set up the load model of power system according to the load model parameters picking out, and determine the part throttle characteristics of power system according to this load model.The part throttle characteristics of power system provided in an embodiment of the present invention determines method and system, improve the precision of load model parameters identification by a kind of follow-on PSO, pick out optimal load model parameters, the part throttle characteristics that load model such that it is able to set up accurate description power system load characteristic carries out power system determines, improves accuracy and the effectiveness of power system load characteristic research.

Description

A kind of part throttle characteristics of power system determines method and system

Technical field

The present invention relates to Operation of Electric Systems, emulation and analytical technology, more particularly, to a kind of power system Part throttle characteristics determines method and system.

Background technology

In recent years, China Express Railway cause has obtained swift and violent development, high-speed railway traction power supply load There is the features such as impact is strong, higher hamonic wave is enriched.For its aggregate power load characteristic of accurate description, The load model structure of reasonable can be passed through and obtain accurate model parameter, set up for power train The load model of operation, emulation and the stability analysis of system.Wherein, parameter identification is the key of load modeling Step, its result can directly affect the accuracy of model, therefore explores a kind of effective parameter identification side Method is significant.

Presently used load model parameters discrimination method mainly has the mathematical method of routine and intelligent optimization to calculate This two big class of method.Conventional mathematical method has calculating speed faster, but it to the continuity of a function, Nonconvex property, the calculating of differentiability have high requirement, and it also exists and is easy to be absorbed in locally optimal solution etc. Shortcoming.Intelligent optimization algorithm is processing non-linear, the optimization problem upper body such as multivariate, discontinuous, non-convex Reveal very strong optimizing ability.

Wherein, particle swarm optimization algorithm (Particle Swarm Optimization, abbreviation PSO) is in recent years A kind of new optimized algorithm growing up, it adopts parallel search mechanism, by the transmission of optimal information, Make population Fast Convergent, eventually find optimal solution, be a kind of global search method based on swarm intelligence. Similar with other heuritic approaches (as genetic algorithm, ant group algorithm), it is also from RANDOM SOLUTION, lead to Cross iteration and find optimal solution, and evaluate the quality of solution by fitness, but the rule of PSO is more simple, Need the parameter of adjustment less, facilitate implementation.However, PSO there is also, ergodic is not enough, be easily absorbed in office The shortcomings of portion is optimum be not so that search precision is high.

Content of the invention

The embodiment of the present invention provides a kind of part throttle characteristics of power system to determine method and system, by one kind Follow-on PSO improves the precision of load model parameters identification, picks out optimal load model parameters, The load that load model such that it is able to set up accurate description power system load characteristic carries out power system is special Property determine.

The part throttle characteristics of power system provided in an embodiment of the present invention determines method, including：

According to the constraints of load model parameters, one number of particles of random initializtion is the population of N G, wherein each particle are the array of d load model parameters of an inclusion；

According toCarry out particle group hunting, wherein, I=1,2 ... ..., N,WithFor carrying out the position of particle i and speed during kth time particle group hunting,WithFor carrying out the position of particle i and speed after kth time particle group hunting,For carrying out The global optimum of described population G during k particle group hunting,Search for carrying out kth time population The individual optimal value of particle i, c during rope₁For the first Studying factors, c₂For the second Studying factors, r₁、r₂ For the random number between [0,1],For carrying out inertia weight value during kth time particle group hunting, and carry out In population search procedure, inertia weight value w of particle cluster algorithm is S-type with population searching times k Successively decrease；

When reaching pre-conditioned, stop described particle group hunting, and determined according to Search Results described negative Lotus model parameter；

Set up the load model of power system according to described load model parameters；

Determine the part throttle characteristics of described power system according to described load model.

The part throttle characteristics of power system provided in an embodiment of the present invention determines system, can be used for realizing above-mentioned The part throttle characteristics of power system determines method, and this system includes：Load model parameters determining unit, load Model sets up unit and part throttle characteristics determining unit.

Wherein, load model parameters determining unit can be used for：According to the constraints of load model parameters, One number of particles of random initializtion is population G of N, and wherein each particle is that inclusion d is negative The array of lotus model parameter；According toCarry out grain Subgroup is searched for, wherein, i=1,2 ... ..., N,WithDuring for carrying out kth time particle group hunting The position of particle i and speed,WithFor carrying out the position of particle i after kth time particle group hunting And speed,For carrying out the global optimum of described population G during kth time particle group hunting, For carrying out the individual optimal value of particle i during kth time particle group hunting, c₁For the first Studying factors, c₂For Second Studying factors, r₁、r₂For the random number between [0,1],For carrying out kth time particle group hunting When inertia weight value, and carry out in population search procedure, inertia weight value w of particle cluster algorithm with The S-type change of population searching times k；When reaching pre-conditioned, stop particle group hunting, and root Determine load model parameters according to Search Results.

Described load model is set up unit and be can be used for：Set up power system according to described load model parameters Load model.

Described part throttle characteristics determining unit can be used for：Described power system is determined according to described load model Part throttle characteristics.

Based on above-mentioned, the part throttle characteristics of power system provided in an embodiment of the present invention determines method and system, Improve the precision of load model parameters identification by a kind of follow-on PSO, pick out optimal load Model parameter, the load model such that it is able to set up accurate description power system load characteristic carries out power train The part throttle characteristics of system determines, improves accuracy and the effectiveness of power system load characteristic research.Meanwhile, The scheduling that accurate part throttle characteristics prediction can help to electrical network carries out electric power allotment, because electric energy is difficult to Substantial amounts of storage, therefore has good load model, and that is, accurately the coupling of model parameter can be saved greatly The resource of amount, can effectively carry out the pool with distribution that generates electricity, have certain economic benefit.

Brief description

In order to be illustrated more clearly that the present invention or technical scheme of the prior art, below will to embodiment or In description of the prior art the accompanying drawing of required use be briefly described it should be apparent that, below describe In accompanying drawing be some embodiments of the present invention, for those of ordinary skill in the art, do not paying On the premise of creative work, other accompanying drawings can also be obtained according to these accompanying drawings.

Fig. 1 is the principle schematic of power system load model；

Fig. 2 is the equivalent circuit schematic diagram of induction machine；

Fig. 3 is the part throttle characteristics flow chart that determines method of power system provided in an embodiment of the present invention；

Fig. 4 is the parameter identification principle schematic of load model；

Fig. 5 is the image comparison schematic diagram of the inertia weight function that linear decrease and S type successively decrease；

Fig. 6 is that the part throttle characteristics of power system provided in an embodiment of the present invention determines system schematic；

The characteristic Simulation schematic diagram of the active power that Fig. 7 absorbs when running for power system load；

The characteristic Simulation schematic diagram of the reactive power that Fig. 8 absorbs when running for power system load.

Specific embodiment

Purpose, technical scheme and advantage for making the embodiment of the present invention are clearer, below in conjunction with this Accompanying drawing in bright embodiment, is clearly and completely described to the technical scheme in the embodiment of the present invention, Obviously, described embodiment is a part of embodiment of the present invention, rather than whole embodiments.It is based on Embodiment in the present invention, those of ordinary skill in the art are obtained under the premise of not paying creative work The every other embodiment obtaining, broadly falls into the scope of protection of the invention.

All electrical equipments of power system are generically and collectively referred to as load, industrial load, agricultural can be divided into negative by purposes Lotus, Commercial Load etc., can be divided into induction conductivity, synchronous motor, illumination to set by the type of electrical equipment Standby, air-conditioning equipment etc..When load operation, the active and reactive power that it absorbs can be with load bus On voltage and the fluctuation of frequency and change, this is referred to as the voltage of load, frequency characteristic, in order to describe The math equation of this characteristic of load is referred to as load model.Load group would generally be regarded as an entirety, Fig. 1 is the principle schematic of electric integrated system loading model, as shown in figure 1, by busbar voltage U and , as system input quantity, the active-power P of load absorption and reactive power Q are defeated as it for system frequency f Output.

Due to induction machine (also known as asynchronous machine) be in aggregate power load most common be also accounting example Big load, therefore would generally select the load model as power system for the model of induction machine.Fig. 2 For the equivalent circuit schematic diagram of induction machine, as shown in Fig. 2 R_sAnd X_sIt is respectively the resistance of stator winding And reactance, X_mMutual induction reactance (also known as the reactance of Exciting Windings for Transverse Differential Protection) for stator winding and rotor windings, R_rWith X_rIt is respectively resistance and the reactance of rotor windings.The mathematical model of asynchronous machine can be expressed as：

In equation,For the busbar voltage of asynchronous machine,For the transient potential of asynchronous machine,For different The bus current of step motor,For stator open circuit time constant, X is the synchronous reactance of asynchronous machine, X ' is the transient state reactance of asynchronous machine, and H is the inertia time constant of asynchronous machine, and s is asynchronous machine Slippage.T_mFor the load torque of motor, T_m=T_L(A(1-s)²+ B (1-s)+C), T_LAsynchronous machine negative Lotus coefficient, T_eThe electromagnetic torque producing for motor, whereinX=X_s+X_m,A+B+C=1..

In the research process of asynchronous machine, on all four based on the magnetomotive force producing under different coordinates Principle, that is, the stator and rotor field synchronous rotation of motor, can set up one and have synchronous rotary speed Rotating coordinate system, this rotating coordinate system is referred to as dq rotating coordinate system.Fasten in dq rotational coordinates, institute There is the signal of telecommunication can be described as constant, facilitate the research of motor problem.

Fasten in dq rotational coordinates, U_d、U_q For the busbar voltage of d axle, q axle, E '_d、E′_qFor the transient potential of d axle, q axle, i_d、i_qFor d axle, q The stator current of axle.Equivalent circuit in conjunction with asynchronous machine and the mathematical model of asynchronous machine, can arrange Draw：

Fasten in dq rotational coordinates, the voltage equation of asynchronous machine can be expressed as：

The power of asynchronous machine is：

Therefore, the active and reactive power that the change of the input according to load model (busbar voltage) causes The change of P, Q, needs the model parameter of the asynchronous machine of identification to have 8, that is, [R_s,X_s,X_m,R_r,X_r,H,A,B].

Following embodiments of the present invention will be defined as example to the present invention's by the part throttle characteristics of asynchronous machine Technical scheme does exemplary illustration.

Fig. 3 is the part throttle characteristics flow chart that determines method of power system provided in an embodiment of the present invention.As Shown in Fig. 3, the method can include：

S31, according to the constraints of load model parameters, one number of particles of random initializtion is N's Population G, wherein each particle are the array of d load model parameters of an inclusion.

S32, carries out particle group hunting.

Specifically, can basisCarry out particle Group hunting.Wherein, i=1,2 ... ..., N,WithFor carrying out grain during kth time particle group hunting The position of sub- i and speed,WithFor carry out after kth time particle group hunting the position of particle i and Speed,For carrying out the global optimum of described population G during kth time particle group hunting,For Carry out the individual optimal value of particle i during kth time particle group hunting, c₁For the first Studying factors, c₂For Two Studying factors, r₁、r₂For the random number between [0,1],For carrying out during kth time particle group hunting Inertia weight value, and carry out in population search procedure, inertia weight value w of particle cluster algorithm is with particle The S-type change of group hunting number of times k.

S33, when reaching pre-conditioned, stops particle group hunting, and determines load mould according to Search Results Shape parameter；

S34, sets up the load model of power system according to load model parameters；

S35, determines the part throttle characteristics of power system according to load model.

In order to be illustrated more clearly that embodiment of the present invention, can be former in conjunction with the parameter identification of load model Reason is further elaborated.Fig. 4 is the parameter identification principle schematic of load model, refer to figure 4.

Obtain the measured data of load system, the such as input quantity of actual load system and output first.Its In, the input quantity of actual load system includes busbar voltage U and incoming frequency f, and output includes active Power P and reactive power Q.

As described above, particle cluster algorithm be a kind of from RANDOM SOLUTION, optimal solution is found by iteration, and Evaluate the optimized algorithm of the quality of solution by fitness.Therefore, load mould is carried out by particle cluster algorithm During shape parameter identification, first have to the constraints according to parameter to be identified, one population of random initializtion Mesh is population G of N, and each particle is an array including d parameter to be identified.With to be identified Parameter number d as particle in particle cluster algorithm the dimension in search space, and to population calculate Population invariable number N in method, the first Studying factors c₁, the second Studying factors c₂, greatest iteration searching times k_iter Or particle search precision δ is configured.General, c₁=c₂=2.

, in asynchronous motor parameter identification, each particle both corresponds to 8 taking asynchronous machine as a example Parameter [R to be identified_s,X_s,X_m,R_r,X_r, H, A, B] value.Parameter [R to be identified_s,X_s,X_m,R_r,X_r,H,A,B] Constraints, can be for example span of parameter to be identified etc..In actual applications, asynchronous electricity The span of the model parameter of machine may refer to shown in table 1.

The span of the model parameter of table 1 induction machine

Parameter	R_s	X_s	X_m	R_r	X_r	H	A	B
									Maximum	0.8	1.0	5.0	0.8	1.0	3.0	1.0	1.0
Minima	0.1	0.1	1.0	0.05	0.01	0.5	-0.5	-0.5

Further, can be according to the span of parameter to be identified to the particle dominant vector of particle and grain Sub- state vector is encoded, and that is, the particle position to particle and particle rapidity are controlled encoding.In grain Within the restriction range of sub- control variable and particle state variable, one population invariable number of random initializtion is N Population, that is, initialization population in i-th particle particle position x_idWith particle rapidity v_id, with Form current population, and be used for the 1st iterative calculation as the state of first generation particle, and maximum is set Rate limitation v_idmax, not out-of-limit with the speed that guarantees particle.It is understood that in the present embodiment All have：I=1,2 ... ..., N.

Exemplary, in the present embodiment, the particle position x of each particle in population_idAnd particle rapidity v_idCan be initialized as follows respectively：

Wherein,The function of one group of random number, x is uniformly produced for one between (0,1)_idmaxAnd x_idmin Represent maximum and the minima of particle i, v respectively_idmaxRepresent that the maximal rate of particle i limits.

For each particle i in population G, the value of the model parameter being obtained according to random initializtion x_id, and input bus voltage value U and incoming frequency f, can be in conjunction with the mathematical model of asynchronous machine The transient internal voltage E ' of the corresponding asynchronous motor of corresponding differential equation particle i_d、E′_q, and then Voltage equation according to asynchronous machine calculates corresponding stator current i_d、i_q, and finally calculate particle i correspondence Asynchronous motor absorb active-power P_iAnd reactive power Q_i.

It is then possible to evaluate the fitness of each particle initialized by fitness function.As one Plant preferred embodiment, for each particle i, can be according to active-power P_iAnd reactive power Q_iWith Difference between the active-power P of actual measurement and reactive power Q, determines the fitness f of this particle i_i.

Exemplary, function can be selectedTo evaluate the adaptation of each particle Degree.

As k=1, after obtaining the fitness of N number of particle, determine that there is in N number of particle minimum fitness f_minParticle be random initializtion population G global optimum g_best, that is, carry out the 1st grain The g during search of subgroup_best.And set the current location x of each particle_idDuring for the 1st particle group hunting Individual optimal value p of particle i_ibest.

According to formulaCarry out the 1st population Search, i.e. now k=1, the position obtaining each particle is updated by iterationAnd speed

It should be noted that r₁、r₂For the random number between [0,1].

It is further to note that w is inertia weight value when carrying out particle group hunting.Due to population In search procedure, larger inertia weight tends to in global search, and less inertia weight tendency In Local Search.According to correlation technique, using linear decrease inertia weight in PSO algorithm iteration, On the one hand, only within the short period starting iteration, just there is larger inertia weight, this makes particle Group also may not travel through all of region in initial search and just has begun to toward local contraction；Separately On the one hand, in an iterative process, inertia weight is changed with identical speed all the time, is unfavorable for that population is entered Row Local Search.

For this reason, the present embodiment constructs the inertia weight function that a S type as shown in Figure 5 successively decreases, In particle group hunting, inertia weight value w is successively decreased with population searching times k is S-type.Fig. 5 is linearly to pass Subtract the image comparison schematic diagram of the inertia weight function successively decreasing with S type, as shown in figure 5, what S type successively decreased Inertia weight function makes the regional extent of larger inertia weight expand, and in search latter stage, energy Less inertia weight is kept to carry out fine search.

As a kind of optional embodiment of the present embodiment, can basisDetermine inertia weight when carrying out kth time particle group hunting ValueWherein, tanh is hyperbolic tangent function, k_iterFor maximum search number of times, a is for adjusting inertia Weighting function is used for the position of Tuning function curve in maximum, minima transitional region steepness, b, leads to The value crossing adjustment a, b can obtain different function curves, to adapt to different application scenarios.According to this The application scenarios of embodiment, inertia weight parameter a, b can be respectively set to 0.15 and 15.It is, In the present embodiment, specifically can basisDetermine into Inertia weight value w during row kth time particle group hunting^k.

As another kind of optional embodiment of the present embodiment, can basisCalculate Carry out the average fitness of population during kth time particle group huntingFitness by particle iWith kind Group mean fitness valueIt is compared.IfThen illustrate that this particle does not find optimum Solution, or it is absorbed in local optimum, its inertia weight should be increased so as to searching in a wider context or jumping out Local optimum；And ifThen illustrate that this particle relativelys close to excellent solution region, should be reduced it Inertia weight is so as to carry out fine search in regional area.

For this reason, the present embodiment also proposes a kind of search strategy of self-adaptative adjustment inertia weight.Exemplary, Can pass throughDetermine the adjustment amount of inertia weight.

Wherein, two parameters of c, d are constant value, and its span is [0.1,0.6].M is to judge vector, WhenWhen, M=1；WhenWhen, M=-1.

Therefore self adaptation inertia weight, can basisDetermine.Self adaptation inertia weight strategy Can be according to the position adjust automatically flight speed of particle itself, such that it is able to improve the convergence rate of population.

It should be noted that after completing the 1st particle group hunting, further it should checking and updating after Particle positionAnd speedWhether within the scope of setting, whether meet the constraint of particle Condition.If being unsatisfactory for the constraints of particle, it should be modified.

As a kind of optional embodiment, the position of particle after updatingAnd speedDiscontented The position of the particle after renewal during the constraints of sufficient particle, can be revisedAnd speedIt is equal to about The boundary value of bundle condition, for exampleOrAndPermissible It is understood by, because the boundary value of the position of particle has two, when being modified, can be randomly selected it Middle any side dividing value is as the correction value of particle position.

Further, the position according to particle iAnd speedAnd input bus voltage value U and Incoming frequency f, calculates the active power that the corresponding asynchronous motor of particle i absorbsAnd reactive powerAnd according to active powerAnd reactive powerAnd the active-power P of actual measurement and reactive power Q between Difference, determine the fitness of this particle i

Determine in the N number of particle after carrying out updating for the first time that there is minimum fitnessParticle be first The global optimum of population G after secondary renewalCarry out during the 2nd particle group hunting

Further, for the individual optimal value of particle i during the 2nd particle group huntingCan basis Following methods determine：

WhenWhen, determineThe position of corresponding particleFor being somebody's turn to do during the 2nd particle group hunting The individual optimal value of particle i

WhenWhen, determine f_iThe position x of corresponding particle_idIndividual optimal value for this particle i

Same reason, population is carried out the 3rd time, the 4th ... particle group hunting, Zhi Daoda Stop particle group hunting to during default end condition, and according to last population Search Results Lai really Constant load model parameter.It is appreciated that final load model parameters correspond to carries out last particle After group hunting, the position of the corresponding particle of minimum fitness determined according to the fitness of the particle after updating Put, the global optimum determined according to the population after final updating.

It should be noted that default end condition can be for example the maximum search number of times pre-setting k_iter, or default search precision.It is appreciated that search precision can be fitted by the minimum of population Angle value is answered to be described.

After picking out optimal load model parameters, further, can be according to the load mould picking out Shape parameter sets up the load model of power system, and passes through the part throttle characteristics to power system for this load model Carry out accurate description.

Finally it is worth mentioning that, because particle cluster algorithm is absorbed in local optimum and precocious phenomenon in population Global optimumMiddle embodiment the most obvious.Therefore, in another embodiment of the invention, enter one Step, on the basis of above-described embodiment, during particle search, can also to the overall situation of population The figure of meritIt is updated, guides the particle change of flight direction in population, enter its in search space He scans in region, makes population further excavate potential optimal solution.

Exemplary, can be during particle search, in the global optimum of populationMiddle increase by Individual random disturbance quantity μ, according toDetermine and carry out kth time particle group hunting The global optimum of Shi Suoshu population GWherein, μ be withHave same dimension and Obey the stochastic variable of standard normal distribution,For carry out during kth time particle group hunting withRight The initial global optimum of population G answered.

The part throttle characteristics of the power system that the above embodiment of the present invention provides determines method, is improved by one kind The PSO of type improves the precision of load model parameters identification, picks out optimal load model parameters, from And the load model that can set up accurate description power system load characteristic carries out the part throttle characteristics of power system Determine, improve accuracy and the effectiveness of power system load characteristic research.Meanwhile, accurate load The scheduling that Predicting Performance Characteristics can help to electrical network carries out electric power allotment.Because electric energy is difficult to substantial amounts of storage, Therefore there is good load model, that is, accurately the coupling of model parameter can save substantial amounts of resource, energy Effectively carry out the pool with distribution that generates electricity, there is certain economic benefit.

Fig. 6 is that the part throttle characteristics of power system provided in an embodiment of the present invention determines system schematic, and this is System can be used to realize the part throttle characteristics determination side of the power system that embodiment illustrated in fig. 3 of the present invention provides Method, here is omitted.

As shown in fig. 6, the part throttle characteristics of the power system of the present embodiment offer determines that system can include： Load model parameters determining unit 61, load model set up unit 62 and part throttle characteristics determining unit 63.

Wherein, load model parameters determining unit 61 can be used for：Constraint bar according to load model parameters Part, one number of particles of random initializtion is population G of N, and wherein each particle is an inclusion d The array of individual load model parameters；According toEnter Row particle group hunting；When reaching pre-conditioned, stop described particle group hunting, and according to Search Results Determine described load model parameters.Wherein, i=1,2 ... ..., N,WithFor carrying out kth time The position of particle i and speed during particle group hunting,WithFor carrying out after kth time particle group hunting The position of particle i and speed,For carrying out the overall situation of described population G during kth time particle group hunting Optimal value,For carrying out the individual optimal value of particle i during kth time particle group hunting, c₁For first Practise the factor, c₂For the second Studying factors, r₁、r₂For the random number between [0,1],For carrying out kth Inertia weight value during secondary particle group hunting, and carry out in population search procedure, particle cluster algorithm used Weighted value w is with the S-type change of population searching times k for property.

Load model sets up what unit 62 can be used for picking out according to load model parameters determining unit 61 Load model parameters set up the load model of power system.

Part throttle characteristics determining unit 63 can be used for determining power system according to the load model of above-mentioned foundation Part throttle characteristics.

As a kind of specific embodiment, in actual applications, load model parameters determining unit 61 has Body can be used for：According to the position carrying out particle i during kth time particle group huntingDetermine particle i's FitnessDetermine the minimum fitness of population G according to the fitness of N number of particleAnd root According toDetermine the global optimum of population G when carrying out kth time particle group huntingAccording to grain The fitness of sub- iWhenWhen, determineThe position of corresponding particleFor carrying out The individual optimal value of this particle i during k particle group huntingWhenWhen, determineRight The position of the particle answeredFor carrying out the individual optimal value of this particle i during kth time particle group huntingAccording toDetermine that carrying out kth time population searches Inertia weight value during ropeWherein, k_iterFor maximum search number of times.

As another kind of specific embodiment, in actual applications, load model parameters determining unit 61 Specifically can be also used for：According to the position carrying out particle i during kth time particle group huntingDetermine particle The fitness of iDetermine the minimum fitness of population G according to the fitness of N number of particle And according toDetermine the global optimum of population G when carrying out kth time particle group huntingRoot Fitness according to particle iWhenWhen, determineThe position of corresponding particleFor entering The individual optimal value of this particle i during row kth time particle group huntingWhenWhen, determineThe position of corresponding particleFor carrying out the individual optimum of this particle i during kth time particle group hunting ValueDetermine the average fitness of N number of particle according to the fitness of N number of particleWhenWhen, according toDetermination is carried out The inertia weight value of kth time particle group huntingWhenWhen, according toDetermine and carry out kth time particle group hunting Inertia weight valueWherein, k_iterFor maximum search number of times, c, d are value between [0.1,0.6] Constant.

Further, as one kind preferred embodiment, in actual applications, load model parameters are true Order unit 61 specifically can be also used for：DetermineWhen corresponding particle is to carry out kth time particle group hunting The initial global optimum of population GAccording to formulaDetermine Carry out the global optimum of population G during kth time particle group huntingWherein, μ be withTool There is same dimension and obey the stochastic variable of standard normal distribution.

In addition, as a kind of optional embodiment, in actual applications, load model parameters determine single Unit 61 specifically can be also used for：Obtain the measured data of load system, such as busbar voltage U, input frequency Rate f, active-power P and reactive power Q；Position according to particle iDescribed busbar voltage U and Incoming frequency f, determines the corresponding active power of particle iAnd reactive powerAccording to active powerReactive powerDifference and active-power P, reactive power Q between, determines that particle i's is suitable Response

The part throttle characteristics of the power system that the present embodiment provides determines system, can be used to realize figure of the present invention The part throttle characteristics of the power system that 3 illustrated embodiments provide determines method, and it realizes principle and technique effect Similar, here is omitted.

Further, adopt the load of the power system shown in Fig. 6 special in another embodiment of the present invention Property determines that system carries out power system load characteristic Simulation.Population Size is set to 30, that is, initial at random Change 30 groups of load model parameters；The maximal rate of particle is limited to 0.1 times of parameter value scope, maximum Searching times are set to 1500 times.

It is respectively adopted following three kinds of algorithms during emulation and carry out load model parameters identification：

1) Linear recurring series particle cluster algorithm (Linearly Decreasing Inertia Weight PSO, Abbreviation LDW-PSO), in PSO algorithm iteration, adopt the inertia weight of linear decrease；

2) S type inertia weight particle cluster algorithm (PSO With S-Curve Inertia Weight, S-PSO), Adopt in PSO algorithm iteration Determine inertia weight；

3) improved S type inertia weight particle cluster algorithm (Improved PSO With S-Curve Inertia Weight, abbreviation S-IPSO), adopt in PSO algorithm iterationDetermine inertia weight, and adoptUpdate global optimum.

The simulation result of the part throttle characteristics of power system is as shown in table 2：

The part throttle characteristics simulation result of table 2 power system

The characteristic Simulation schematic diagram of the active power that Fig. 7 absorbs when running for power system load, Fig. 8 is The characteristic Simulation schematic diagram of the reactive power that power system load absorbs when running.Refer to Fig. 7 and Fig. 8 Shown although the load model parameters that picked out by the Identification of parameter of above-mentioned three kinds of load models, Practical power systems part throttle characteristics can be carried out with matching substantially, but it will be apparent that pass through S-PSO Determine in precision in the characteristic in the active power and reactive power of system with S-IPSO, with respect to LDW-PSO has larger improvement, especially in sudden load change, by provided in an embodiment of the present invention negative Lotus property determination method can more accurately describe its mutation process, and the waveform simulating is born closer to actual Lotus operation curve, determines method in dynamic load thus demonstrating part throttle characteristics provided in an embodiment of the present invention Effectiveness in characteristic research.

One of ordinary skill in the art will appreciate that：Realize all or part step of above-mentioned each method embodiment Suddenly can be completed by the related hardware of programmed instruction.Aforesaid program can be stored in a computer can Read in storage medium.This program upon execution, executes the step including above-mentioned each method embodiment；And Aforesaid storage medium includes：ROM, RAM, magnetic disc or CD etc. are various can be with store program codes Medium.

Finally it should be noted that：Various embodiments above is only in order to illustrating technical scheme rather than right It limits；Although being described in detail to the present invention with reference to foregoing embodiments, this area common Technical staff should be understood：It still can be modified to the technical scheme described in foregoing embodiments, Or equivalent is carried out to wherein some or all of technical characteristic；And these modifications or replacement, and Do not make the scope of the essence disengaging various embodiments of the present invention technical scheme of appropriate technical solution.

Claims

1. a kind of part throttle characteristics of power system determines method it is characterised in that including：

According to

\{\begin{matrix} v_{i d}^{k + 1} = w_{i d}^{k} \cdot v_{i d}^{k} + c_{1} \cdot r_{1} \cdot (p_{i b e s t}^{k} - x_{i d}^{k}) + c_{2} \cdot r_{2} \cdot (g_{b e s t}^{k} - x_{i d}^{k}) \\ x_{i d}^{k + 1} x_{i d}^{k} + v_{i d}^{k + 1} \end{matrix}

Carry out particle group hunting, wherein, I=1,2 ... ..., N,WithFor carrying out the position of particle i and speed during kth time particle group hunting,WithFor carrying out the position of particle i and speed after kth time particle group hunting,For carrying out The global optimum of described population G during k particle group hunting,Search for carrying out kth time population The individual optimal value of particle i, c during rope₁For the first Studying factors, c₂For the second Studying factors, r₁、r₂ For the random number between [0,1],For carrying out inertia weight value during kth time particle group hunting, and carry out In population search procedure, inertia weight value w of particle cluster algorithm is S-type with population searching times k Successively decrease；

2. method according to claim 1 is it is characterised in that described basis

\{\begin{matrix} v_{i d}^{k + 1} = w_{i d}^{k} \cdot v_{i d}^{k} + c_{1} \cdot r_{1} \cdot (p_{i b e s t}^{k} - x_{i d}^{k}) + c_{2} \cdot r_{2} \cdot (g_{b e s t}^{k} - x_{i d}^{k}) \\ x_{i d}^{k + 1} = x_{i d}^{k} + v_{i d}^{k + 1} \end{matrix}

Carry out particle group hunting, including：

According to the position carrying out particle i during kth time particle group huntingDetermine the fitness f of particle i_i ^k；

Determine the minimum fitness of described population G according to the fitness of N number of particleAnd according toDetermine the global optimum of described population G when carrying out kth time particle group hunting

Fitness f according to particle i_i ^k, work as f_i ^k<f_i ^k-1When, determine f_i ^kThe position of corresponding particle For carrying out the individual optimal value of this particle i during kth time particle group huntingWork as f_i ^k>f_i ^k-1When, Determine f_i ^k-1The position of corresponding particleFor carrying out the individuality of this particle i during kth time particle group hunting Optimal value

According toDetermine and carry out kth time particle group hunting When inertia weight valueWherein, k_iterFor maximum search number of times.

3. method according to claim 1 is it is characterised in that described basis

\{\begin{matrix} v_{i d}^{k + 1} = w_{i d}^{k} \cdot v_{i d}^{k} + c_{1} \cdot r_{1} \cdot (p_{i b e s t}^{k} - x_{i d}^{k}) + c_{2} \cdot r_{2} \cdot (g_{b e s t}^{k} - x_{i d}^{k}) \\ x_{i d}^{k + 1} = x_{i d}^{k} + v_{i d}^{k + 1} \end{matrix}

Carry out particle group hunting, including：

Determine the average fitness of described N number of particle according to the fitness of N number of particle

When

f_{i}^{k} > f_{a v e r}^{k}

When, according to

w_{i d}^{k} = (1 + c) d + 0.3 \times (1 - \tanh (0.15 \times (60 k / k_{i t e r} - 15))) + 0.2

Really Surely carry out the inertia weight value of kth time particle group huntingWherein, k_iterFor maximum search number of times, c, D is constant between [0.1,0.6] for the value；

When

f_{i}^{k} \leq f_{a v e r}^{k}

When, according to

w_{i d}^{k} = (- 1 + c) d + 0.3 \times (1 - \tanh (0.15 \times (60 k / k_{i t e r} - 15))) + 0.2

Determine the inertia weight value carrying out kth time particle group huntingWherein, k_iterFor maximum search number of times, C, d are constant between [0.1,0.6] for the value.

4. according to the method in claim 2 or 3 it is characterised in that described basisDetermine into The global optimum of described population G during row kth time particle group huntingIncluding：

Determine describedCorresponding particle is the first of described population G when carrying out kth time particle group hunting Beginning global optimum

According to formulaDetermination carries out described grain during kth time particle group hunting The global optimum of subgroup GWherein, μ be withThere is same dimension and obedience standard The stochastic variable of normal distribution.

5. method according to claim 4 is it is characterised in that described basis carries out kth time particle The position of particle i during group huntingDetermine the fitness f of particle i_i ^k, including：

Obtain load system measured data, described measured data include busbar voltage U, incoming frequency f, Active-power P and reactive power Q；

Position according to described particle iDescribed busbar voltage U and described incoming frequency f, determine institute State the corresponding active-power P of particle i_i ^kAnd reactive power

According to described active-power P_i ^k, reactive powerAnd active-power P, reactive power Q between Difference, determines the fitness f of described particle i_i ^k.

6. a kind of part throttle characteristics of power system determines system it is characterised in that including：Load model is joined Number determining unit, load model set up unit and part throttle characteristics determining unit；Wherein,

Described load model parameters determining unit is used for：

According to

\{\begin{matrix} v_{i d}^{k + 1} = w_{i d}^{k} \cdot v_{i d}^{k} + c_{1} \cdot r_{1} \cdot (p_{i b e s t}^{k} - x_{i d}^{k}) + c_{2} \cdot r_{2} \cdot (g_{b e s t}^{k} - x_{i d}^{k}) \\ x_{i d}^{k + 1} = x_{i d}^{k} + v_{i d}^{k + 1} \end{matrix}

Carry out particle group hunting, wherein, I=1,2 ... ..., N,WithFor carrying out the position of particle i and speed during kth time particle group hunting,WithFor carrying out the position of particle i and speed after kth time particle group hunting,For carrying out The global optimum of described population G during k particle group hunting,Search for carrying out kth time population The individual optimal value of particle i, c during rope₁For the first Studying factors, c₂For the second Studying factors, r₁、r₂ For the random number between [0,1],For carrying out inertia weight value during kth time particle group hunting, and carry out In population search procedure, inertia weight value w of particle cluster algorithm is S-type with population searching times k Change；

Described load model is set up unit and is used for：Set up the negative of power system according to described load model parameters Lotus model；

Described part throttle characteristics determining unit is used for：The negative of described power system is determined according to described load model Lotus characteristic.

7. system according to claim 6 is it is characterised in that described load model parameters determine list Unit specifically for：

According to

w_{i d}^{k} = 0.3 \times (1 - \tanh (0.15 \times (60 k / k_{i t e r} - 15))) + 0.2

Determine and carry out kth time particle group hunting When inertia weight valueWherein, k_iterFor maximum search number of times.

8. system according to claim 6 is it is characterised in that described load model parameters determine list Unit specifically for：

When

f_{i}^{k} > f_{a v e r}^{k}

When, according to

w_{i d}^{k} = (1 + c) d + 0.3 \times (1 - \tanh (0.15 \times (60 k / k_{i t e r} - 15))) + 0.2

When

f_{i}^{k} < f_{a v e r}^{k}

When, according to

w_{i d}^{k} = (- 1 + c) d + 0.3 \times (1 - \tanh (0.15 \times (60 k / k_{i t e r} - 15))) + 0.2

Really Surely carry out the inertia weight value of kth time particle group huntingWherein, k_iterFor maximum search number of times, c, D is constant between [0.1,0.6] for the value.

9. the system according to claim 7 or 8 is it is characterised in that described load model parameters are true Order unit is specifically additionally operable to：

10. system according to claim 9 is it is characterised in that described load model parameters determine Unit is specifically additionally operable to：

Position according to described particle iDescribed busbar voltage U and described incoming frequency f, determine institute State the corresponding active-power P of particle i_i ^kAnd reactive power Q_i ^k；

According to described active-power P_i ^k, reactive power Q_i ^kAnd active-power P, reactive power Q between Difference, determines the fitness f of described particle i_i ^k.