Identifikasi Jenis Asap Menggunakan Spektrofotometer Dan Jaringan Syaraf Tiruan

Abstract

Spectroscopy method can be used to identify the type of gas by passing light into the sample gas and light are described using a monochromator. The resulting light captured using the detector that will produce different spectra for each gas. The light source can use incandescent lamps or using Light Emitting Diode (LED). If the tested gas is a type of gas that is difficult to be taken, such as the type of toxic gas, gas from the output of the volcano, so this way you will have trouble, and the data obtained is not real time. In this study designed a system of identification of gas or smoke in the air using a spectrophotometer using solar light. The spectrum of light that has been absorbed by the gas or fumes captured using a telescope, then described using a monochromator produces gray level curve representing the absorption of each wavelength of light. The sample used in this study is the smoke of burning oil, sulfur and dry leaves. The curve of each sample generated, analyzed and identified the type of smoke using Artificial Neural Network (ANN) with backpropagation training algorithm. The architecture consists of input layer, hidden and output. Input node number is 197, the number of neurons in the first hidden layer is 150, the number of neurons in the second hidden layer is 20, and the number of neurons in the output layer is 4. In this ANN learning process entails iterating as many as 900 epoch and the results can distinguish the smoke oil , sulfur and dry leaves. The success rate of 70% for oil and dried leaves, 80% for sulfur.