CN111275160A - Photovoltaic array parameter identification method based on population optimization improved particle swarm algorithm - Google Patents

Photovoltaic array parameter identification method based on population optimization improved particle swarm algorithm Download PDF

Info

Publication number
CN111275160A
CN111275160A CN202010068478.3A CN202010068478A CN111275160A CN 111275160 A CN111275160 A CN 111275160A CN 202010068478 A CN202010068478 A CN 202010068478A CN 111275160 A CN111275160 A CN 111275160A
Authority
CN
China
Prior art keywords
population
particle
value
fitness
iteration
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202010068478.3A
Other languages
Chinese (zh)
Other versions
CN111275160B (en
Inventor
丁坤
翁帅
李辰阳
王立
陈富东
李元良
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Changzhou Campus of Hohai University
Original Assignee
Changzhou Campus of Hohai University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Changzhou Campus of Hohai University filed Critical Changzhou Campus of Hohai University
Priority to CN202010068478.3A priority Critical patent/CN111275160B/en
Publication of CN111275160A publication Critical patent/CN111275160A/en
Application granted granted Critical
Publication of CN111275160B publication Critical patent/CN111275160B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/004—Artificial life, i.e. computing arrangements simulating life
    • G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/06—Energy or water supply

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Health & Medical Sciences (AREA)
  • Theoretical Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Economics (AREA)
  • General Physics & Mathematics (AREA)
  • General Health & Medical Sciences (AREA)
  • Human Resources & Organizations (AREA)
  • Biophysics (AREA)
  • Strategic Management (AREA)
  • Tourism & Hospitality (AREA)
  • Marketing (AREA)
  • General Business, Economics & Management (AREA)
  • Water Supply & Treatment (AREA)
  • Public Health (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Biomedical Technology (AREA)
  • Primary Health Care (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Photovoltaic Devices (AREA)

Abstract

The invention discloses a photovoltaic array parameter identification method based on a population optimization improved particle swarm algorithm in the technical field of photovoltaic power generation, and aims to solve the problems that in the prior art, a photovoltaic cell parameter identification error is large and the method is not suitable for occasions with high precision requirements. Initializing population distribution positions by using chaotic mapping, calculating the optimal values of each particle fitness value, individual and population, adopting a self-adaptive inertial weight adjustment strategy for different populations, judging whether the algorithm is premature convergence according to the difference value of the variance of the fitness values of adjacent iteration populations, carrying out variation operation on the premature population, and outputting the optimal solution of the parameter to be identified when the algorithm meets the end requirement. The method initializes the distribution position of the population by presetting the feasible solution interval of the parameter to be identified and utilizing chaotic cube mapping, calculates the fitness value of each particle, the individual and the optimal value of the population by an improved particle swarm algorithm, can effectively avoid the problem of early-maturing convergence, and has higher identification precision on the parameter to be identified of the photovoltaic cell.

Description

Photovoltaic array parameter identification method based on population optimization improved particle swarm algorithm
Technical Field
The invention belongs to the technical field of photovoltaic power generation, and particularly relates to a photovoltaic array parameter identification method based on a population optimization improved particle swarm algorithm.
Background
Photovoltaic power generation is one of important renewable energy power generation forms, the development is rapid in recent years, the photovoltaic power generation becomes one of important ways for guaranteeing energy supply and building a low-carbon society in China, and a photovoltaic array is an important component of a photovoltaic system, so that the research on the power generation performance of the photovoltaic array is particularly necessary. The I-V curve can express the macroscopic characteristics of the photovoltaic array, each parameter can reflect the inherent characteristics of the array, and the I-V equation of the array can be obtained by determining the parameters, so that the fault diagnosis and the health state management of the photovoltaic array are facilitated to further ensure the power generation performance of the photovoltaic array, and therefore, the parameter identification of the photovoltaic array model is very meaningful.
The I-V equation of the photovoltaic cell is a complex nonlinear transcendental equation, parameters cannot be solved through simple calculation, the traditional photovoltaic cell parameter identification method mostly adopts simplified I-V equations, differential derivation is utilized, and factory parameters of a photovoltaic module are combined to carry out approximate estimation on the parameters to be identified, but the method has larger error and is not suitable for occasions with higher precision requirements.
Disclosure of Invention
The invention aims to provide a photovoltaic array parameter identification method based on a population optimization improved particle swarm algorithm, and aims to solve the problems that in the prior art, the photovoltaic cell parameter identification method has a large approximate estimation error on photovoltaic cell parameters and is not suitable for occasions with high precision requirements.
In order to achieve the purpose, the technical scheme adopted by the invention is as follows: a photovoltaic array parameter identification method based on a population optimization improved particle swarm optimization algorithm comprises the following steps,
a. determining parameters to be identified according to the photovoltaic array five-parameter basic model;
b. obtaining I-V curve data of the photovoltaic array under a certain temperature and irradiance by combining a five-parameter basic model of the photovoltaic array;
c. determining a feasible solution interval of each parameter to be identified according to the I-V curve data, and defining a fitness function;
d. initializing a population position by utilizing chaotic cube mapping, and calculating the fitness value of each particle at the initial moment through a fitness function in combination with each basic parameter of a particle swarm algorithm;
e. dividing the population into three sub-populations of better, ordinary and poorer types according to the fitness value;
f. different subgroups of particles self-adaptively adjust inertia weight, and update the speed and position of each particle;
g. calculating the fitness value of each particle through a fitness function, and updating the individual optimal position and the group optimal position;
h. if the iteration times reach the maximum or the precision meets the requirement, outputting a global optimal solution, namely an optimal value of the parameter to be identified; if the iteration times do not reach the maximum and the precision does not meet the requirements, judging whether the current population falls into precocity convergence, if so, randomly mutating the population, updating the particle fitness value, the individual optimal position and the population optimal position, otherwise, not performing mutation operation, and repeating the steps e-h.
The photovoltaic array five-parameter basic model comprises the following steps:
Figure BDA0002376645070000021
wherein V is the output voltage of the photovoltaic array and I is the output current of the photovoltaic array, IscEquivalent photoproduction current for photovoltaic modules, I0Is reverse saturation current of the photovoltaic module, e is a natural constant, A is an ideal factor of the photovoltaic module, and RsIs equivalent series resistance, R, of a photovoltaic moduleshIs an equivalent parallel resistance of the photovoltaic module, NsNumber of modules connected in series in a photovoltaic array, NpThe number of components connected in parallel in a photovoltaic array,q is an electronic charge, KbThe temperature is a Boltzmann constant, and T is a Kelvin temperature value when the photovoltaic module works;
the parameter to be identified is Isc、I0、A、RsAnd Rsh。
The step c comprises the following steps:
c1, determining a feasible solution interval of each parameter to be identified, wherein the specific method comprises the following steps:
a ranges from [0.5,2 ];
Iscthe maximum value of the actually measured current is taken as the center of the interval, and the variation range is 10 percent;
I0l 'calculated by the formula'0The value is the center of the interval, the variation range is 50%, and the calculation formula is as follows:
Figure BDA0002376645070000031
wherein, I0,refIs reverse saturation current of photovoltaic module under standard test condition, Isc,refFor short-circuit current, V, of photovoltaic modules under standard test conditionsoc,refFor open circuit voltage of photovoltaic modules under standard test conditions, ArefIs an ideal factor, T, of the photovoltaic module under standard test conditionsrefIs the temperature under standard test conditions, EgCalculated I for forbidden bandwidth0' can be taken as I0Taking the central value of the interval;
Rstaking R under standard test conditionssThe standard value is the center of the interval, and the variation range is 20 percent;
taking the absolute value of the reciprocal of the derivative value of the measured I-V curve at the short-circuit current point as RshThe range of variation of the feasible solution interval center is 20 percent;
c2, defining a fitness function, selecting the root mean square error of the measured current and the predicted current as the fitness function, wherein the formula is as follows:
Figure BDA0002376645070000032
wherein f is the calculated fitness, m is the number of data points, IiTo measure the current value, Ii' is a predicted current value corresponding to the measured voltage.
In the step d, the population position is initialized by using chaotic cubic mapping, and the chaotic cubic mapping formula is as follows:
xn+1＝4xn 3-3xn(4)
wherein x isnIs the position of the particle in a dimension of the solution space, xn+1The position of the next particle in the corresponding dimension. x is the number ofnThe value range is [ -1,1 [ ]]And is not 0; randomly generating the position of a particle to ensure that the position of each dimension is not 0, generating a chaotic sequence by utilizing the position of a first particle, initializing the positions of all the rest particles, and independently calculating each dimension;
each basic parameter of the particle swarm optimization comprises a population scale, the maximum iteration number, the dimension of a solution space, and the boundaries of particle positions and speeds.
The step e comprises the following steps:
e1, setting a threshold, directly dividing all the particles into poorer sub-populations when the optimal fitness value is larger than the set threshold in the initial iteration stage, or calculating the average value f of the fitness values of all the particles in the populationsavTurning to the rest sub-steps;
e2, making the fitness value larger than favThe particles of (a) are divided into a better sub-population N1The fitness value is less than favClassified as a common sub-population N2；
e3 calculating sub-populations N separately1And N2Particle average fitness value fav1And fav2；
e4, better sub-population N1Middle fitness value greater than fav1Is divided into a poor sub-population N3Normal sub-population N2Middle fitness value less than fav2The particles of (a) are divided into a better sub-population N1，N1And N2Dividing the rest particles into normal sub-population N2；
e5, determining the poor resultGroup N3If the number of particles is less than 20% of the population size, sorting the particle fitness values if the number of particles is less than 20%, and dividing the lower 20% of the particle fitness values into better sub-populations N1The higher fitness value of the last 30% of the particles is classified as a poor sub-population N3The balance being the common sub-population N2(ii) a Otherwise the classification in step e4 is retained.
In step f, the velocity update formula of the particles is as follows:
Figure BDA0002376645070000051
wherein,
Figure BDA0002376645070000052
the velocity of the ith particle in the D-dimension in the (k + 1) th iteration is represented by ω, i is the inertial weight, i is 1,2,3 … N, N is the number of particles, D is 1,2, … D, D is the dimension of the solution space, and k is the current iteration number; c. C1Is a social learning factor of a particle, is a constant, c2An individual learning factor that is a particle, being a constant; r is1Is the interval [0,1]Random number in between, random coefficient representing movement of particle to the best position of population history, r2Is the interval [0,1]Random numbers in between, representing the random coefficients of the particles moving to the individual historical best positions;
Figure BDA0002376645070000053
is the position of the ith particle in the d dimension in the k iteration;
Figure BDA0002376645070000054
the speed of the ith particle in the d dimension in the kth iteration is taken as the speed of the ith particle in the kth iteration;
Figure BDA0002376645070000055
the optimal position of the ith particle in the d dimension in the kth iteration is taken;
Figure BDA0002376645070000056
the position of the population-optimal particle in the d dimension in the k iteration is determined;
the position update formula of the particle is:
Figure BDA0002376645070000057
wherein,
Figure BDA0002376645070000058
is the position of the ith particle in the d dimension in the (k + 1) th iteration;
the inertial weight adopts a self-adaptive population adjustment strategy, and the generation method comprises the following steps:
Figure BDA0002376645070000059
wherein, ω isminIs the minimum value of the inertial weight, ωmaxIs the maximum value of the inertial weight, i is the particle number, fiIs the fitness value of the ith particle, favAnd fbestRespectively is the average value of the population fitness of the current iteration and the optimal fitness value of the population.
In the step h, randomly mutating the population by the following method:
h1, calculating the difference value between the variance of the fitness value of the current iteration population and the variance of the fitness value of the last iteration population and taking an absolute value;
h2, setting early maturing threshold L1And the variation threshold L of the particle2If the variance of the population fitness values differs by less than a given threshold value L1And when the current iteration does not meet the algorithm end condition, calculating the distance of the position of each particle of the current iteration moving than the distance of the particle of the last iteration moving, wherein the moving distance of the particles in the common sub-population and the poor sub-population is smaller than L2The particles of (a) are subjected to position variation operation and boundary limitation is applied, and the variation formula is as follows:
Figure BDA0002376645070000061
wherein,
Figure BDA0002376645070000062
for the position of the ith particle in the d-dimension in the kth iteration,
Figure BDA0002376645070000063
is the position of the particle after mutation, and t is [0, 1]]Random number of cells, xmaxIs the upper bound, x, of the d-dimensional position of the particleminThe lower bound of the d-dimensional position of the particle,
Figure BDA0002376645070000064
the position of the optimal particle in the d-dimension for the k-th iteration.
Compared with the prior art, the invention has the following beneficial effects: the method initializes the distribution position of the population by presetting the feasible solution interval of the parameter to be identified and utilizing chaotic cube mapping, calculates the fitness value of each particle, the individual and the optimal value of the population by an improved particle swarm algorithm, can effectively avoid the problem of early-maturing convergence, and has higher identification precision on the parameter to be identified of the photovoltaic cell.
Drawings
Fig. 1 is a schematic flow chart of a photovoltaic array parameter identification method based on a population optimization improved particle swarm optimization algorithm according to an embodiment of the present invention;
FIG. 2 is a graph comparing an I-V curve predicted using five parameters obtained by embodiments of the present invention with an actual measured I-V curve.
Detailed Description
The invention is further described below with reference to the accompanying drawings. The following examples are only for illustrating the technical solutions of the present invention more clearly, and the protection scope of the present invention is not limited thereby.
The invention obtains the actually measured I-V curve data through the 1 x 22 type photovoltaic array so as to verify the effectiveness of the method.
Step 1: establishing a photovoltaic array five-parameter model shown in the following formula, and determining parameters to be identified;
Figure BDA0002376645070000071
wherein V is the output voltage of the photovoltaic array and I is the output current of the photovoltaic array, IscEquivalent photoproduction current for photovoltaic modules, I0Is reverse saturation current of the photovoltaic module, e is a natural constant, A is an ideal factor of the photovoltaic module, and RsIs equivalent series resistance, R, of a photovoltaic moduleshIs an equivalent parallel resistance of the photovoltaic module, NsNumber of modules connected in series in a photovoltaic array, NpThe number of the components connected in parallel in the photovoltaic array, q is the electronic charge, and the value of q is 1.602e-19C，KbIs the Boltzmann constant, KbIs taken to be 1.38e-23J/K, and T is a Kelvin temperature value when the photovoltaic module works; temperature T under Standard test conditionsref298K and irradiance SrefThe value is 1000W/m2Finally, determining the parameter to be identified as Isc、I0、A、Rs、Rsh。
Step 2: inputting photovoltaic module nameplate parameters, and actually measuring to obtain actual output voltage V and output current I of the photovoltaic array at a certain temperature T and irradiance S, wherein the actually measured photovoltaic module working temperature T of 52.2 ℃ and irradiance S of 780.4W/m are obtained in the embodiment2Array I-V curve data.
And step 3: and determining a feasible solution interval of each parameter to be identified, and defining a fitness function. The step 3 comprises the following specific steps:
step 3-1: determining a feasible solution interval of each parameter to be identified, wherein the specific method comprises the following steps:
a ranges from [0.5,2 ];
Iscthe maximum value of the actually measured current is taken as the center of the interval, and the variation range is 10 percent;
I0l 'calculated by the formula'0The value is the center of the interval, the variation range is 50%, and the calculation formula is as follows:
Figure BDA0002376645070000081
wherein, I0,refFor standard testing of photovoltaic modulesReverse saturation current under conditions Isc,refFor short-circuit current, V, of photovoltaic modules under standard test conditionsoc,refFor open circuit voltage of photovoltaic modules under standard test conditions, ArefIs an ideal factor, T, of the photovoltaic module under standard test conditionsrefIs the temperature under standard test conditions, EgCalculated I for forbidden bandwidth0' can be taken as I0Taking the central value of the interval;
Rstaking R under standard test conditionssThe standard value 8.932 Ω is the center of the interval, and the variation range is 20%;
taking the absolute value of the reciprocal of the derivative value of the measured I-V curve at the short-circuit current point as RshThe range of variation of the feasible solution interval center is 20 percent;
step 3-2: defining a fitness function, selecting the root mean square error of the measured current and the predicted current as the fitness function, wherein the formula is as follows:
Figure BDA0002376645070000082
wherein f is the calculated fitness value, m is the number of data points, IiTo measure the current value, Ii' is a predicted current value corresponding to the measured voltage.
And 4, step 4: inputting basic parameters of a particle swarm algorithm: the population scale is 50, the maximum iteration number is 700, the dimension of a solution space is 5, the position range of each dimension of the particles is [ -1,1], the maximum speed limit is 0.3, the minimum speed limit is-0.3, then the population position is initialized by utilizing chaotic cubic mapping, and the fitness value of each particle at the initial moment is calculated. The chaotic cube mapping formula is:
xn+1＝4xn 3-3xn(4)
wherein x isnIs the position of the particle in a dimension of the solution space, xn+1For the position of the next particle in the corresponding dimension, xnThe value range is [ -1,1 [ ]]And is not 0; randomly generating the position of a particle to ensure that the position of each dimension is not 0, and using the first particleThe chaotic sequence is generated, the positions of all the other particles are initialized, and each dimension is independently calculated.
And 5: dividing the population into three sub-populations of better, ordinary and poorer types according to the fitness value; the method comprises the following specific steps:
step 5-1: setting a threshold, in this embodiment, setting the threshold to be 100A, directly dividing all particles into poor sub-populations when the optimal fitness value is greater than the set threshold in the initial iteration stage, otherwise, calculating the average value f of the fitness values of all particles in the populationsavTurning to the rest sub-steps;
step 5-2: the fitness value is larger than favThe particles of (a) are divided into a better sub-population N1The fitness value is less than favClassified as a common sub-population N2；
Step 5-3: computing sub-populations N separately1And N2Particle average fitness value fav1And fav2；
Step 5-4: will better sub-population N1Middle fitness value greater than fav1The particles of (a) are divided into a relatively poor sub-population, a common sub-population N2Middle fitness value less than fav2The particles of (A) are divided into better sub-populations, N1And N2Dividing the rest particles into normal sub-population N2。
Step 5-5: judging poor sub-population N3If the number of particles is less than 20% of the population size, sorting the particle fitness values if the number of particles is less than 20%, and dividing the lower 20% of the particle fitness values into better sub-populations N1The higher fitness value of the last 30% of the particles is classified as a poor sub-population N3The balance being the common sub-population N2(ii) a Otherwise the classification in step 5-4 is retained.
Step 6: updating the speed and position of the particles and applying boundary limits;
the velocity update formula of the particles is:
Figure BDA0002376645070000091
wherein,
Figure BDA0002376645070000092
the velocity of the ith particle in the D-dimension in the (k + 1) th iteration is represented by ω, i is the inertial weight, i is 1,2,3 … N, N is the number of particles, D is 1,2, … D, D is the dimension of the solution space, and k is the current iteration number; c. C1Is a social learning factor of a particle, is a constant, c2An individual learning factor that is a particle, being a constant; r is1Is the interval [0,1]Random number in between, random coefficient representing movement of particle to the best position of population history, r2Is the interval [0,1]Random numbers in between, representing the random coefficients of the particles moving to the individual historical best positions;
Figure BDA0002376645070000101
is the position of the ith particle in the d dimension in the k iteration;
Figure BDA0002376645070000102
the speed of the ith particle in the d dimension in the kth iteration is taken as the speed of the ith particle in the kth iteration;
Figure BDA0002376645070000103
the optimal position of the ith particle in the d dimension in the kth iteration is taken;
Figure BDA0002376645070000104
the position of the population-optimal particle in the d dimension in the k iteration is determined;
the position update formula of the particle is:
Figure BDA0002376645070000105
wherein,
Figure BDA0002376645070000106
is the position of the ith particle in the d dimension in the (k + 1) th iteration;
the inertial weight adopts a self-adaptive population adjustment strategy, and the generation method comprises the following steps:
Figure BDA0002376645070000107
wherein, ω isminIs the minimum value of the inertial weight, ωmaxIs the maximum value of the inertial weight, in this embodiment ωminValue of 0.4, omegamaxThe value is 0.9, i is the number of the particle, fiIs the fitness value of the ith particle, favAnd fbestRespectively is the average value of the population fitness of the current iteration and the optimal fitness value of the population.
And 7: and calculating the fitness value of each particle, and updating the individual optimal position and the group optimal position.
And 8: and judging whether the iteration times reach the maximum or whether the precision meets the requirement, if so, turning to the step 10, otherwise, turning to the step 9.
And step 9: and judging whether the current population is trapped in precocity convergence, if so, performing mutation operation on the population, updating the particle fitness value, the individual optimal position and the population optimal position, and otherwise, not performing mutation operation. After this step, the loop continues to step 5; step 9 comprises the following specific sub-steps:
step 9-1: and calculating the difference (taking the absolute value) between the variance of the population fitness value of the current iteration and the variance of the population fitness value of the last iteration.
Step 9-2: setting a premature threshold L1And the variation threshold L of the particle2If the variance of the population fitness values differs by less than a given threshold value L1And when the current iteration does not meet the algorithm end condition, calculating the distance of the position of each particle of the current iteration moving than the distance of the particle of the last iteration moving, wherein the moving distance of the particles in the common sub-population and the poor sub-population is smaller than L2The particles of (a) are subjected to position variation operation and boundary limitation is applied, and the variation formula is as follows:
Figure BDA0002376645070000111
wherein,
Figure BDA0002376645070000112
for the position of the ith particle in the d-dimension in the kth iteration,
Figure BDA0002376645070000113
the position of the particle after mutation (which is not consistent with the expression in claim 6), and t is [0, 1]]Random number of cells, xmaxIs the upper bound, x, of the d-dimensional position of the particleminThe lower bound of the d-dimensional position of the particle,
Figure BDA0002376645070000114
the position of the optimal particle in the d-dimension for the k-th iteration. Updating the fitness value of the variant particles after the variation; if the difference of the variance of the population fitness values is not less than a given threshold value L1No mutation is performed.
Step 10: and outputting a global optimal solution, namely the optimal value of the parameter to be identified, and ending the algorithm.
Through the implementation mode, parameter identification of the photovoltaic array can be completed, in the embodiment, the comparison between an I-V curve obtained through the prediction of the identified parameter result and an I-V curve actually measured by the array is shown in an attached figure 2, the two curves are basically overlapped, and the root mean square error of the current is only 0.0125A, so that the technical scheme of the photovoltaic array parameter identification method based on the population optimization improved particle swarm optimization provided by the invention is feasible. The method initializes the distribution position of the population by presetting the feasible solution interval of the parameter to be identified and utilizing chaotic cube mapping, calculates the fitness value of each particle, the individual and the optimal value of the population by an improved particle swarm algorithm, can effectively avoid the problem of early-maturing convergence, and has higher identification precision on the parameter to be identified of the photovoltaic cell.
The above description is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, several modifications and variations can be made without departing from the technical principle of the present invention, and these modifications and variations should also be regarded as the protection scope of the present invention.

Claims (7)

1. A photovoltaic array parameter identification method based on a population optimization improved particle swarm optimization algorithm is characterized by comprising the following steps of,
a. determining parameters to be identified according to the photovoltaic array five-parameter basic model;
b. obtaining I-V curve data of the photovoltaic array under a certain temperature and irradiance by combining a five-parameter basic model of the photovoltaic array;
c. determining a feasible solution interval of each parameter to be identified according to the I-V curve data, and defining a fitness function;
d. initializing a population position by utilizing chaotic cube mapping, and calculating the fitness value of each particle at the initial moment through a fitness function in combination with each basic parameter of a particle swarm algorithm;
e. dividing the population into three sub-populations of better, ordinary and poorer types according to the fitness value;
f. different subgroups of particles self-adaptively adjust inertia weight, and update the speed and position of each particle;
g. calculating the fitness value of each particle through a fitness function, and updating the individual optimal position and the group optimal position;
h. if the iteration times reach the maximum or the precision meets the requirement, outputting a global optimal solution, namely an optimal value of the parameter to be identified; if the iteration times do not reach the maximum and the precision does not meet the requirements, judging whether the current population falls into precocity convergence, if so, randomly mutating the population, updating the particle fitness value, the individual optimal position and the population optimal position, otherwise, not performing mutation operation, and repeating the steps e-h.
2. The method for identifying the parameters of the photovoltaic array based on the population optimization improved particle swarm algorithm according to claim 1, wherein the five-parameter basic model of the photovoltaic array is as follows:
Figure FDA0002376645060000011
wherein V is the output voltage of the photovoltaic array and I is the output current of the photovoltaic array, IscEquivalent photoproduction current for photovoltaic modules, I0Is light ofReverse saturation current of the photovoltaic module, e is a natural constant, A is an ideal factor of the photovoltaic module, and RsIs equivalent series resistance, R, of a photovoltaic moduleshIs an equivalent parallel resistance of the photovoltaic module, NsNumber of modules connected in series in a photovoltaic array, NpIs the number of components connected in parallel in a photovoltaic array, q is the electron charge, KbThe temperature is a Boltzmann constant, and T is a Kelvin temperature value when the photovoltaic module works;
the parameter to be identified is Isc、I0、A、RsAnd Rsh。
3. The method for identifying the parameters of the photovoltaic array based on the population optimization improved particle swarm algorithm according to claim 1, wherein the step c comprises the following steps:
c1, determining a feasible solution interval of each parameter to be identified, wherein the specific method comprises the following steps:
a ranges from [0.5,2 ];
Iscthe maximum value of the actually measured current is taken as the center of the interval, and the variation range is 10 percent;
I0l 'calculated by the formula'0The value is the center of the interval, the variation range is 50%, and the calculation formula is as follows:
Figure FDA0002376645060000021
wherein, I0,refIs reverse saturation current of photovoltaic module under standard test condition, Isc,refFor short-circuit current, V, of photovoltaic modules under standard test conditionsoc,refFor open circuit voltage of photovoltaic modules under standard test conditions, ArefIs an ideal factor, T, of the photovoltaic module under standard test conditionsrefIs the temperature under standard test conditions, EgCalculated I for forbidden bandwidth0' can be taken as I0Taking the central value of the interval;
Rstaking R under standard test conditionssThe standard value is the center of the interval, and the variation range is 20 percent;
the measured I-V curveThe absolute value of the reciprocal of the derivative value at the short-circuit current point is taken as RshThe range of variation of the feasible solution interval center is 20 percent;
c2, defining a fitness function, selecting the root mean square error of the measured current and the predicted current as the fitness function, wherein the formula is as follows:
Figure FDA0002376645060000031
wherein f is the calculated fitness, m is the number of data points, IiTo measure the current value, Ii' is a predicted current value corresponding to the measured voltage.
4. The method for identifying the parameters of the photovoltaic array based on the population optimization improved particle swarm algorithm according to claim 1, wherein in the step d, the population position is initialized by using chaotic cubic mapping, and the chaotic cubic mapping formula is as follows:
xn+1＝4xn 3-3xn(4)
wherein x isnIs the position of the particle in a dimension of the solution space, xn+1The position of the next particle in the corresponding dimension. x is the number ofnThe value range is [ -1,1 [ ]]And is not 0; randomly generating the position of a particle to ensure that the position of each dimension is not 0, generating a chaotic sequence by utilizing the position of a first particle, initializing the positions of all the rest particles, and independently calculating each dimension;
each basic parameter of the particle swarm optimization comprises a population scale, the maximum iteration number, the dimension of a solution space, and the boundaries of particle positions and speeds.
5. The method for identifying the parameters of the photovoltaic array based on the population optimization improved particle swarm algorithm according to claim 1, wherein the step e comprises the following steps:
e1, setting a threshold, directly dividing all the particles into poor sub-populations when the optimal fitness value is larger than the set threshold in the initial iteration stage, otherwise, calculating the fitness of all the particles in the populationsAverage value f of valuesavTurning to the rest sub-steps;
e2, making the fitness value larger than favThe particles of (a) are divided into a better sub-population N1The fitness value is less than favClassified as a common sub-population N2；
e3 calculating sub-populations N separately1And N2Particle average fitness value fav1And fav2；
e4, better sub-population N1Middle fitness value greater than fav1Is divided into a poor sub-population N3Normal sub-population N2Middle fitness value less than fav2The particles of (a) are divided into a better sub-population N1，N1And N2Dividing the rest particles into normal sub-population N2；
e5, judging the poor sub-population N3If the number of particles is less than 20% of the population size, sorting the particle fitness values if the number of particles is less than 20%, and dividing the lower 20% of the particle fitness values into better sub-populations N1The higher fitness value of the last 30% of the particles is classified as a poor sub-population N3The balance being the common sub-population N2(ii) a Otherwise the classification in step e4 is retained.
6. The method for identifying the parameters of the photovoltaic array based on the population optimization improved particle swarm algorithm according to claim 1, wherein in the step f, the velocity updating formula of the particles is as follows:
Figure FDA0002376645060000041
wherein,
Figure FDA0002376645060000042
the velocity of the ith particle in the D-dimension in the (k + 1) th iteration is represented by ω, i is the inertial weight, i is 1,2,3 … N, N is the number of particles, D is 1,2, … D, D is the dimension of the solution space, and k is the current iteration number; c. C1Is a social learning factor of a particle, is a constant, c2An individual learning factor that is a particle, being a constant; r is1Is the interval [0,1]Random number in between, random coefficient representing movement of particle to the best position of population history, r2Is the interval [0,1]Random numbers in between, representing the random coefficients of the particles moving to the individual historical best positions;
Figure FDA0002376645060000043
is the position of the ith particle in the d dimension in the k iteration;
Figure FDA0002376645060000044
the speed of the ith particle in the d dimension in the kth iteration is taken as the speed of the ith particle in the kth iteration;
Figure FDA0002376645060000045
the optimal position of the ith particle in the d dimension in the kth iteration is taken;
Figure FDA0002376645060000046
the position of the population-optimal particle in the d dimension in the k iteration is determined;
the position update formula of the particle is:
Figure FDA0002376645060000047
wherein,
Figure FDA0002376645060000048
is the position of the ith particle in the d dimension in the (k + 1) th iteration;
the inertial weight adopts a self-adaptive population adjustment strategy, and the generation method comprises the following steps:
Figure FDA0002376645060000051
wherein, ω isminIs the minimum value of the inertial weight, ωmaxIs the maximum value of the inertial weight, i is the particle number, fiIs the fitness value of the ith particle, favAnd fbestAre respectively the current stacksThe population fitness average value and the population optimal fitness value of the generation.
7. The method for identifying the parameters of the photovoltaic array based on the population optimization improved particle swarm algorithm according to claim 1, wherein in the step h, the population is randomly varied by the following method:
h1, calculating the difference value between the variance of the fitness value of the current iteration population and the variance of the fitness value of the last iteration population and taking an absolute value;
h2, setting early maturing threshold L1And the variation threshold L of the particle2If the variance of the population fitness values differs by less than a given threshold value L1And when the current iteration does not meet the algorithm end condition, calculating the distance of the position of each particle of the current iteration moving than the distance of the particle of the last iteration moving, wherein the moving distance of the particles in the common sub-population and the poor sub-population is smaller than L2The particles of (a) are subjected to position variation operation and boundary limitation is applied, and the variation formula is as follows:
Figure FDA0002376645060000052
wherein,
Figure FDA0002376645060000053
for the position of the ith particle in the d-dimension in the kth iteration,
Figure FDA0002376645060000054
is the position of the particle after mutation, and t is [0, 1]]Random number of cells, xmaxIs the upper bound, x, of the d-dimensional position of the particleminThe lower bound of the d-dimensional position of the particle,
Figure FDA0002376645060000055
the position of the optimal particle in the d-dimension for the k-th iteration.
CN202010068478.3A 2020-01-21 2020-01-21 Photovoltaic array parameter identification method based on population optimization improved particle swarm algorithm Active CN111275160B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010068478.3A CN111275160B (en) 2020-01-21 2020-01-21 Photovoltaic array parameter identification method based on population optimization improved particle swarm algorithm

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010068478.3A CN111275160B (en) 2020-01-21 2020-01-21 Photovoltaic array parameter identification method based on population optimization improved particle swarm algorithm

Publications (2)

Publication Number Publication Date
CN111275160A true CN111275160A (en) 2020-06-12
CN111275160B CN111275160B (en) 2022-09-13

Family

ID=70999051

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010068478.3A Active CN111275160B (en) 2020-01-21 2020-01-21 Photovoltaic array parameter identification method based on population optimization improved particle swarm algorithm

Country Status (1)

Country Link
CN (1) CN111275160B (en)

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111814399A (en) * 2020-07-08 2020-10-23 温州大学 Model parameter optimization extraction method and measurement data prediction method for solar photovoltaic cell system
CN113595132A (en) * 2021-06-25 2021-11-02 福州大学 Photovoltaic online parameter identification method based on global maximum power point tracking and hybrid optimization algorithm
CN113779733A (en) * 2021-09-18 2021-12-10 国网江苏省电力有限公司常州供电分公司 Photovoltaic module model parameter hybrid optimization identification method
CN113919474A (en) * 2021-09-13 2022-01-11 中国电建集团江西省电力建设有限公司 Photovoltaic operation and maintenance data oriented reduction method
CN115204048A (en) * 2022-07-19 2022-10-18 贵州大学 Photovoltaic power generation system model identification system and method
CN115409387A (en) * 2022-08-30 2022-11-29 华中科技大学 Reservoir optimal scheduling method and system based on improved differential evolution

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104933307A (en) * 2015-06-15 2015-09-23 西北工业大学 Solar cell implicit equation parameter identification method based on particle swarm optimization algorithm
CN106202914A (en) * 2016-07-07 2016-12-07 国网青海省电力公司 Based on the photovoltaic cell parameter identification method improving particle cluster algorithm

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104933307A (en) * 2015-06-15 2015-09-23 西北工业大学 Solar cell implicit equation parameter identification method based on particle swarm optimization algorithm
CN106202914A (en) * 2016-07-07 2016-12-07 国网青海省电力公司 Based on the photovoltaic cell parameter identification method improving particle cluster algorithm

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111814399A (en) * 2020-07-08 2020-10-23 温州大学 Model parameter optimization extraction method and measurement data prediction method for solar photovoltaic cell system
CN111814399B (en) * 2020-07-08 2024-03-19 温州大学 Model parameter optimization extraction method and measurement data prediction method of solar photovoltaic cell system
CN113595132A (en) * 2021-06-25 2021-11-02 福州大学 Photovoltaic online parameter identification method based on global maximum power point tracking and hybrid optimization algorithm
CN113595132B (en) * 2021-06-25 2023-11-17 福州大学 Photovoltaic online parameter identification method based on maximum power point and hybrid optimization algorithm
CN113919474A (en) * 2021-09-13 2022-01-11 中国电建集团江西省电力建设有限公司 Photovoltaic operation and maintenance data oriented reduction method
CN113919474B (en) * 2021-09-13 2022-03-15 中国电建集团江西省电力建设有限公司 Photovoltaic operation and maintenance data oriented reduction method
CN113779733A (en) * 2021-09-18 2021-12-10 国网江苏省电力有限公司常州供电分公司 Photovoltaic module model parameter hybrid optimization identification method
CN115204048A (en) * 2022-07-19 2022-10-18 贵州大学 Photovoltaic power generation system model identification system and method
CN115409387A (en) * 2022-08-30 2022-11-29 华中科技大学 Reservoir optimal scheduling method and system based on improved differential evolution

Also Published As

Publication number Publication date
CN111275160B (en) 2022-09-13

Similar Documents

Publication Publication Date Title
CN111275160B (en) Photovoltaic array parameter identification method based on population optimization improved particle swarm algorithm
CN104573879B (en) Photovoltaic plant based on optimal similar day collection goes out force prediction method
CN108471143A (en) Micro-grid multi-energy method for optimizing scheduling based on positive and negative feedback particle cluster algorithm
CN106126863B (en) Photovoltaic cell parameter identification method based on artificial fish-swarm and the algorithm that leapfrogs
CN111934360B (en) Virtual power plant-energy storage system energy collaborative optimization regulation and control method based on model predictive control
CN110287540B (en) Photovoltaic cell parameter identification method based on elite crowd differential evolution algorithm
CN110334870B (en) Photovoltaic power station short-term power prediction method based on gated cyclic unit network
CN107103154A (en) A kind of photovoltaic module model parameter identification method
CN104734175A (en) Method for intelligently modifying wind speed and power curve of wind turbine generator
TWI391807B (en) A maximum power tracking system and method for photovoltaic power generation systems
CN110286708B (en) Maximum power tracking control method and system for photovoltaic array
CN105356451A (en) Probability trend calculating method considering large-scale photovoltaic grid-connected power generation system
CN106355511A (en) Active power distribution network reconstruction method taking new energy and electric vehicle access into consideration
CN107066760A (en) A kind of photovoltaic module parameter identification method
CN112465271A (en) Energy storage battery model selection method for energy storage stabilizing wind power fluctuation scene
Yong et al. The MPPT control method by using BP neural networks in PV generating system
Zhang et al. A genetic algorithm approach to parameter estimation for PV modules
CN109145503B (en) High-precision dynamic modeling method for photovoltaic power station cluster
Figueiredo et al. Comparison between P&O-based and PSO-based MPPT algorithms for photovoltaic system under partially shaded conditions
CN112114254B (en) Power battery open-circuit voltage model fusion method
CN109543269A (en) A kind of photovoltage model parameter identification method for trusting section reflection based on mixing
CN105633971A (en) Intelligent power distribution network random power flow method based on imperialism competition algorithm
CN117540638A (en) Equivalent modeling method, medium and system for photovoltaic power station
CN113779733A (en) Photovoltaic module model parameter hybrid optimization identification method
Chenhong et al. Wind speed prediction research based on time series model with residual correction

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant