Project Description
I need to reproduce the complete control model from a published paper on grid-connected photovoltaic inverters—specifically the PI + Resonant current controller—and then push the work further by adding an intelligent ANN-based control structure. Everything must be built and run in MATLAB/Simulink so that I can validate the original results first, then make a clear side-by-side comparison between the classical and the new approach.
The task starts with an exact replication of the paper: identical parameters, same test grid, and matching waveform plots. Once that baseline is confirmed, you will design, train, and integrate an artificial neural network controller, tune it, and show how it behaves under identical scenarios. I am especially interested in response time, harmonic distortion, and efficiency improvements, but feel free to highlight any other advantages your model reveals. The simulation should be flexible enough to toggle standard test grid settings, weather variability, and load changes; if you already have reusable blocks or data sets for these, great—otherwise please create them.
Deliverables
• Fully commented Simulink model of the original PI + Resonant control system
• ANN-enhanced controller integrated into the same model (clearly modularised)
• MATLAB scripts/functions for training and running the ANN
• A concise technical report (plots, numerical tables, discussion of results, and your methodology)
• Any custom data or lookup tables used in the study
Acceptance criteria
• Baseline simulation matches key figures from the reference paper within a reasonable margin
• ANN controller runs stably over all tested scenarios and demonstrates measurable improvement in at least one target metric
• All files open and execute without errors in the stated MATLAB/Simulink version
If you have already tackled similar grid-connected inverter problems or hybrid control schemes, let me know—sample outputs or references are welcome.