The utilization of solar energy in Indonesia holds great potential due to high sunlight intensity throughout the year. However, partial shading remains a major challenge in photovoltaic (PV) systems, as it causes efficiency loss by generating multiple maximum power points (MPPs). This study designs and implements a Maximum Power Point Tracking (MPPT) Boost Converter using a clustering method with the Modified Perturb and Observe (P&O) algorithm on a Floating Solar Farm. PV panels are divided into two clusters: Cluster 1 for uniform conditions and Cluster 2 for shading conditions. Test results show that clustering enables simultaneous operation of both clusters, ensuring more stable and efficient power output. Under uniform conditions, the highest power output reached 30.39 W at 12:00. Under 60% shading, the system still produced 6.20 W at 12:00 and 6.01 W at 15:00. The highest total output was 34.18 W at 10:00 under 60% shading. This demonstrates that clustering significantly improves power extraction efficiency compared to single-topology systems by optimizing two peak points concurrently. Therefore, this approach offers an effective solution to shading problems in solar panel systems, especially in floating solar farm applications.
Keywords: MPPT, Boost Converter, Clustering, Modified P&O, Floating Solar Farm, Partial Shading.