CN115203857A - Amorphous alloy transformer parameter optimization design method based on particle swarm optimization - Google Patents

Amorphous alloy transformer parameter optimization design method based on particle swarm optimization Download PDF

Info

Publication number
CN115203857A
CN115203857A CN202210919265.6A CN202210919265A CN115203857A CN 115203857 A CN115203857 A CN 115203857A CN 202210919265 A CN202210919265 A CN 202210919265A CN 115203857 A CN115203857 A CN 115203857A
Authority
CN
China
Prior art keywords
particle
amorphous alloy
value
particle swarm
fitness
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.)
Withdrawn
Application number
CN202210919265.6A
Other languages
Chinese (zh)
Inventor
刘道生
钟伟
王永胜
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.)
Jiangxi University of Science and Technology
Original Assignee
Jiangxi University of Science and Technology
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 Jiangxi University of Science and Technology filed Critical Jiangxi University of Science and Technology
Priority to CN202210919265.6A priority Critical patent/CN115203857A/en
Publication of CN115203857A publication Critical patent/CN115203857A/en
Withdrawn legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/10—Geometric CAD
    • G06F30/17—Mechanical parametric or variational design
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
    • 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
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2111/00—Details relating to CAD techniques
    • G06F2111/04—Constraint-based CAD

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Geometry (AREA)
  • General Physics & Mathematics (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • Computer Hardware Design (AREA)
  • Artificial Intelligence (AREA)
  • Software Systems (AREA)
  • Computing Systems (AREA)
  • Medical Informatics (AREA)
  • General Health & Medical Sciences (AREA)
  • Data Mining & Analysis (AREA)
  • Computational Linguistics (AREA)
  • Mathematical Physics (AREA)
  • Biophysics (AREA)
  • Biomedical Technology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Molecular Biology (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Computational Mathematics (AREA)
  • Mathematical Analysis (AREA)
  • Mathematical Optimization (AREA)
  • Pure & Applied Mathematics (AREA)
  • Soft Magnetic Materials (AREA)

Abstract

The invention discloses an amorphous alloy transformer parameter optimization design method based on a particle swarm optimization algorithm. The transformer designed by the method is an efficient energy-saving amorphous alloy dry-type transformer, and has better energy-saving effect and better safety and reliability. The parameter mathematical optimization model of the amorphous alloy dry type transformer is established from the perspective of reducing the manufacturing cost and reducing the loss, and the optimal solution searching is carried out by improving the particle swarm algorithm to obtain the electromagnetic parameters of the amorphous alloy dry type transformer with low loss and low manufacturing cost; the automation degree is high and the design period is short. The method can obtain the amorphous alloy dry-type transformer design scheme with low manufacturing cost meeting the performance parameter requirement in a short time, reduce the production cost of the amorphous alloy dry-type transformer and improve the loss performance of the amorphous alloy dry-type transformer; has important significance for the popularization and the application of the energy-saving amorphous alloy dry-type transformer.

Description

Amorphous alloy transformer parameter optimization design method based on particle swarm optimization
Technical Field
The invention relates to an amorphous alloy transformer parameter optimization design method based on a particle swarm optimization, in particular to a multi-constraint discrete multivariable optimization method, and belongs to the field of heuristic algorithms.
Background
The transformer is an expensive power device in the power system, and plays a crucial role in power transmission and distribution of the power system. With the development of the transformer industry, the demand of the energy-saving transformer is increasingly urgent. Compared with the traditional silicon steel sheet iron core transformer, the amorphous alloy transformer has excellent no-load performance and occupies a great demand in the distribution transformer market. In recent years, the price of the amorphous alloy transformer rises along with the price rise of raw materials such as an iron core, a lead, a clamping piece and the like required by manufacturing the amorphous alloy transformer, and certain barrier effects are generated on the popularization and the application of the amorphous alloy transformer. On the premise of ensuring the performance of the amorphous alloy transformer, the amorphous alloy transformer is optimally designed, so that the reduction of the production cost of the amorphous alloy transformer becomes important.
The optimization design of the amorphous alloy transformer is a multivariable, multi-objective and nonlinear mixed type problem. When manual calculation is adopted, the optimization degree of calculation is influenced by too many human and environmental factors, so that the defects of production material waste and product quality stability are caused, the automation degree of the traditional transformer design method is low, the design period is long, and the optimal transformer design scheme is difficult to obtain. With the development of computer technology and artificial intelligence, many solving algorithms of transformer optimization design models are proposed at present, such as genetic algorithm, frog-leaping algorithm, bacterial foraging algorithm, differential evolution algorithm, particle swarm algorithm, quantum particle swarm algorithm and the like. The algorithms are applied to the parameter optimization design of the amorphous alloy transformer, so that the design period of the transformer is greatly shortened and the design efficiency is improved.
Compared with other swarm intelligent algorithms, the particle swarm algorithm has the characteristics of simple structure and less dependent parameters. In practical application, due to a solution space search mode of the particle swarm algorithm, the search speed of the solution space is high, but the global search of the solution space has certain defects. Therefore, in order to solve the problems of the particle swarm optimization in the optimization of the amorphous alloy transformer, the invention provides corresponding improvement measures for the conventional particle swarm optimization so as to improve the global search capability of the particle swarm optimization and obtain the particle swarm optimization with stronger constraint condition processing capability and optimizing capability for the amorphous alloy transformer.
Disclosure of Invention
The invention aims to provide a particle swarm algorithm-based amorphous alloy transformer parameter optimization design, which is a multivariable, multi-target and nonlinear mixed problem, aims at solving the problem that the overall search capability of the particle swarm algorithm in the actual optimization design process is insufficient, solves the amorphous alloy transformer design scheme meeting various constraint conditions by using the improved particle swarm algorithm, and can reduce the main material cost of the amorphous alloy transformer to the maximum extent.
The purpose of the invention is realized by the following technical scheme:
an amorphous alloy transformer parameter optimization design method based on a particle swarm algorithm comprises the following steps:
the method comprises the following steps:
determining the value range parameters of the performance parameter set constraint conditions and the optimized parameter variables according to the requirements of users of the amorphous alloy dry-type transformer or performance parameter standards;
step two:
generating codes of optimization variables of all dimensions and corresponding actual value groups, then randomly generating positions of first generation of particle swarms, and setting fitness functions and penalty functions for the particle swarms;
step three:
comparing the calculated particle fitness with the optimal position of each particle and the optimal positions of all the particles, and replacing and updating the optimal position and the global optimal position of each particle if the calculated particle fitness is better than the optimal position of the previous generation;
step four:
and executing a particle cross search strategy according to the update condition of the global optimal position of the current particle, and if a new particle superior to the current particle is found, replacing the current particle by the new particle.
Step five:
and judging whether the set maximum iteration number N is reached, if not, repeating the third step and the fourth step, updating the population position, and calculating a fitness function value.
Step six:
and stopping when the maximum iteration number N is reached, and outputting the optimal design scheme in the population.
The second step of the invention specifically comprises the following steps:
the constraint condition is processed by using a penalty function method, and the specific constraint condition processing method and fitness function calculation are as follows.
The constraints can be expressed as:
g j (X i )≤0,j＝1,...,q
and q is the number of the constraint conditions.
The constraint violation degree of the ith particle on the jth constraint is expressed as:
G j (X i )＝max{g j (X i ),0},1≤j≤q
the constraint violation degree of the particle is
Figure BDA0003776931730000031
Due to the difference of the constraint conditions, the constraint violation degree v (X) of some constraint conditions on the individual can occur i ) The conditions that play a dominant role, so the order of magnitude of each constraint is made the same by the normalization method. In the normalization process, the maximum value of the first generation population where the particles violate the respective constraints is first found.
Figure BDA0003776931730000032
Gmax j To violate the maximum of the jth constraint, the other parameters are as defined above.
Violation degree standard value Gnorm of ith particle to jth constraint condition j (X i ) The definition is as follows:
Figure BDA0003776931730000041
the constraint violation degree standard value v of the particle is calculated in the whole subsequent iteration process norm (X i ) The definition is as follows:
Figure BDA0003776931730000042
similarly, the fitness function and v are also normalized by a normalization method norm (X i ) Are of the same order of magnitude. First find the minimum and maximum objective function values for the first generation population.
Figure BDA0003776931730000043
Figure BDA0003776931730000044
f(X i ) Is a fitness function value, f min And f max Respectively the minimum value and the maximum value of the objective function of the first generation population particles.
Then, the standard value of the objective function of each generation of the individual x is calculated as follows:
Figure BDA0003776931730000045
f norm (x i ) The other parameters are the same as above for the standard values of the objective function.
The fitness function calculation mode is as follows:
fitness(X i )＝f norm (X i )+λ(t)v norm (X i )
fitness(X i ) The fitness value is called fitness for short; λ (t) is a penalty factor.
The smaller the fitness of the particle, the better the particle is represented.
Solutions which do not meet the constraint condition are also called infeasible solutions, and the existence of the infeasible solutions is allowed by the iteration process so as to guarantee the global search capability in the early stage. The punishment coefficient at the early stage of iteration is a small value, after the post-iteration stage, the population is converged to a feasible solution, and a large punishment coefficient is needed at the moment. And adjusting the penalty coefficient in a linear increasing mode.
λ(t)＝λ min +(λ max -λ min )/T×t
In the formula of min And λ max Minimum and maximum penalty factors, respectively.
The third step of the invention specifically comprises the following steps:
the optimization target is an economic index, namely the main material cost of the amorphous alloy dry-type transformer:
f(X)＝C h +C l +C co +C cl
wherein, C h Cost for the high voltage winding; c l Cost for low voltage winding; c co The cost of the iron core; c cl For clip cost, a smaller objective function represents a better design solution.
The fourth step of the invention specifically comprises the following steps:
when the convergence of the particle swarm falls into stagnation, new particles are generated through the intersection among the particles, and a new global optimal solution is searched through the new particles. The cross search among the particles makes up the defects of a particle swarm algorithm solution space search mode.
And (3) an iteration process, wherein each particle is crossed with another random particle by a probability Pr determined by the convergence condition of the current particle swarm, and the probability Pr is defined as:
Pr＝μ×Re
wherein Re is p g Continuously non-updated algebra, mu is the coefficient for increasing the cross probability
The crossover operation is defined as:
Figure BDA0003776931730000051
wherein X new1 And X new1 Is a new particle generated by the crossover operation; x 1 Is the current particle, X 2 Is X 1 An outer random particle; rnd is a 6-dimensional random vector of one (0,1) interval.
Mixing X new1 And X new2 And p g By contrast, if X new1 And X new2 Preferably, p is replaced g 。
The sixth step of the invention specifically comprises the following steps:
setting iteration times N of the population, carrying out a particle cross search strategy according to the particle updating condition in the step three and the step four, and replacing a new particle with the new particle if the new particle superior to the new particle is found; and repeating iteration for N times, sequencing the cost prices of the main materials of the final population from low to high, outputting the design scheme of the amorphous alloy transformer with the lowest main material cost according with the performance standard, and generating a calculation list.
The amorphous alloy transformer parameter optimization design method based on the particle swarm optimization comprises an amorphous alloy dry-type transformer parameter setting module, a particle swarm optimization operation parameter and optimization variable upper and lower boundary setting module, a particle swarm optimization engine module and an optimization result output module, and the amorphous alloy dry-type transformer parameter setting module, the particle swarm optimization engine module and the optimization result output module have the following composition structure and connection relation:
the amorphous alloy dry-type transformer parameter setting module is used for setting and optimizing parameters such as rated capacity, iron core materials, iron core forms, insulation grades and the like of the amorphous alloy dry-type transformer, and the module allows manual parameter adjustment to meet the actual product requirements;
the particle swarm algorithm operation parameter and optimization variable upper and lower limit setting module is used for setting parameters such as the number of particles, the dimension, the acceleration factor C1, the acceleration factor C2, the maximum inertia weight, the minimum inertia weight and the like of the particle swarm algorithm, and setting the upper limit and the lower limit of the iron core stack thickness, the upper limit and the lower limit of the low-voltage layer number, the upper limit and the lower limit of the high-voltage line width (line thickness) and the upper limit and the lower limit of the low-voltage line width (line thickness);
the particle swarm algorithm engine module is used for carrying out particle swarm algorithm related calculation, including random encoding, generating an initial variable population, calculating a target function value (a target function with a penalty function) of the generated initial population and the fitness value of population individuals, updating the population position, finally generating a new filial generation population, and carrying out iterative calculation until a convergence condition is met and stopping calculation;
the optimization result output module is used for outputting a plurality of amorphous alloy transformer optimization schemes which are obtained by the particle swarm algorithm and meet the performance requirements of customers or manufacturers, and selecting the optimal amorphous alloy transformer optimization design scheme according to the requirements of low manufacturing cost and low loss.
Advantageous effects
The method selects the iron core stack thickness, the low-voltage layer number, the high-voltage wire width, the high-voltage wire thickness, the low-voltage wire width and the low-voltage wire thickness as optimization variables, optimizes the amorphous alloy transformer by adopting an improved particle swarm algorithm under constraint conditions, takes the main material cost (the cost of the iron core, the low-voltage wire, the high-voltage wire, the clamping piece and other accessories) as a target function, can obtain a plurality of amorphous alloy transformer optimization design schemes under the condition of meeting performance standards, selects the design scheme with low main material cost from the design schemes, and can output the schemes in a form of calculation sheet.
Drawings
FIG. 1 is an algorithm flow chart of a particle swarm algorithm-based amorphous alloy transformer parameter optimization method
FIG. 2 is a flow chart of amorphous alloy dry type transformer design
FIG. 3 is a graph showing the convergence success rate of the particle swarm optimization algorithm when the cross probability μ takes different values
FIG. 4 is a structural design diagram of an amorphous alloy transformer parameter optimization device based on a particle swarm optimization
Detailed Description
The present invention will be described in detail with reference to the accompanying drawings and examples, and it should be noted that the technical problems and effects of the present invention are also described. All other embodiments, which can be obtained by a person skilled in the art based on the embodiments of the present invention without any inventive step, shall fall within the scope of protection of the present invention.
As shown in fig. 1, the present invention comprises the steps of:
step 1, determining performance parameters according to amorphous alloy dry type transformer users or performance parameter standards, and setting a value range parameter of a constraint condition and an optimized parameter variable;
the parameters of the system to be determined before optimization are rated capacity, iron core form, connection group and insulation grade;
for the design scheme of the whole amorphous alloy dry-type transformer, the selected optimized parameters comprise six of iron core overlapping thickness, low-voltage layer number, low-voltage line width, low-voltage line thickness, high-voltage line width and high-voltage line thickness;
generating a first generation population by utilizing a random process according to the second step, and setting a fitness function and a penalty function for the population;
a first generation population consisting of six-dimensional random vectors is generated through a random process, and the upper and lower value limit values of each-dimensional random variable are determined by a line gauge table in a system.
For each particle in the population, calculation and inspection are performed according to the transformer calculation process shown in fig. 2, and the specific steps are as follows:
1) According to the selected amorphous alloy model and iron core structure form, the formula
Figure BDA0003776931730000081
Estimating the diameter of the iron core, and designing the sections of the iron core column and the iron yoke, the turn potential and the number of turns.
Wherein D is the estimated core diameter; k is the iron core coefficient; s φ Is rated capacity; a. The j To estimate the net sectional area of the core; e.g. of a cylinder t To estimate the turn potential; f is the frequency of 50Hz; b is m 1.25T was taken for the core flux density estimation.
2) By the formula
Figure BDA0003776931730000082
And calculating the turns of the high-voltage winding and the low-voltage winding to be integrated.
Wherein N is d The number of low-voltage winding turns; u shape dφ A low voltage phase voltage; u shape gφ Is a high voltage phase voltage.
3) According to the winding structure form, the specification of the wire is determined, and the formula is as follows:
Figure BDA0003776931730000083
and calculating the axial dimension and the radial dimension of the winding.
Wherein Z gyzx Is the axial dimension of the high-voltage winding; w mdk Each section of the high-voltage winding is wide; l is a radical of an alcohol zbdk 、L djdk Respectively forming a middle space and an inter-section space of the high-voltage winding; h gyfx The radial dimension of the high-voltage winding is; h is dxjy The high-voltage wire is insulated thick; h is gyjy The thickness of the interlayer insulation of the high-voltage winding; x is the number of gydx The high-voltage winding stacking thickness coefficient; z dyzx Is the axial size of the low-voltage winding; w lk The width of the low-voltage winding wire is; b is dybr The low-voltage conducting wire is wound in parallel; h dyfx The radial total size of the low-voltage winding; h dyfx1 The radial size of the low-voltage winding at the inner side of the air passage; h dyfx2 The radial size of the low-voltage winding outside the air passage; c dy1 The number of layers of the low-voltage winding on the inner side of the air passage is; c dy2 The number of layers of the low-voltage winding outside the air passage is; h is dyjy The insulation thickness between layers of the low-voltage winding is set; x is the number of dydx Is a low voltage winding stack thickness factor.
And calculating short-circuit impedance, no-load loss, load loss and winding temperature rise, randomly calculating relevant parameters of the winding again when the short-circuit impedance, the no-load loss, the load loss and the winding temperature rise are not met, and readjusting the diameter of the iron core if the short-circuit impedance, the no-load loss, the load loss and the winding temperature rise are not met.
Generating a population with the size of 100 according to the method, and setting the maximum iteration algebra T to 4000 generations; learning factor c 1 And c 2 Set to 2; minimum inertial weight w min Set to 0.3, maximum inertial weight w max Set to 1.1; as shown in fig. 3, the convergence success rate is highest when the cross probability increase coefficient μ is 0.5%, the cross probability increase coefficient μ is set to 0.5%,
after the initial population is generated, setting a fitness function calculation mode as follows:
fitness(X i )＝f norm (X i )+λ(t)v norm (X i )
wherein fitness (X) i ) The fitness value is called fitness for short; λ (t) is a penalty coefficient.
The smaller the fitness of the particle, the better the particle is represented.
The main material cost of the amorphous alloy dry type transformer is set as a fitness function, namely:
f(X)＝C h +C l +C co +C cl
wherein, C h Cost for the high voltage winding; c l Cost for low voltage windings; c co The core cost; c cl Is the clip cost.
Smaller objective functions represent better design solutions.
And according to the third step, comparing the calculated particle fitness with the optimal position of each particle and the optimal position of the whole particles, replacing and updating the optimal position of each particle and the optimal position of the whole particles if the calculated particle fitness is better than the optimal position of the previous generation, generating new particles through the intersection among the particles when the convergence of the particle swarm is stagnated, and searching a new global optimal solution through the new particles.
And (3) an iteration process, wherein each particle is crossed with another random particle by a probability Pr determined by the convergence condition of the current particle swarm, and the probability Pr is defined as:
Pr＝μ×Re
wherein Re is p g Mu is a cross probability increasing coefficient of an algebraic number which is not updated continuously.
The crossover operation is defined as:
Figure BDA0003776931730000101
wherein X new1 And X new1 Is a new particle generated by the crossover operation; x 1 Is the current particle, X 2 Is X 1 An outer random particle; rnd is a 6-dimensional random vector of one (0,1) interval.
X is to be new1 And X new2 And p g By comparison, if X new1 And X new2 More preferably, the method has the advantages that,then replace p g 。
And (5) searching the optimal fitness of the population according to the fourth step and the third step, and storing and replacing the optimal fitness of the population.
And (5) repeatedly executing the operation of the step three and the operation of the step four for N times on the population according to the step five and the set maximum iteration number 4000.
According to the sixth step: and stopping when the maximum iteration times 4000 are reached in the calculation, sequencing the cost prices of the main materials of the final population from low to high according to an optimization result output module in the figure 4, outputting the design scheme of the amorphous alloy transformer with the lowest main material cost meeting the performance standard, and generating a calculation list.

Claims (5)

1. A particle swarm algorithm-based amorphous alloy transformer parameter optimization design method is characterized by comprising the following steps:
the method comprises the following steps: determining performance parameters according to amorphous alloy dry type transformer users or standard requirements, and setting a value range of a constraint condition and an optimized parameter variable;
step two: generating codes of optimization variables of all dimensions and corresponding actual value arrays, then randomly generating the positions of first generation of particle swarms, and setting a fitness function and a penalty function for the particle swarms;
and (3) processing the constraint condition by adopting a penalty function method, wherein the specific constraint condition processing method and the fitness function are calculated as follows:
the constraints can be expressed as:
g j (X i )≤0,j＝1,...,q
q is the number of constraint conditions;
the constraint violation degree of the ith particle on the jth constraint is expressed as:
G j (X i )＝max{g j (X i ),0},1≤j≤q
the constraint violation of the particle is:
Figure FDA0003776931720000011
due to the difference of the constraint conditions, the constraint violation degree v (X) of some constraint conditions on the individual can occur i ) The conditions that play a dominant role, so the order of magnitude of each constraint is made the same by the normalization method. In the standardization process, firstly, the maximum value of each constraint condition violated by particles in the first generation of population is found;
Figure FDA0003776931720000012
Gmax j for the maximum violating the jth constraint, the other parameters have the same meanings as above. Violation degree standard value Gnorm of ith particle to jth constraint condition j (X i ) The definition is as follows:
Figure FDA0003776931720000021
the constraint violation degree standard value v of the particle is calculated in the whole subsequent iteration process norm (X i ) The definition is as follows:
Figure FDA0003776931720000022
similarly, the fitness function and v are also normalized by a normalization method norm (X i ) Are of the same order of magnitude. First find the minimum and maximum objective function values for the first generation population:
Figure FDA0003776931720000023
Figure FDA0003776931720000024
f(X i ) Is a fitness function value, f min And f max Respectively the minimum value and the maximum value of a first generation population particle objective function;
then, the standard value of the objective function of each generation of the individual x is calculated as follows:
Figure FDA0003776931720000025
f norm (x i ) The standard value of the objective function is obtained, and the other parameters have the same meanings as above;
the fitness function calculation mode is as follows:
fitness(X i )＝f norm (X i )+λ(t)v norm (X i )
fitness(X i ) Is a fitness value, which is called fitness for short; λ (t) is a penalty coefficient; the smaller the fitness of the particles is, the better the representative particle is;
solutions which do not meet the constraint condition are also called infeasible solutions, and the existence of the infeasible solutions is allowed by the iteration process so as to guarantee the global search capability in the early stage. The punishment coefficient at the early stage of iteration is a small value, after the post-iteration stage, the population is converged to a feasible solution, and a large punishment coefficient is needed at the moment. Adjusting the penalty coefficient by adopting a linear increasing mode;
λ(t)＝λ min +(λ max -λ min )/T×t
in the formula of lambda min And λ max Minimum and maximum penalty coefficients, respectively;
step three: comparing the calculated particle fitness with the optimal position of each particle and the optimal position of the whole particles, and replacing and updating the optimal position of each particle and the optimal position of the whole particles if the calculated particle fitness is better than the optimal position of the previous generation;
step four: judging whether the set maximum iteration number N is reached, if not, repeating the steps to update the population positions, and calculating a fitness function value;
step five: and stopping when the maximum iteration number N is reached, and outputting the optimal design scheme in the population.
2. The method for optimally designing the parameters of the amorphous alloy transformer based on the particle swarm optimization algorithm according to claim 1, further characterized in that:
an adaptive particle cross-search strategy is proposed. When the convergence of the particle swarm falls into stagnation, new particles are generated through the intersection among the particles, and a new global optimal solution is searched through the new particles. The cross search among the particles makes up the defects of a particle swarm algorithm solution space search mode;
and (3) an iteration process, wherein each particle is crossed with another random particle by a probability Pr determined by the convergence condition of the current particle swarm, and the probability Pr is defined as:
Pr＝μ×Re
wherein Re is p g Algebra that is not updated continuously; μ is a cross probability increasing coefficient.
The crossover operation is defined as:
Figure FDA0003776931720000031
wherein X new1 And X new1 Is a new particle generated by the crossover operation; x 1 Is the current particle; x 2 Is X 1 An outer random particle; rnd is a 6-dimensional random vector in the interval (0,1);
mixing X new1 And X new2 And p g By comparison, if X new1 And X new2 Preferably, p is replaced g 。
3. The method for optimally designing the parameters of the amorphous alloy transformer based on the particle swarm optimization algorithm, according to claim 1, is further characterized in that:
the optimization target is an economic index, namely the main material cost of the amorphous alloy dry-type transformer:
f(X)＝C h +C l +C co +C cl
wherein, C h Cost for the high voltage winding; c l Cost for low voltage windings; c co The core cost; c cl For clip cost, a smaller objective function represents a better design solution.
4. The method for optimally designing the parameters of the amorphous alloy transformer based on the particle swarm optimization algorithm according to claim 1, further characterized in that: the main technical performance indexes, the material and the process technology and other constraint conditions are as follows:
1) No-load loss: p 0 (X)≤P 0r ；
2) Load loss: p k (X)≤P kr ；
3) No-load current: i is 0 (X)≤I 0r ；
4) Impedance voltage: u shape k min ≤U k (X)≤U k max ；
5) Temperature rise of the low-voltage winding: t is l (X)≤T l max ；
6) Temperature rise of the high-voltage winding: t is a unit of h (X)≤T h max ；
7) Low-voltage winding wire current density:
Figure FDA0003776931720000041
8) Current density of the high-voltage winding wire:
Figure FDA0003776931720000042
9) Magnetic flux density: b m min ≤B m (X)≤B m max ；
10 Efficiency): eta (X) is not less than eta r ；
Wherein P is 0 (X)、P 0r Respectively representing the actual value and the rated value of no-load loss;
P k (X)、P kr respectively representing the actual value and the rated value of the load loss;
I 0 (X)、I 0r respectively an actual value and a rated value of the no-load current;
U k (X) For actual value of short-circuit impedance, U k min 、U k max Respectively taking the minimum and maximum values of the short-circuit impedance;
T l (X) is the actual value of the temperature rise of the low-voltage winding, T l max The maximum limit value of the temperature rise of the low-voltage winding is set;
T h (X) is the actual temperature rise value of the high-voltage winding, T h max The maximum limit value of the temperature rise of the high-voltage winding is set;
J l (X)、J h (X) are the actual current density values of the low-voltage wire and the high-voltage wire, J l max 、J h max Respectively the maximum limit values of the current density of the low-voltage wire and the high-voltage wire;
B m (X) is an actual value of the magnetic flux density of the core, B m min 、B m max Respectively taking the minimum and maximum values of the magnetic flux density of the iron core;
η(X)、η r actual efficiency and rated efficiency, respectively.
5. A particle swarm algorithm-based amorphous alloy transformer parameter optimization design method is characterized by comprising the following steps: the system comprises an amorphous alloy dry type transformer parameter setting module, a particle swarm algorithm, an optimized variable upper and lower boundary setting module, a particle swarm algorithm engine module and an optimized result output module, and the composition structure and the connection relation of the system are as follows:
the amorphous alloy dry-type transformer parameter setting module is used for setting and optimizing parameters such as rated capacity, iron core materials, iron core forms, insulation grades and the like of the amorphous alloy dry-type transformer, and the module allows manual parameter adjustment to meet the actual product requirements;
the particle swarm algorithm and optimization variable upper and lower limit setting module is used for setting parameters such as the number of particles, the dimension, the acceleration factor C1, the acceleration factor C2, the maximum inertia weight, the minimum inertia weight and the like of the particle swarm algorithm and setting the upper limit and the lower limit of the iron core stacking thickness, the upper limit and the lower limit of the low-voltage layer number, the upper limit and the lower limit of the high-voltage line width (line thickness) and the upper limit and the lower limit of the low-voltage line width (line thickness);
the particle swarm algorithm engine module is used for performing particle swarm algorithm-related calculation, including random coding, generating an initial variable population, calculating the fitness value of a target function value (a target function with a penalty function) and a population individual of the generated initial population, updating the population position, finally generating a new offspring population, and performing iterative calculation until a convergence condition is met and stopping the calculation;
the optimization result output module is used for outputting a plurality of amorphous alloy transformer parameter optimization schemes which are obtained by the particle swarm algorithm and meet the standard requirements of customers or performance parameters, and selecting the optimal amorphous alloy transformer parameter optimization design scheme according to the requirements of low manufacturing cost and low loss.
CN202210919265.6A 2022-08-02 2022-08-02 Amorphous alloy transformer parameter optimization design method based on particle swarm optimization Withdrawn CN115203857A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210919265.6A CN115203857A (en) 2022-08-02 2022-08-02 Amorphous alloy transformer parameter optimization design method based on particle swarm optimization

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210919265.6A CN115203857A (en) 2022-08-02 2022-08-02 Amorphous alloy transformer parameter optimization design method based on particle swarm optimization

Publications (1)

Publication Number Publication Date
CN115203857A true CN115203857A (en) 2022-10-18

Family

ID=83586905

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210919265.6A Withdrawn CN115203857A (en) 2022-08-02 2022-08-02 Amorphous alloy transformer parameter optimization design method based on particle swarm optimization

Country Status (1)

Country Link
CN (1) CN115203857A (en)

Similar Documents

Publication Publication Date Title
CN110021940B (en) Capacitor optimal configuration method based on improved particle swarm optimization
CN117291004A (en) Transformer parameter optimization design method based on MSPBO algorithm
CN113972667A (en) Power distribution network reactive power optimization method based on one-dimensional convolutional neural network
CN111614110B (en) Receiving-end power grid energy storage optimization configuration method based on improved multi-target particle swarm optimization
CN114091330A (en) Optimal design method for medium and high speed grade magnetic gear of high-power wind electromagnetic gear box
CN112560331B (en) Energy-saving and material-saving optimization design system and method for amorphous alloy dry type transformer
CN105956332B (en) transformer electromagnetic optimization design method based on expert system
Versèle et al. Multiobjective optimal design of transformers for isolated switch mode power supplies
CN109586278B (en) Method for evaluating power supply capacity of alternating current-direct current hybrid power distribution network
CN115203857A (en) Amorphous alloy transformer parameter optimization design method based on particle swarm optimization
CN114036669A (en) Optimization design method for low-speed magnetic gear of high-power wind electromagnetic gear box
CN106257477B (en) A kind of intermediate frequency amorphous alloy transformer optimization method based on multi-objective genetic algorithm
CN112052601B (en) Optimal fraction radar chart-based power transmission and distribution network voltage sequence optimization method
Versele et al. Multiobjective optimal design of high frequency transformers using genetic algorithm
CN108053250B (en) Genetic algorithm-based transformer price optimization method and device
Liu et al. TOC optimization design of amorphous metal core distribution transformer based on NSGA-II
CN115859518A (en) Three-dimensional wound core transformer optimization design method based on improved gull optimization algorithm
CN117291005A (en) Transformer parameter optimization method based on SPBO algorithm under multi-constraint condition
Georgilakis et al. A neural network framework for predicting transformer core losses
DingPing et al. Power system reactive power optimization based on MIPSO
Georgilakis et al. Spotlight on transformer design
Hooshmand et al. Optimal choice of fixed and switched capacitors for distribution systems by genetic algorithm
Ding et al. Multi-objective optimal design of traction transformer using improved NSGA-II
Tian et al. Deep Reinforcement Learning for DC-DC converter parameters optimization
Kul et al. FEM-based modeling and optimization of dry-type transformers with metaheuristic algorithms

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
WW01 Invention patent application withdrawn after publication
WW01 Invention patent application withdrawn after publication

Application publication date: 20221018