Analisis Sifat Magnetik Kalsium Ferit yang Disintesis Menggunakan Metode Metalurgi Serbuk

Abstract

Sintesis kalsium ferit telah dilakukan dengan metode metalurgi serbuk.  Kalsium ferit dihasilkan dari bahan dasar  Fe2O₃ hasil oksidasi dari pasir besi dan CaCO₃ yang diproduksi oleh Merck. Variasi persentase massa CaCO₃ yang digunakan yaitu 5%, 10%, 15%, 20%, 25% dan 30%.  Variasi persentase massa CaCO₃ dilakukan untuk melihat pengaruhnya terhadap sifat magnet sampel yang dihasilkan.  Hasil sintesis dikarakterisasi dengan menggunakan X-Ray Diffraction (XRD) dan Vibrating Sampel Magnetometer (VSM). Hasil pola difraksi dengan XRD menunjukkan fasa kalsium ferit (Ca-Fe-O) yang terindikasi dari pola puncak XRD antara lain CaFe₂O₄ dan CaFe₄O₇ pada persentase massa 5% dan 10%. Sifat magnetisasi saturasi (M_S) yang dihasilkan dari karakterisasi VSM yaitu 0,58 em/g pada persentase massa 5% dan 0,68 em/g pada persentase massa 10%. Magnetisasi remanen (M_R) yang dihasilkan pada persentase massa 5% yaitu 0,23 em/g dan pada persentase massa 10% yaitu 0,24 em/g. Nilai medan koersivitas (H_C) yang berada pada range 19936 A/m hingga 30303 A/m.

Research on daily rainfall predictions have made by using artificial neural networks with some backpropagation and radial basis training functions. This study used daily rainfall data from the Meteorology Climatology and Geophysics Agency in Class II of the Minangkabau International Airport Padang Pariaman from 2008 to 2018. The purposes of the study is to compare the predicted performance of rainfall in Backpropagation and Radial Neural Networks and determine which one the best artificial neural network architecture for rainfall predictions at Minangkabau International Airport is. For the backpropagation method, optimization is performed on the number of hidden layers, the number of neurons in the hidden layer, the transfer function, the training function and the amount of input data on the training data. For the radial basis optimization method is performed on the number of hidden layer neurons, the amount of input data on the training data and the spread value. From this study found the best results for the backpropagation method were obtained with trainlm and architectural training functions (60-70-6-1) with a prediction accuracy level of 86.4876%. The best results for the radial basis method by value of a spread is 0.01 with architecture (60-120-1) and a predictive accuracy rate of 95.3107%. Thus the best method for the prediction of daily rainfall in the area of the Minangkabau International Airport is the radial basis method.