Today’s increased level of automation in manufacturing also requires the automation of material and equipment inspection with as little as possible human intervention. To stay competitive while meeting industry standards, companies strive to achieve both quantity and quality in production without compromising one over the other. However, manual quality inspection of workpieces typically allows only for the analysis of individual samples from a given batch of products. Further, scheduled preventive maintenance of machinery may either lead to unnecessary downtime if performed too early, or, if performed too late, to unexpected equipment failure. To this end, Machine Learning based predictive maintenance has emerged to help determine and monitor conditions of workpieces, machine components and process flows, predict ideal maintenance schedules and recommend actions to take.
Predictive maintenance with mobile IoT sensors and the Cloud - a perfect match
One opportunity that allows companies of all sizes to make use of the recent advances in Machine Learning driven predictive maintenance applications is their integration with mobile IoT sensors powered by Cloud technology. This allows for simple, non-invasive prototyping and experimentations as well as full condition monitoring service deployment at reasonable costs. For example, Amazon Monitron provided by Amazon Web Services (AWS) is one easy to setup Cloud based condition monitoring solution. Monitron mobileIoT sensors are attached to your machines in order to capture vibration and temperature data from machine components, such as bearings, gearboxes, motors or pumps. Collected sensor data is sent to AWS via a Monitron gateway connected to your WiFi network for storage and analysis. To ensure highest level of security and the safety of your data, the data is also encrypted end-to-end.
Source: Amazon Web Services (AWS)
Cost-effective machine component surveillance using Amazon Monitron on AWS infrastructure
The subsequent architectural diagram illustrates a more detailed use case to analyse and visualise incoming sensor data using Amazon Cloud based technologies. Amazon Monitron sensors measure and detect anomalies from your machine components. Both measurement data and as well as Machine Learning based anomaly detections are collected by AWS services (Kinesis Streams and Firehose) and stored in Amazon S3. AWS Glue crawlers analyse the Amazon Monitron data in Amazon S3, create metadata schemata and tables in Athena. Finally, Grafana uses Athena to query the Amazon S3 data, allowing for the creation of user friendly dashboards for a convenient visualisation of both measurement data and machine health status.
Source: Amazon Web Services (AWS)
In summary, with the emergence of vast improvements in the area of Artificial Intelligence as well as Cloud technologies, companies such as yours can begin to benefit from and employ such technologies during the production cycle to automate quality inspection, as well as monitor machine conditions in a cost-effective manner, thereby minimising human intervention, and optimising factory capacities. In the words of chess grandmaster Garry Kasparov who famously lost to IBM’s ‘Deep Blue ’computer in 1997: "Human plus machine means finding a better way to combine better interfaces and better processes."