Planes and pumps, buildings and bridges. Each has unique story to tell, if we have the capability to listen.
That surprising truth was the basis for my keynote at Hannover Messe 2018, the world’s largest trade show for industrial technology. If you didn’t attend, I’d like to share some of the highlights.
It starts with some older concepts but newer execution
First, the digital twin concept isn’t really new, it was introduced by Dr. Michael Grieves in his book “Virtually Perfect: Driving Innovative and Lean Products through Product Lifecycle Management”, but many of the underlying ideas have been around even longer. However, there are three major factors that make digital twin solutions practical now.
- Velocity– the rate at which we can acquire data to inform the twin. The rise of IoT has
made it relatively easy to collect large amounts of data in near real-time.
- Resolution– the level of detail available in the twin. This is a function of how much digital data is now available for physical assets.
- Learning– proven machine learning techniques now allow the twin to learn from the data and refine its model. Learning allow us to calibrate a general model to each individual thing.
With these advances, now, our “things” have much to tell us about their health and history. But we need more than just analytics or visualization of physical components. We need to cross a digital divide. And we need to manage the added dependencies and complexities of software, and its part in a connected product lifecycle. Plus, keep in mind that Digital Twins provide different benefits to designers, makers and operators of things.
The ability to learn from data will be the innovation that makes Digital Twins take off. The use of AI – to augment human skills and intuition – means today’s engineers can more clearly understand the ‘systems-of-systems.’ That allows them to shorten design times, manage larger ecosystems, dynamically re-calibrate, and create better experiences through software-driven devices.
Second, understand the business outcomes
When you combine cognitive sensing, artificial intelligence and Watson Internet of Things, you can build a smarter digital twin. Not only is the technology intriguing, but there are several compelling business reasons why it makes sense, too. With a robust digital twin, you can:
- Leverage operational insights to improve product performance.
- Quickly identify problems, root causes and potential impacts.
- Automate discovery of new knowledge and insights, which can enable new business models and new user experiences, and deliver true differentiation.
Understand all the dimensions of a Digital Twin
Like most technologies, first comes the business value, then comes the actual implementation. As you evaluate your own digital twin opportunities, consider these dimensions:
- Model – an abstraction of the thing or process, its component parts, its behavior and its relationships to other things or processes. A digital twin may be composed of many different models that describe different aspects.
- Interface – how the user/role interacts with the twin. Different users and use cases may require different visualizations or interaction modes. There is no “one” universal interface or visualization for digital twin.
- Analytics – how value is extracted/derived from the twin to create a business outcome. A digital twin commonly incorporates many different data sources. And the volume and complexity of data is beyond human capacity to process.
- Lifecycle – The representation of the digital twin changes depending on where in this lifecycle the particular user of the twin is. Most products start with engineering specifications where the twin model represents a prototype for a whole group of things being built (a model of “the many”). As each thing is built it gets its own individual digital twin model (a model of ”the one”). The collection of data and artifacts created through the lifecycle is called the Digital Thread.
- Granularity – the scope of the digital twin. Twins can be very small, representing a single part or component in a complex system, or very large, representing an aggregate of many parts or even many things (cyber-physical systems)
Take heed of lessons learned from digital twin adoption
There is a maturity model to take heed of with Digital Twin adoption. This key? Bringing your data together and doing something smart with it. To make your digital twin adoption go more smoothly, I have four suggestions. They’ve proven effective for other companies, and they’ll help you achieve a more positive outcome.
- Focus on current business processes. Often, digital transformation requires a new way of working and a shift in organizational norms. It also fosters new working relationships. For example, it’s no longer “engineering versus operations.” Instead, it’s “engineering AND operations.” Start by improving collaboration across all roles, teams and divisions to build stronger teams and to get the most from your technology.
- Add incremental value to address top pain points. Be prepared for some unexpected bumps. Set aside “just in case” resources so that you can address the ones causing the most problems.
- Develop a data Integration strategy. To make the most of any digital transformation, you need to understand the value of your data. It’s not enough to just have the data; you must extract value from it, too. And the best way to get the most value is by combining your IoT data with other existing data sources and tools to give you insights that matter to the business.
- Move to Software-Driven and Agile.We are increasingly living in a world driven by software. Nimble organizations embrace agile because it lets them react and respond more efficiently to market changes, customer expectations and new technology. Software is essential to integrating digital twins in to your business processes, and software enables agility. Understanding and planning for this shifts that digitalization causes is now a critical organizational skill.
So the key things to consider are: Who does it serve? How can we get there incrementally? And what expertise gaps do we have? This will help chart the right course.