PREDIKSI KERUSAKAN KINERJA STRUKTUR BANGUNAN GEDUNG MENGGUNAKAN JARINGAN SARAF TIRUAN DENGAN METODE PURELIN

Abstract

An earthquake is one of the natural events whose magnitude cannot be predicted. This is can happen because the direction of the earthquake work depends on the movement of the soil that supports it. This is generally the biggest threat to construction, especially buildings. Buildings are expected to be built because they can use even a small area of land. However, over time, it is common for buildings to collapse due to earthquakes, so a more detailed analysis is needed to design a better earthquake-resistant building. Time history analysis is one of the analyzes used to evaluate buildings against earthquakes. However, time history analysis has a weakness, namely the duration of the analysis tends to be long, so determining whether a structure is still able to function according to plan is difficult to measure. Analysis of artificial neural networks by utilizing structural response data is expected to be able to predict the structural performance of building structures. The purelin method reads data linearly but in this case, predicts based on previous data or is known as the Backpropagation Analysis method.