【拟合】基于粒子群算法实现热瞬态阻抗曲线拟合附matlab代码

✅作者简介：热爱科研的Matlab仿真开发者，修心和技术同步精进，matlab项目合作可私信。

🍎个人主页：Matlab科研工作室

🍊个人信条：格物致知。

⛄ 内容介绍

文章阐述MATLAB的曲线拟合工具箱(Curve Fitting Toolbox)的交互式图形界面和程序调用两种方式的使用方法,并以某柴油发动机稳态工况下的试验数据绘制万有特性图为例,介绍利用曲线拟合工具箱进行曲面拟合的步骤和方法,并给出关键代码,探讨利用该工具箱改进拟合效果和绘图质量.对拟合结果的数理统计校验表明拟合结果完全符合要求,MATLAB的曲线拟合工具箱为发动机特性的曲面拟合提供更大的便利.

⛄ 部分代码

% Your digitized plot goes here. You can use e.g. Plot Digitizer or many

% other digitizers available on Matlab Central.

% Here the trench IGBT module SKM145GB066D from Semikron will serve as an

% example. Note that the thermal chain parameters are given explicitely for

% this particular module but this is not always the case. In fact, it's usually not

% the case. More on thermal modelling:

% http://www.plexim.com/download/documentation .

x_igbt_zth = [1.02e-05, 1.46e-05, 1.93e-05, 2.52e-05, 3.14e-05, 3.92e-05,...

    5.03e-05, 6.71e-05, 8.89e-05, 0.0001185, 0.000159193, 0.000215443, 0.00029157,...

    0.000403431, 0.000541969, 0.000722727, 0.000970911, 0.00132372, 0.00184512, 0.002553,...

    0.00350648, 0.00514681, 0.00733474, 0.0112534, 0.01664, 0.0225198, 0.0332994, 0.0444054,...

    0.0632823, 0.0831518, 0.104528, 0.135335, 0.171386, 0.206112, 0.272833, 0.361154,...

    0.492388, 0.65661, 0.831518, 0.992647];

y_igbt_zth = [0.000438281, 0.000600275, 0.000779386, 0.00100595, 0.00123818, 0.00152402,...

    0.00189824, 0.0024793, 0.0031811, 0.00408155, 0.0052994, 0.00675922, 0.00877605,...

    0.0111273, 0.0141926, 0.0176776, 0.0216298, 0.026623, 0.0333577, 0.0410584,...

    0.0487691, 0.0603848, 0.0747672, 0.0936806, 0.117378, 0.138597, 0.168579, 0.195541,...

    0.228164, 0.252384, 0.271013, 0.28419, 0.292749, 0.299781, 0.30336, 0.30336,...

    0.30336, 0.305166, 0.30336, 0.30336];

% rng('shuffle');

asur = 0; % 1 --> the asynchronous update rule; 0 --> the synchronous update rule

evaporation_constant = 1; % Static optimization problem (SOP)

swarm_size = 50;

init_position = 3; % note that parameters are coded using exponential function;

% the initial area covered by the swarm is then [0.001, 1000];

% this should work for most thermal models of

% semiconductor switches and heat sinks in power electronic systems. If the

% optimizer fails to identify your device, you should try setting

% init_position = 5 .

init_speed = 1;

diversity_limit = 0; % Static optimization problem (SOP)

velocity_clamping = 2;

num_of_iter = 10000; % don't worry -- it will take seconds to complete the task at hand

% Constricted PSO

correction_factor = 2.05;

Kappa=2/abs(2-2*correction_factor-sqrt((2*correction_factor)^2-8*correction_factor));

% The problem

num_of_D = 6; % 6 parameters of the Foster-type RC ladder network to be identified

% More declarations

swarm = zeros(swarm_size,4,num_of_D);

swarm_diversity = zeros(1,num_of_D);

swarm_dir = zeros(1,num_of_D);

%% Initial swarm position and velocity

for index=1:swarm_size,

    swarm(index, 1, :) = init_position*(rand(num_of_D,1)-0.5)*2;

    swarm(index, 2, :) = init_speed*(rand(num_of_D,1)-0.5)*2;

end

swarm(:, 4, 1) = 10e10;          % initial best value --> some value out of practical limits

%% The swarm is turned on

for iter = 1 : num_of_iter,

    for n = 1 : swarm_size,

        r1 = 10^swarm(n, 1, 1); % resistances

        r2 = 10^swarm(n, 1, 2);

        r3 = 10^swarm(n, 1, 3);

        c1 = 10^swarm(n, 1, 4); % capacitances

        c2 = 10^swarm(n, 1, 5);

        c3 = 10^swarm(n, 1, 6);

        %% In this example the 3rd order Foster chain is used.

        % Particles are rated here using sum of squared errors.

        % Obviously you can define your own cost function.

        fitness = 0;

        for i = 1:numel(x_igbt_zth),

⛄ 运行结果

⛄ 参考文献

[1]黄兵锋, 解方喜, 傅佳宏. MATLAB曲线拟合工具箱在发动机特性拟合中的应用[J]. 湖北文理学院学报, 2014, 35(5):5.

⛄ 完整代码

❤️部分理论引用网络文献，若有侵权联系博主删除

❤️ 关注我领取海量matlab电子书和数学建模资料