Predictive Maintenance Using Deep Learning

Overview

Predictive maintenance allows equipment operators and manufacturers to assess the condition of machines, diagnose faults, and estimate time to failure. Because machines are increasingly complex and generate large amounts of data, many engineers are exploring deep learning approaches to achieve the best predictive results. 

In this talk, you will discover how to use deep learning for:

• Anomaly detection of industrial equipment using vibration data 
• Condition monitoring of an air compressor using audio data

 You’ll also see demonstrations of:

• Data Preparation: Generating features using Predictive Maintenance Toolbox™ and extracting features automatically from audio signals using Audio Toolbox™
• Modeling: Training audio and time-series deep learning models using Deep Learning Toolbox™

Highlights

Explore deep learning approaches to predictive maintenance by detecting anomalies and identifying faults in industrial equipment sensor data.

About the Presenters

Rachel Johnson is the Product Manager for Predictive Maintenance Toolbox at MathWorks. Previously, she was a Senior Application Engineer supporting the Aerospace and Defense Industry.  Rachel spent her pre-MathWorks days at the Office of Naval Intelligence where she used MATLAB and Simulink for missile analysis and simulation. She has also taught high school math, physics, and engineering.  Rachel holds a B.S.E. in Aerospace Engineering from Princeton University, an M.S. in Aerospace Engineering from the University of Maryland, and an M.A.T. in Mathematics Education from Tufts University.

Sudheer Nuggehalli is a Technical Consultant focused on Artificial Intelligence and Data Analytics at MathWorks. Previously, he was an Application Engineer supporting the Government and Defense Industry with technical focuses in machine learning / deep learning, software and application development, and predictive maintenance. Sudheer holds a B.S.E. and M.S.E. in Electrical Engineering with a focus in control systems from the University of Michigan.