PURWARUPA SISTEM DETEKSI DINI KEBAKARAN BERBASIS ELECTRONIC NOSE; PROTOTYPE OF EARLY FIRE DETECTION SYSTEM BASED ON ELECTRONIC NOSE
WICAKSANA, SATRIA, Lelono, Danang
2015 | Skripsi | FMIPA UGMFire is one of the disaster phenomenon that has killed many people every year in worldwide. It is important to develop existing fire detection system in order to reduce losses caused by fire. In this study, the electronic nose designed from a combination of five pieces of metal oxide gas sensors applied to a prototype early fire detection system. The signals obtained from the same sample has a high correlation, while the signals from the different samples exhibit a unique patterns. Therefore, the back propagation neural network as a pattern recognition used on the system. The result, by using signals on the odor on state for 20 seconds, the accuracy achieved is 98.33% for the classification of 3 groups, namely fire (wood, paper and plastic), non-fire (cigarettes and repellent) and clear (clean air), while the accuracy of 93.33% is achieved for classification 6 groups to each sample.
Kata Kunci : electronic nose; early fire detection; neural network; back propagation
