<![CDATA[The demand for electricity in Indonesia is continuously increasing due to the growing economy over time. In accordance with Presidential Instruction No. 10 of 2005 and Presidential Regulation No. 5 of 2006, regulations have been issued regarding energy conservation and national energy usage policies. Therefore, this paper discusses an electricity energy audit and predicts the electricity load at the PT. Telkom Indonesia Building in Lhokseumawe using Artificial Neural Network (ANN) algorithms. Energy audit involves the inspection, survey, and analysis of energy flows to identify energy-saving opportunities in buildings, aiming to reduce the input energy into the system without compromising system output. The research aims to identify electricity energy-saving opportunities. Additionally, the paper conducts predictions of electricity usage before and after the audit using the Artificial Neural Network (ANN) algorithm. Therefore, an audit of this load is needed. After inspecting the location, it was identified that the air conditioning units used exceed the required AC capacity. Consequently, a proposal for new devices is necessary to achieve potential electricity usage savings in the building. Based on the training and validation of the neural network using the Bilayered Neural Network (BNN) model with 3 layers, the most suitable model was obtained. The obtained values include an RMSE of 1210.7, R-Squared of 1.00, MSE of 1.4658e+06, MAE of 901.19, Prediction Speed ~1500obs/Sec, and Training Time of 4.8719 Sec. The results indicate potential savings of up to 17,023 kWh/month or 163,327 kWh/year. The required Payback Period to recover the invested capital is estimated at 14 months.]]>
