PV power estimation for solar panel systems

Hi, I work on a project that is related to solar panel power estimation. Simply, I have environmental variables like temperature, humidity, radiation, etc. By using each of these sensor values, their individual effect on power as a formulation is requirred to predict power value later on. If I have x and y variables, how can I come up with a polynomial or logarithmic approach that explains x-y relation?

Thanks

Answer 1

[Sayan]

Hi @SÜLEYMAN ÖZDUVEN,

I understand from your query that it is required to obtain a polynomial or logarithmic approach using the values of the output PV power as a function of temperature, humidity, radiation, etc. In Simulink, the variation power with temperature and radiation can be obtained. However, variation of power with humidity is not supported. The below-mentioned steps can be followed to obtain the result.

1. Use the block PV Array (available at Simscape/Electrical/Specialized Power Systems/Sources) in Simulink to simulate the behavior of a PV array and connect it to the desired circuit model.
2. Provide the desired waveform/values of irradiance "Ir" and temperature "T" as the input to the block.
3. Connect the measurement port "m" to a "To File" block (available at Simulink/Sinks). Change the "Save format" field of the block as "Array". The measurement port output is a five-element vector. They can be found in the following PV Array documentation.
4. Run the simulation.
5. Load the ".mat" in MATLAB command window.

The array variable loaded in the workspace contains the values of "PV voltage," "PV current," "irradiance," and "temparature" in its 2nd, 3rd, 5th, and 6th rows, respectively, for each sample sample time stored in the 1st row of the array. The "PV voltage" and "PV current" can be extracted from the array and multiplied to obtain the power output ("y" variable) of the PV array. After obtaining the power, the relation with the temperature and irradiance ("x" variables) can be obtained using polynomial or logarithmic approach in the following way:

• Polynomial Approach: To obtain polynomial regression "polyfit" and "polyval" functions can be used like the following code snippet.

coefficients = polyfit(x, y, degree);%obtain the coefficients of the polynomial according to the required degree

% Generate the polynomial equation
polynomial = polyval(coefficients, x);

% Plot the polynomial curve
plot(x, polynomial);

• Logarithmic Approach: To obtain logarithmic regression "fittype" and "fit" functions can be used like the following code snippet.

x_log = log(x);
y_log = log(y);

% Define the logarithmic regression equation
logarithmicEquation = fittype('Desired function', 'Name', 'Value');%Define the the rquired Desired function and other parameters as required

% Fit the logarithmic regression model
fitResult = fit(x_log, y_log, logarithmicEquation);\
%Define the required logarithmic curve and plot the result

Further informations on the above used blocks/functions can be found in the following MATLAB documentations:

• PV Array block:https://www.mathworks.com/help/releases/R2021b/physmod/sps/powersys/ref/pvarray.html?s_tid=doc_ta
• "polyfit" function: https://www.mathworks.com/help/releases/R2021b/matlab/ref/polyfit.html
• "polyval" function: https://www.mathworks.com/help/releases/R2021b/matlab/ref/polyval.html
• "fittype" function: https://www.mathworks.com/help/releases/R2021b/curvefit/fittype.html
• "fit" function: https://www.mathworks.com/help/releases/R2021b/curvefit/fit.html

Hope this answers the query.