Soc battery using coulomb counting code

Can someone help me ro modify this code " Calculate batery Soc using coulomb counting method "

% Assuming synthData is already defined with fields `t` and `I`
time = synthData.t;  % Time in seconds

Unable to resolve the name 'synthData.t'.

current = synthData.I; % Current in Amperes

% Constants
C_battery = 2000; % Battery capacity in milliampere-hours (mAh)
initial_SOC = 100; % Initial State of Charge in percentage
C_rate = C_battery / 1000; % Convert capacity from mAh to Ah

% Initialize State of Charge (SoC)
SOC = zeros(1, length(current)); % Preallocate SoC array
SOC(1) = initial_SOC; % Set initial SoC

% Calculate the change in state of charge using Coulomb counting
for i = 2:length(current)
    % Time step
    dt = time(i) - time(i-1); % Time interval in seconds
    
    % Calculate the change in charge (in Ah)
    delta_Q = current(i) * dt / 3600; % Convert current to Ah
    
    % Update the State of Charge
   
    SOC_estim = ((1-(1/(x*3600))*cumtrapz(Time, -Current))*100); % Convert Ah change to percentage
    SOC(i) = max(0, min(SOC(i), 100)); % Ensure SoC is between 0 and 100
end

% Plot the results
figure;
plot(time, SOC, '-o', 'LineWidth', 1.5, 'MarkerSize', 4);
xlabel('Time (seconds)');
ylabel('State of Charge (SoC) [%]');
title('Battery State of Charge over Time');
grid on;
xlim([min(time) max(time)]); % Set x-axis limits
ylim([0 100]); % Set y-axis limits

[Umar]

Hi @Anis Aoudia,

What is your source of information, could you provide the link to “ Calculate batery Soc using coulomb counting method” which will help resolve this issue, “ time = synthData.t;”

Answer 1

[Samar]

Hello @Anis Aoudia,

By modifying I believe you are trying to optimize the code such that the execution time is reduced. So, I tried to find which part of the code is taking the most amount of time using MATLAB’s Profile function. Here is the MathWorks documentation for more information:

www.mathworks.com/help/releases/r2024a/matlab/ref/profile.html

Since the time and current values were not provided, I assumed random vectors of different sizes (10,1000 and 1e5). From this I inferred that the for loop takes the most amount of time to execute. So, I tried vectorization of the code snippet present in the for loop.

(In the code provided an undefined variable “x” was used to calculate value of “SOC_estim”. I assume it’s a typo and I used “C_rate” for calculation).

Vectorization seems to improve the performance of code if size of time and current vectors cross 1000. For vectorization I used the element wise operators. Their usage is explained in the link below:

www.mathworks.com/help/releases/r2024a/matlab/matlab_prog/matlab-operators-and-special-characters.html

To calculate the difference between adjacent elements of a vector, the function “diff” can be used. This function can be used to find “dt”. Refer to the link given below for more information.

www.mathworks.com/help/matlab/ref/double.diff.html

Refer the link given below to determine the State of Charge by Coulomb Charging.

www.mathworks.com/help/releases/r2024a/simscape-battery/ref/socestimatorcoulombcountingvariablecapacity.html#mw_b1a55a66-c359-40ec-ac65-f2d269f60f59

Let me know in case of doubts if any.