- Register


Home>IIot & Smart Technology>Big Data>Next level machine learning

Editor's Pick


Next level machine learning

23 February 2023

HOME TO some of the most advanced manufacturing equipment in the world, The Manufacturing Technology Centre (MTC) is taking machine learning to the next stage with a world first Guided Analytics
 demonstrator supported by SolutionsPT, AVEVA and Reekoh. David Baskett talks about the project

THE MTC is a high-quality environment for the development and demonstration of new technologies on an industrial scale, and provides unique opportunities for manufacturers to develop new and innovative processes and technologies. The MTC's mission to reach engineering excellence through technology is what aligns the organisation so well with leading industrial IT solutions partner and distributor of  AVEVA software, SolutionsPT.

David Baskett, technology and management strategy at SolutionsPT, says: “Working with the MTC is a perfect fit for SolutionsPT. Not only because of the opportunity to work with their skilled engineers to create demonstrators that showcase the very latest software capabilities in practice, but because their drive to help the UK manufacturing sector matches so closely to our own vision of digital transformation success for every UK enterprise. I have been involved in multiple projects with the MTC and they always lead to impactful, beneficial results. The Guided Analytics demonstrator is one I am particularly excited to be talking about.”

As part of the AVEVA Insight Cloud Platform for Industrial Operations, Guided Analytics takes machine learning to the next stage of its evolution and brings the technology within reach of a much broader cohort of industrial enterprises.

Machine Learning (ML) is a branch of artificial intelligence (AI) and computer science which focuses on the use of data and algorithms to imitate the way that humans learn, gradually improving accuracy and enabling (or automating) better decision making, powered by big data.

Slow uptake

Industry has an ever-increasing reliance on data and those enterprises not putting this data to work risk becoming uncompetitive. However, uptake of ML in industry has been slow, partly due to its complexity. ML algorithms traditionally require an in-depth understanding of data science which is outside of the skill set most industrial enterprises have at their disposal. It is here that Guided Analytics can help by bringing powerful ML capabilities into the domain of IT and OT operators already working in industrial companies.

It is important to note that Guided Analytics is not a predictive analytics add-on, it is an in-built feature of AVEVA Insight. It is also not to be confused with Automated Analytics, the default Insight training model, which is based on unsupervised machine learning. Guided Analytics is supervised ML, that enables the creation of custom training models where the user builds unique anomaly detection models without the pre-requisite of extensive data-science or coding experience.

Once the on-boarding process is complete, an end user can enable Guided Analytics and begin training their model in just a few clicks through the configuration-based low-code environment. This was one of the many reasons AVEVA Insight was chosen for the MTC demonstrator. Enabled through AVEVA Technology Partner Reekoh, plug and play integration between IT, OT and the business layer means immediate support out of the box without coding expertise needed onsite.

“This out of the box connectivity really sets the system apart and has benefits that reach far beyond the demonstrator. Creating an edge to enterprise connectivity approach, the system will seamlessly connect to the Enterprise Asset Management (EAM) while adhering to all certified security requirements, bringing in data from all IoT devices through the Message Queuing Telemetry Transport (MQTT) no matter how geographically spread those assets are," says Baskett.

“Showcasing the potential of Guided Analytics was the driving force behind the demonstrator model. The technology testbed shows the use of intelligent anomaly detection and its potential for wider industry use on a range of assets and processes. Spotting anomalies means unplanned downtime events can be avoided, operators can be alerted of sudden energy spikes or even the smallest change in a number of variables, and Overall Equipment Effectiveness (OEE) can be improved dramatically," he continues.

As industry adopts a more data-led approach in the era of digital transformation, machine learning will be an integral technology. Guided Analytics is the output of machine learning, an easy-to-use way to build data models, and, more importantly, a straightforward way to get valuable, tailored insight from them. Using a cutting machine as the data source, the demonstrator shows an in-depth understanding of the state of the asset. Leaning towards an intelligent predictive maintenance model, the data describes the health of the machine from several inputs including energy consumption, temperature, and vibrations.

“As is often the case in manufacturing facilities, it is not easy to immediately see the area that is contributing to anomalies or failure conditions. If that was the case, you would simply deploy maintenance there and then. Guided Analytics works for complex production processes. This gives operators the tools they need to assess and maintain every asset on the factory floor, whether there are two machines or two hundred. The implications for improving Overall Operation Effectiveness (OEE) places AVEVA Insight and its Guided Analytics functionality at the forefront of digital transformation technologies.” Says David

Human data load

As industrial leaders know, an abundance of data can be a help or a hinderance; when it helps, it informs efficient operations, but if it is too complex, it can cause operator overload, which undermines its usefulness.

To help ensure that Guided Analytics prevents data overload for operators, cloud graphics are available straight out of the box with no extended coding necessary. This functionality also helps overcome the skills issue often implicated in the take-up of new digital technologies. The demonstrator proves the remarkable ease of use, with configuration of a Guided Analytics model literally taking seconds, and with no time spent on coding or pipeline configuration.

The graphics act as a data filter, ensuring the operator is only seeing what is necessary to complete their job function with the opportunity to drill down for further detail through simple, intuitive dashboards. Previously, this high-level of machine learning would have required a data scientist to create the models along with reams of code, but the simple-to-use demonstrator shows how these models can be applied by operators with even the most basic IT or OT background. The low code integration approaches enables Guided Analytics to deliver insight and show benefits within a few clicks, capabilities previously unheard of in the control system industry.

SolutionsPT and the MTC both agree that most valuable asset of any production facility is the human workforce. But for that human workforce, their most valuable asset isn’t a high-cost piece of equipment, it’s time. Since time cannot be made, only be saved, Guided Analytics offers a valuable insight into the productivity benefits of digital transformation.

Machine learning is vital to the future of industry and Guided Analytics offers a direct pathway to the benefits. Companies don’t need to bring in data scientists or make huge time investments into complicated coding. The simplified model enables access to asset operating conditions whilst reducing the cognitive load on operators.

Guided Analytics shows the potential of machine learning within industrial settings, it also shows that the benefits associated with machine learning are within reach for most companies. There is no need for a data scientist or mountains of code to set algorithms and enable advanced maintenance strategies. Manufacturers can see Guided Analytics in action with the MTC, providing a low-risk environment to explore the benefits, advance digital transformation and be supported by the experts at every stage.

David Baskett is technical strategy manager at SolutionsPT


Key Points

  • Industry has an ever-increasing reliance on data and those enterprises not putting this data to work risk becoming uncompetitive
  • Guided Analytics brings machine learning within reach of a much broader cohort of industrial enterprises
  • Using a cutting machine as the data source, the demonstrator shows an in-depth understanding of the state of the asset