Turning sensor data into intelligence: OSIsoft on improving the value of data

OSIsoft and Element Analytics are making it easier for data scientists and engineers to access the operational data they need to perform advanced analytics and create predictive models

Rebecca Lambert
Rebecca Lambert
By Rebecca Lambert on 17 July 2015
Turning sensor data into intelligence: OSIsoft on improving the value of data

This article was first published in the Summer 2015 issue of OnWindows

Two megatrends are promising to transform the face of manufacturing as we know it – the internet of things (IoT) and predictive analytics. Imagine the opportunity if there was a way of bringing these two trends together in one environment – giving manufacturers an opportunity to gain an improved understanding of their operations and, at the same time, generate novel insights.

Connecting pervasive sensor-based data with advanced analytics can now accelerate the value of data. Turning data into intelligence, and reactive decision-making into proactive insights. The combination enables operations to predict events before they happen and prevent catastrophic failures without lengthy and expensive testing. That’s exactly what OSIsoft and its partner Element ­Analytics are committed to making a reality.

Element Analytics provides a data analysis layer on top of OSIsoft’s PI System infrastructure to take advantage of decades of historical and real-time data. Today, the PI System is deployed in over 125 countries, at over 17,000 sites capturing and sharing over one billion streams of sensor data based with visualisation and analytics tools for real-time decision support and operational intelligence.

“OSIsoft is the gold standard data infrastructure for the enterprise to capture and connect sensor-based data to systems for ­decision-making and analytical insights,” says David Mount, founder of Element Analytics. “We add the analytics layer on top of the data infrastructure to help manufacturers and industrials identify patterns in their data using a combination of PI AF, Asset Analytics, Event Frames and Azure Machine Learning. They could do this ­manually in PI, but we’re helping to semi-automate this process and standardise the data. Being able to deal with data from multiple sources in one place makes it much easier to use the data in advanced analytics investigations.”

In particular, Element Analytics is working with OSIsoft to provide services within the Azure Marketplace. “OSIsoft has developed a ­template-based one-click solution, making it easy to deploy the PI System on Azure,” explains ­Prabal Acharyya, worldwide director for ­Microsoft and IoT Alliances at OSIsoft. “We’re expanding on that by providing a subscription-based service on Azure Marketplace and helping customers get their data ready for analysis by ­providing services that contextualise and standardise their data. With these tools, and Element’s integration with OSIsoft’s PI System, Coresight, Microsoft PowerBI, Azure IoT Services and ­Azure Machine Learning, customers can easily access predictive analytics with a few clicks.”

Sameer Kalwani, head of product at Element Analytics, explains: “Until now, manufacturers have been used to dealing with data in a much more reactive way. They’ve used it to retrospectively review data and determine what happened and why. But in order to reduce overall downtime, and increase asset lifetime and performance, manufacturers are now finding that they need take a different and more proactive approach.”

At a time when asset reliability is being highlighted as one of the biggest opportunities for manufacturers to reduce their costs, it’s important that they have the means to look at data proactively. “They need to stop issues before they take place and improve performance while reducing process-based risk,” adds Kalwani. “They need the data working for them.”

UpWind Solutions, which services and maintains wind turbines in the US, is using Element Analytics’ software in combination with OSIsoft’s Connected Services enablement to do exactly that, and better serve its customers in the process.

The Connected Services agreement allows UpWind to subscribe and access in real time all of its customers’ PI System data. Element Analytics then takes this data, and normalises it to make it easier for UpWind to build predictive models in Azure Machine Learning and determine when maintenance issues will take place. “Using OSIsoft Connected Services, Element Analytics was able to help UpWind improve operations using PI Coresight and Power BI dashboards with Azure IoT Services to decrease downtime, increase asset lifetime and advance asset performance. And they achieved all of this within just four weeks,” says Acharyya.

According to Kalwani, data is the most valuable asset manufacturers have. “Historical analysis using a customer’s own data helps them understand trends, gain insight into best practices for efficiency improvement and identify patterns that lead to failure. To do this effectively, the data needs to be in a form that is organised and readily integrates with analytical tools like Azure Machine Learning and PowerBI. The Element Analytics tools built on PI System infrastructure reduce data science implementation times significantly and enable customers to see results in weeks, not months or years.

“Real-time and predictive analytics are providing dramatic improvements in efficiency,” he adds. “This is a paradigm shift that’s being driven by IoT, cheaper data and cloud computing, and those who don’t start implementing these sorts of analytical tools will miss out.” At a time when many manufacturers have data pouring in from numerous sources, but don’t necessarily know how to manage it all, OSIsoft and Element Analytics are providing an environment that is both cost effective and easy to implement. The result is a data infrastructure that allows manufacturers to gain historical, ­real-time and predictive insights in one place.

“It seems like the industry is finally at a turning point, going from people mining the data for answers, to data proactively providing insight to operators,” says Kalwani.

Number of views (11141)/Comments (-)

Comments are only visible to subscribers.