Predictive maintenance (PdM) techniques are designed to help determine the condition of in-service equipment in order to estimate when maintenance should be performed. Some of the main components that are necessary for implementing predictive maintenance are data collection and pre-processing, fault detection, time to failure prediction, maintenance scheduling and resource optimization. Predictive maintenance has also been considered to be one of the driving forces for improving productivity and one of the ways to achieve „just-in-time“ in manufacturing.
We address four levels for managing an overall successful PdM implementation
- Machine Level: Identify and evaluate the local machine infrastructure as well as the existing raw data.
- Data Collection Level: Structuring raw data and separate it from Machine Level.
- Data Analysis Level: Understand the business needs in its environment for a global PdM implementation.
- Machine Learning Level: Modelling and run the adequate algorithms to get the best results for the business needs.
Manufacturers increasingly collect big data from Internet of Things (IoT) sensors in their factories and products and using different algorithms for the collected data to detect warning signs of expensive failures before they occur. As the data volume is huge, the use of Machine learning Model Based Condition Monitoring for predictive maintenance programs is becoming increasingly popular over time. It allows for the automation of data collection and analysis tasks, providing round the clock condition monitoring and warnings about faults as they develop.
I think we need to address the entire implementation process. So, the customer sees it as an “all-round carefree package”. In our project we developed machine learning model using Microsoft Azure ML to predict the break down of a record where records were created from a machine using sensors.
See the following PREZI Presentation