Perancangan Algoritma Prediksi Iluminansi di Ruang Kuliah TN7 DTNTF UGM
Mohammad Itqon Alexander, Prof. Dr. Ir. Faridah, S.T., M.Sc., IPU., ASEAN Eng.; Ir. Memory M. Waruwu, S.T., M.Eng., IPM.
2026 | Skripsi | FISIKA TEKNIK
High energy consumption contributed to environmental issues, making energy conservation essential, particularly in the building sector. Lighting control can be implemented through indoor illuminance monitoring. However, installing physical sensors remains challenging. The Building Management System (BMS) of DTNTF UGM already provides outdoor illuminance data and the status of light group switches. Therefore, this study designs an illuminance prediction algorithm for lecture hall TN7 based on these available inputs.
The algorithm was constructed based on the room characteristics and the lighting system configuration. Measurements of natural and artificial lighting were conducted at 32 points within the room. The algorithm was evaluated under various lighting conditions. The algorithm was implemented in a Python program and integrated into the soft sensor module embedded in the BMS DTNTF UGM.
The results showed a higher daylight factor at measurement points near the windows with greater standard deviations. Validation of the daylight factor produced an MAE of 9 lux. The illuminance contribution from artificial lighting was higher at measurement points corresponding to the positions of the light groups. Algorithm testing resulted in an MAPE of 14% for overcast skies, which is below the threshold detectable by humans. In total, 81.25% of measurement points achieved an MAPE of less than 20%. The algorithm can be deployed in the BMS DTNTF UGM, with prediction results visualised as a heatmap on the GUI dashboard. The soft sensor module operates in real-time for outdoor illuminance data, while dummy data is used for the status of light group switches.
Kata Kunci : Pencahayaan, iluminansi, daylight factor, algoritma prediksi