How to change an offline parameter estimation to online task ?

I have a PV mathematical model contain unkown parameters

I have used an algorithm (metaheuristics ) and extract these parameters using measured data.

The problem is that these parameters are not constant, so each time it changes due to external conditions so i have to estimate it again and again ..

I have read about recursion algorithm (finite and infinite) which i want to use in my case to estimate the parameters and update it continuesly based on online data

How to perform such approach keeping same algorithm and adapt it for online estimation?

Answer 1

[MULI]

Hi Yasser,

As you want to use a recursive algorithm to continuously estimate and update the parameters in real-time, you can follow these steps:

• Obtain an initial estimate of the parameters using your existing metaheuristic algorithm.
• Implement a recursive estimation method such as Recursive Least Squares (RLS) or a Kalman Filter.
• Integrate the recursive algorithm into your current workflow, allowing for real-time updates.
• Use the command-line functions or Simulink blocks provided by the “System Identification Toolbox” to set up the online estimation process.

For more information on “System Identification Toolbox” and online parameter estimation you may refer to these links:

https://in.mathworks.com/help/releases/R2020a/ident/index.html

https://in.mathworks.com/help/releases/R2020a/ident/online-estimation.html