Simulink Grid Converter Optimization
Budget: $10 – $30 USD
I already have a fully-functional MATLAB/Simulink model of a grid-connected multilevel converter; now I need the model to think for itself. The goal is to embed two metaheuristic algorithms—Harmony Search (HS) and the Wild Horse optimizer—so the system can automatically fine-tune the PI controller gains and squeeze the best possible performance from the converter.
Here is what the job involves:
• Insert both HS and Wild Horse routines into the existing .slx/.m framework.
• Use them to search the PI gain space and minimise the usual pain points (THD and voltage balance) with the clear target of giving me a better-behaved controller.
• Keep the current NPC/CHB topology intact; only the optimisation layer changes.
• Return a concise “before vs. after” snapshot that proves the improvement—plots, numeric tables, whatever best shows the gains.
Deliverables
1. Updated .slx model and any supporting .m scripts, ready to run without additional libraries.
2. A short comparison report (PDF or live script) illustrating performance prior to optimisation and after each algorithm is applied.
Acceptance will be based on three checks: the file runs cleanly, the optimisation loops converge, and the plotted metrics show a tangible performance uptick.
If you live in MATLAB/Simulink, speak fluent power electronics, and have hands-on experience implementing custom heuristic optimisers, I look forward to seeing how you can push this converter a bit closer to perfect.
Here is what the job involves:
• Insert both HS and Wild Horse routines into the existing .slx/.m framework.
• Use them to search the PI gain space and minimise the usual pain points (THD and voltage balance) with the clear target of giving me a better-behaved controller.
• Keep the current NPC/CHB topology intact; only the optimisation layer changes.
• Return a concise “before vs. after” snapshot that proves the improvement—plots, numeric tables, whatever best shows the gains.
Deliverables
1. Updated .slx model and any supporting .m scripts, ready to run without additional libraries.
2. A short comparison report (PDF or live script) illustrating performance prior to optimisation and after each algorithm is applied.
Acceptance will be based on three checks: the file runs cleanly, the optimisation loops converge, and the plotted metrics show a tangible performance uptick.
If you live in MATLAB/Simulink, speak fluent power electronics, and have hands-on experience implementing custom heuristic optimisers, I look forward to seeing how you can push this converter a bit closer to perfect.