OPTIMIZATION OF ELECTRICITY CONSUMPTION IN COMMERCIAL BUILDINGS: REGRESSION MODEL APPROACH WITH DATA LIMITATIONS

• Main Author: Juliani, Elsha
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/87627
• Institution: Institut Teknologi Bandung

Electricity consumption continues to increase over time, and this has also become an important issue in various sectors, one of which is the building sector. The hospitality sector is one part of the commercial building sector that heavily relies on electricity in its services. Many studies have been conducted to improve the optimization and efficiency of electricity consumption. However, most of the research conducted focuses on issues primarily raised from problems in high-star hotels, whereas the problems in low-star hotels are not the same as those in high-star hotels. There are various limitations, such as technology, infrastructure, and buildings that pose challenges for low-star hotels. Hotel Sawunggaling, as a 1-star hotel, faces the same challenges. With the limited data available, a new approach was taken to optimize the electricity consumption at the hotel. Based on the research conducted, parameters such as the type of hotel room, the type of day, and the month affect the daily electricity consumption of the hotel. By using a regression model approach and utilizing limited data, the data was processed and developed into a daily kWh database for Hotel Bumi Sawunggaling. The regression model results were evaluated with a MAPE of 2.69% and an R² accuracy of 0.97. With this model data, predictions were made using the Random Forest and XG Boost methods, resulting in an R² accuracy of 0.93, an RMSE of 10, and an MAE of 8.5. Next, from the data obtained by implementing savings strategies such as the efficiency of air conditioning and lighting usage, as well as wise user application, savings can be achieved amounting to 8% of total consumption.