Predix - Foundation for the Industrial Internet


New tools mean new rules rules for optimising assets and processes

Finding the ultimate competitive advantage in your existing business or creating a new business model takes different thinking and different tools. In Predix, GE’s Industrial IoT platform, you’ll find a combination of modern IT technologies developed with a deep understanding of how to safely and securely integrate with edge automation and systems. This blend opens up new ways to take advantage of data both within critical assets and processes, and also along the supply chain and service organisations around them.

Fully-built solutions, or a core platform to build on – you choose

Key considerations when unlocking value from an industrial operation’s data are volume, portability and usability. Predix was specifically designed to cope with the flow of high-volume, high fidelity data from industrial processes, and to direct that flow to a platform that would make it easy to add context and structure to supplement the raw streams. The environments modelled around asset and process data in turn make it easier for many stakeholders to extract value – whether a data scientist working with analytics tools, an operations leader using a Predix application – or even another system consuming data that Predix has packaged and structured. Predix now underpins a number of full application sets from GE – Asset Performance Management is one – but is also at the heart of a market of 3rd parties who develop tools, services and apps that run on the platform. This allows customers to define their own systems journey using the mix of COTS and custom development that is most able to support the outcomes they’re pursuing.





Digital Twins

What is a Digital Twin?

It's simplest to describe a digital twin as the data-based representation of a physical item. But when you tie that concept to an industrial IoT platform, you can take advantage of some powerful extensions of that simple idea:

  1. Digital twins are not just for data scientists. The most common way many firms start to use digital twins is to simplify the work of visualisation, reporting and integration of asset or process data. Digital twins give you a place where you can define all the attributes that should be available to “see” or use; the power of the Predix platform is to then manage the way those universal definitions link to the many real-world objects with their variety of local controls, systems and data structures
  2. Twins are born from many sources or combinations of sources. Your CAD or PLM system; your maintenance system; even your financial system; these may all hold data that should form part of the definition of a twin for an asset or a category of assets
  3. Twins become the way you link a unique item to groups or categories to which it belongs – empowering analytics that seek truths that apply across families, not just to single assets
  4. Twins are not isolated – an asset twin may be best understood in relation to twins that reflect the building or environment around it, or the materials that go through it. A platform like Predix allows you to model all the aspects that may be relevant to your twins in ways that help you find the insights that deliver value.

Digital Twin White Paper

Analytics & Machine Learning

Applying advanced tools and techniques to the data

The most exciting opportunities to recover value are often the most difficult to capture. Problems are often multivariate in nature – solving a reliability or quality issue often means analysing a combination of reference data (what are the characteristics of an asset and/or materials), process data and outcome data (test results, reliability statistics, etc). As the number of variables goes higher, the need for tools to sift through the data becomes more pronounced.

Predix offers a number of paths for customers to follow when applying tools to the data they've collected:

  1. Open Source/Custom: As an open platform, Predix easily exposes its data to the analytics tools of your choice. For the customer with data scientists on board, and where the finest “fit” is critical, this may be an important option.
  2. Predix MarketPlace: GE and third parties are adding to a growing library of well defined, prebuilt analytic tools that can be easily integrated into your Predix footprint. See the catalog for the current listing
  3. Embedded in GE applications: In the Asset Performance Management suite, GE has over 150 asset models with associated analytic tools developed. With GE’s history in power and utility operations, there is a particularly rich set of models available for power generation and Balance-of-Plant assets. 

Distributed Intelligence

Putting intelligence where it’s needed

There can be good reasons to bring data to a central platform – whether cloud based, or in your own data centre. It allows people to perform analysis on large data sets that are needed to resolve very fine distinctions leading to significant impacts on asset or process performance. But where it’s not practical or cost effective to constantly feed a central platform, you still want a means to make use of the insights you’ve worked so hard to extract.

Our approach allows for a distributed deployment of models, processing power and analytics. By architecting Predix to run in a mixed of local and central computing resources, GE enables you to easily take advantage of the digital twins you develop – wherever it’s most logical to deploy.



Predix Datasheet



IoT ready by 2020 whitepaper 

  • Digitalising ahead of 2030
  • Putting the AI into maintenance
  • The new approach to SCADA
  • Making sense of industrial data.