How to implement PSO algorithm for sizing of supercapacitor and battery in HESS using Matlab

The PSO Code for HESS Sizing of Ultracapacitors and Battery in Electric Vehicle

[Sam Chak]

Hi @Sri Hari R

Please provide the following information:

• formulas for the supercapacitor and battery (preferably in LaTeX)
• constraints of the sizing requirement
• the parameters and dynamics in HESS
• prefered settings in PSO

If you have the code, do you know how to modify it to suit your application?

[Sri Hari R]

Yes we have got a objective function

we want to minize the weight and split the power based on it

[Marko Stokic]

Hi, @Sam Chak

I am a PhD student working on this topic and it would be very helpful if I could get the code for my further research.

Since Sri Hari R has been inactive for more than 10 months, I am asking if you have any code on this topic or at least the objective function and constraints necessary to develop this model?

Thank you very much!

[Sam Chak]

Hi Marko, I don't have the HESS code. The PSO algorithm is available in the Optimization Toolbox.

I think the general idea is that the size of the battery affects the performance of the system. Thus, I think there should be 3 components for an unconstrained problem:

1. The dynamics of system.

2. The objective function that measures the performance (maybe power and energy dissipated). Can be a quadratic function.

3. The optimizer, e.g., PSO, GA, ACO, ABC, fminunc, fmincon, fminsearch.

[Marko Stokic]

@Sam Chak, thank you very much for the explanation!

Answer 1

[Anshuman]

PSO is a heuristic algorithm, and the actual implementation details can vary widely depending on the specific problem being solved. Below is a simplified example of a MATLAB code snippet that uses PSO to size hybrid energy storage systems (HESS) consisting of ultracapacitors and batteries for an electric vehicle (EV).

% PSO parameters
numParticles = 30; % Number of particles in the swarm
maxIterations = 100; % Maximum number of iterations
inertiaWeight = 0.7; % Inertia weight
cognitiveComponent = 1.5; % Cognitive component constant
socialComponent = 1.5; % Social component constant

% Problem-specific parameters
% Define the ranges for ultracapacitor size and battery size
ucapRange = [10, 1000]; % Range of ultracapacitor size in Farads
battRange = [1, 100]; % Range of battery size in kWh

% Initialize the swarm
position = zeros(numParticles, 2); % Each particle has a position [ucapSize, battSize]
velocity = zeros(numParticles, 2); % Velocity of the particles
pBest = zeros(numParticles, 2); % Best position of each particle
pBestCost = inf(numParticles, 1); % Best cost for each particle
gBest = zeros(1, 2); % Global best position
gBestCost = inf; % Global best cost

% Initialize positions and velocities
for i = 1:numParticles
    position(i, :) = [rand*(ucapRange(2)-ucapRange(1))+ucapRange(1), rand*(battRange(2)-battRange(1))+battRange(1)];
    velocity(i, :) = [0, 0]; % Start with zero velocity
end

% PSO main loop
for iter = 1:maxIterations
    for i = 1:numParticles
        % Evaluate the cost of the current position
        cost = objectiveFunction(position(i, 1), position(i, 2));
        
        % Update personal best
        if cost < pBestCost(i)
            pBest(i, :) = position(i, :);
            pBestCost(i) = cost;
        end
        
        % Update global best
        if cost < gBestCost
            gBest = position(i, :);
            gBestCost = cost;
        end
    end
    
    % Update velocity and position of particles
    for i = 1:numParticles
        velocity(i, :) = inertiaWeight * velocity(i, :) ...
                        + cognitiveComponent * rand * (pBest(i, :) - position(i, :)) ...
                        + socialComponent * rand * (gBest - position(i, :));
        position(i, :) = position(i, :) + velocity(i, :);
        
        % Ensure the particles stay within the defined ranges
        position(i, 1) = max(min(position(i, 1), ucapRange(2)), ucapRange(1));
        position(i, 2) = max(min(position(i, 2), battRange(2)), battRange(1));
    end
    
    % Display iteration information
    disp(['Iteration ' num2str(iter) ': Best Cost = ' num2str(gBestCost)]);
end

% Objective function (to be defined according to the problem)
function cost = objectiveFunction(ucapSize, battSize)
    % Define the cost function based on the HESS sizing criteria
    % This is a placeholder function; you should define your own cost function
    % based on energy, power requirements, cost, weight, etc.
    cost = (ucapSize * 0.5) + (battSize * 200); % Example cost function
end