The global site of the UK's leading magazine for automation, motion engineering and power transmission
19 March, 2024

LinkedIn
Twitter
Twitter link

Now Amazon wants to monitor your industrial equipment

03 December, 2020

Amazon has announced five machine-learning technologies and services designed to help industrial and manufacturing users to embed intelligence into their production processes to improve efficiency, quality control, security, and workplace safety. The services combine machine learning, sensor analysis, and computer vision capabilities to address common technical challenges faced by industry, and are claimed to be the most comprehensive suite of cloud-to-edge industrial machine learning services available.

The five services that Amazon has announced are:
Amazon Monitron, which provides end-to-end machine monitoring using sensors, gateways and machine learning to detect abnormal equipment conditions that may need maintenance;
Amazon Lookout for Equipment, which uses AWS machine-learning models in combination with existing equipment sensors to detect abnormal behaviour and to enable predictive maintenance;
Amazon Lookout for Vision, which applies AWS-trained computer vision models to images and video streams to find anomalies and flaws in products or processes;
• the AWS Panorama Appliance, which will allow users with existing cameras in industrial facilities to use computer vision to improve quality control and workplace safety; and
• the AWS Panorama Software Development Kit (SDK), which allows industrial camera manufacturers to embed computer vision capabilities into new cameras

“Industrial and manufacturing customers are constantly under pressure from their shareholders, customers, governments, and competitors to reduce costs, improve quality, and maintain compliance,” says Swami Sivasubramanian, vice-president of Amazon Machine Learning. “These organisations would like to use the cloud and machine learning to help them automate processes and augment human capabilities across their operations, but building these systems can be error-prone, complex, time-consuming, and expensive.

“We’re excited to bring customers five new machine-learning services, purpose-built for industrial use, that are easy to install, deploy, and get up and running quickly and that connect the cloud to the edge to help deliver the smart factories of the future for our industrial customers,” he adds.

The Monitron device offers end-to-end machine monitoring to detect anomalies and predict when industrial equipment will need maintenance for sites that do not have existing sensor networks. It allows users to cut costs and complexity when building machine-learning-driven predictive maintenance systems, and allows them to focus on their core manufacturing, supply chain, and operations functions.

Amazon Monitron detects when machines are not operating normally, based on fluctuations in vibration or temperature, and tells users when to examine machinery to determine if preventative maintenance is needed. The system includes IoT sensors to capture vibration and temperature data, a gateway to aggregate and transfer data to AWS, and a machine-learning cloud service that can detect abnormal equipment patterns and deliver results “in minutes” with no machine learning or cloud experience required.

Maintenance technicians can start to track machine health within hours, without any development work or specialised training, AWS says. The Monitron devices can be used on a rotating equipment such as motors, pumps, as well as on conveyor belts. They can monitor a few critical machines such as cooling fans or water pumps, or large-scale industrial installations. The service also includes a mobile app for maintenance technicians to monitor equipment behaviour in real time. They
receive alerts of any abnormal equipment conditions from each machine, check up on the health of the machines, and decide if they need to schedule maintenance. Users can enter feedback on the accuracy of the app alerts, and the system will improve over time.

For users that have existing sensors but don’t want to build machine-learning models, Amazon Lookout for Equipment sends their data to AWS to build models for them and return predictions to detect abnormal equipment behaviour. The service analyses the sensor data, assesses normal or healthy patterns, and then builds a model customised for the user’s environment. It uses the machine-learning model to analyse incoming sensor data and identify early warning signs of machine failure.

Amazon Lookout for Vision offers an accurate, low-cost way of using machine learning to process thousands of images an hour to spot defects and anomalies. Users send camera images to the service in batches or in real time, and it identifies anomalies, such as cracks in machine parts, dents in panels, irregular shapes or incorrect product colours.

The Lookout for Vision service can handle variances in camera angle, pose, and lighting arising from changes in working environments. Users can assess machinery or manufactured products by providing as few as 30 images of the baseline “good” state. The service is already available and, starting in 2021, users will be able to run it on AWS Panorama appliances and other AWS Panorama devices in locations where Internet connections are limited or non-existent. https://aws.amazon.com/lookout-for-vision

The Panorama devices allow users to add computer vision to existing cameras. They start by connecting the device to their network. It identifies camera streams and starts interacting with existing industrial cameras automatically. The appliance is integrated with AWS machine-learning and IoT services that can be used to build custom machine-learning models or to analyse video streams. The device extends AWS’ machine-learning capabilities to the edge to help users make predictions locally at sites that don’t have reliable connections.

Each appliance can run computer vision models on multiple camera streams in parallel for applications such as quality control, part identification, and workplace safety. The Panorama appliances work with AWS or third-party pre-trained computer vision models. User-developed computer vision models developed in Amazon SageMaker can also be deployed.

Amazon’s Panorama appliance will allow users with existing cameras in industrial facilities to use computer vision to improve quality control and safety

Finally, the AWS Panorama Software Development Kit (SDK) allows hardware vendors to build cameras that can run computer vision models at the edge. For example, the cameras can detect damaged parts on conveyor belts or spot if machinery is outside of a designated zone. By using the SDK, manufacturers can build cameras with computer vision models that can process higher quality video with better resolution for spotting issues. They can also build more sophisticated models on low-cost devices that can be powered over Ethernet and placed around a site.
https://aws.amazon.com/panorama

Amazon has revealed that BP, Fender Musical Instruments, OSIsoft, GE Healthcare, Siemens Mobility and Adlink Technology are among the first organisations planning to use or offer the new AWS industrial machine-learning services.

BP is working with AWS to build an IoT and cloud platform to improve the efficiency of its operations. “One of the areas we have explored as part of this effort is the use of computer vision to help us further improve security and worker safety,” explains Grant Matthews, BP America’s chief technology officer. “We want to leverage computer vision to automate the entry and exit of trucks to our facility and verify that they have fulfilled the correct order. Additionally, we see possibilities for computer vision to keep our workers safe in a number of ways, from monitoring social distancing, to setting up dynamic exclusion zones, and detecting oil leaks. AWS Panorama offers an innovative approach to delivering all of these solutions on a single hardware platform with an intuitive user experience.”

A Swedish producer of microwaveable pizzas, Dafgards, is using Amazon Lookout for Vision to ensure that its pizzas are covered adequately in cheese and have the correct toppings. “Previously, we installed a machine vision system to detect proper coverage of cheese across a pizza’s surface,” explains Fredrik Dafgård, the company’s head of operational excellence and industrial IoT. “While this system served well for our original inspection requirement, it was unable to detect defects on new product types that include multiple toppings. Amazon Lookout for Vision automates and scales inspection of diverse product types such as a cheese pizza with vegetables. We successfully expanded our quality assurance for new product types with minimal impact to operations.”

The Taiwanese industrial computing and vision specialist Adlink is planning to add AWS Panorama to industrial vision systems. According to Adlink USA’s CEO, Elizabeth Campbell, this “makes for truly plug-and-play computer vision at the edge. In 2021, we will be making AWS Panorama-certified Adlink Neon cameras powered by Nvidia Jetson AGX Xavier available to customers to drive high-quality computer vision powered outcomes much, much faster.”

RS Components is planning to offer Amazon Monitron to its e-commerce customers, and to use it to deliver condition-based monitoring and reliability services via its reliability services business, RS Monition. It will be the company’s first end-to-end wireless vibration and temperature condition-monitoring system. “With the emergence of IoT, we have seen our customers looking to bring real-time condition monitoring capabilities into the factory environment to reduce reactive maintenance and improve asset reliability,” says technical director, Richard Jeffers. He adds that Amazon Monitron will allow users to deploy cost-effective, easy-to-use, continuously improving condition monitoring and to enable predictive maintenance across a wider range of equipment.

Amazon Web Services: Twitter

AdlinkTwitter LinkedIn

RS ComponentsTwitter LinkedIn




Magazine
  • To view a digital copy of the latest issue of Drives & Controls, click here.

    To visit the digital library of past issues, click here

    To subscribe to the magazine, click here

     

Poll

"Do you think that robots create or destroy jobs?"

Newsletter
Newsletter

Events

Most Read Articles