Real-Time Misfire Detection of Large Gas Engine Using Big Data Analytics

In this work application of fuzzy deep learning algorithm to classify vibration signals of a Deutz MWM 8 cylinder large gas engine is presented. The main contribution is to identify the misfire characteristics in realtime embedded microcontroller environment. The substance of the scheduled fuzzy system is a stratified deep neural network that obtains information from both fuzzy and neural representations. The misfire characteristics learnt from these two various views are convolved, forming the ultimate signal to be classified.