Deep Learning as an Early Detection System for Rotary Percussion Drilling Malfunctions

To mine in the underground, method of blasting and blast hole drilling methods are mainly, widely accepted. The done with rotary percussion drill. However, there problems terms difficulty operating mining cost resulting from its failure occurs, thus it is hard for companies to find a way underground efficiently, profitability, safely. From this background, necessary build early detection system drill bit failure. This needs technology CNN (Convolutional Neural Network Smart Mining, which process using information, autonomy, improve safety, reduce costs, site productivity. In research, vibration transmitted as acceleration waveform used input data building system. collected replacing kinds diameter or condition. developing model introduced detect difference between Normal other something error. For Firstly, batch training make recognize pattern. Secondly, validation confirms correct answer rate against data, then, test practiced. Finally, by comparing each accuracy phase 4 types models built different ideal found.