What are Digital Twins & How do They Shake Up Enterprise Maintenance?

The era of digital transformation is upon us. With it comes a flood of maintenance-related information, surging in from various databases and other digital forms.

This information is highly valuable; it includes data about the condition of machines, tools, equipment, parts and many other assets, which can serve as a basis for constructing accurate digital models (i.e. “virtual representations”) of physical assets in the cyber world.

The models – called “Digital Twins” – are increasingly used to simulate and analyze industrial processes. Digital Twins can be seen as a natural extension of conventional models for the digital simulation of maintenance processes based on information derived from the physical world using sensors and Internet-of-Things (IoT) devices. Nevertheless, they can be also seen as a disruptor of digital simulation for enterprise maintenance based on their ability to fuse IoT information with events from digital simulation.

How are Digital Twins Used in Enterprise Maintenance?

Understanding the disruptive power of Digital Twins for maintenance applications begins with understanding their principal use cases in enterprise maintenance:

Use #1: Digital Simulation

A Digital Twin provides the information needed to execute realistic and accurate simulations about the behavior of assets and their maintenance. Simulations based on Digital Twins take in information about risk factors, failure modes, operating scenarios and system configurations, in order to produce maintenance-related KPIs (Key Performance Indicators) such as:

• Maintenance cost
• System downtime and unavailability
• Remaining Useful Life (RUL)
• End of Life (EoL)
• Mean Time Between Failures (MTBF)
• and more.

Such simulations enable the anticipation of future maintenance activities, which is key for developing predictive maintenance systems. At the same time, they can improve the Planning of preventive and condition-based maintenance processes as a means of minimizing downtimes and unscheduled repairs.

Use #2: Support for What-if Analysis

Digital Twins are also used for what-if analysis of alternative maintenance scenarios. By simulating different maintenance scenarios, organizations can evaluate and select the most effective asset management strategy. What-if analysis can be exploited both for long-term planning of maintenance strategies (e.g., comparing a predictive maintenance strategy to a preventive one in terms of return on investment) and for short-term on-the-spot decision-making (e.g., whether it’s time to replace a tool or not).

Use #3: Maintenance System Configuration

In many cases, Digital Twins remain synchronized to the status of the physical assets they represent. Whenever the status of an asset is changed, the Digital Twin model updates to reflect the change. Likewise, whenever the status of the Digital Twin is changed as part of an IT operation, the respective change is reflected in the physical assets based on some IoT or Cyber-Physical System (CPS) that conveys the status of the digital world to the real world.

Based on this synchronization, Digital Twins can be used to configure the operation of assets and related physical systems. For example, if an IoT/CPS application detects a machine’s failure or degradation pattern, it could configure the machine to operate at a reduced speed. This can be done through an IT command to the Digital Twin of the machine, rather than a human operator of the machine.

It’s important to note that the use of Digital Twins for the flexible configuration of maintenance systems hinges on the deployment of proper CPS systems in the plant floor; notably, systems that can configure their physical parts based on information and commands from cyber counterparts.

Use #4: Open the Doors to Innovation in Enterprise Maintenance

In the medium- and long-term, we expect Digital Twins to drive open innovation in enterprise maintenance based on digital technologies. In particular, Digital Twins will be used as a vehicle for testing, validating and evaluating innovative ideas about when and how to maintain or repair an asset, without disrupting plant operations. This will facilitate innovators in their endeavors and will reduce the enterprise maintenance innovation cycles. In this context, IBM’s views Digital Twins as a way of transforming engines and other pieces of equipment to digital innovation platforms.

The Value of Digital Twins for Enterprise Maintenance

As you can see, these use cases provide clear benefits for plant operators and integrators of enterprise maintenance solutions. First, they give them the opportunity to gain insights on the production and asset management processes, such as non-obvious failure or degradation patterns for assets.

Based on such insights and knowledge, plant operators and solution integrators can better plan their maintenance strategies.

Second, Digital Twin simulations facilitate optimal decisions based on the evaluation of alternative maintenance scenarios. This leads to improved Overall Equipment Efficiency (OEE) and ultimately to a better Return on Investment (ROI) for the maintenance solutions.

Third, Digital Twins can be used to increase the automation and cost-effectiveness of the maintenance processes, through increased flexibility in the configuration of maintenance systems and the physical assets that they comprise.

Finally, Digital Twins can greatly facilitate the transition from traditional forms of maintenance (i.e. reactive maintenance and preventive maintenance) to emerging and more effective forms, such as predictive maintenance (PdM), with minimal disruption to shop floor operations.

Digital Twin Design Challenges

Despite the ongoing digitalization of assets and maintenance processes, the construction and use of Digital Twins in maintenance applications remains a challenge.

The main issue concerns the design and construction of proper models for the assets and systems involved. It’s particularly complex to develop such models, as it requires an understanding of many different aspects of the maintenance processes, including:

• Asset physical properties: The design of a Digital Twin should reflect the properties of the physical asset, including its electrical and mechanical properties as specified by the vendor. Therefore, Digital Twin designers should consider engaging the equipment vendor in the construction process.
• Failure modes and criticality: A Digital Twin design must include information about the failure models of the assets, as well as about the failure modes of their parts. Moreover, it should also comprise information about the criticality of each mode. Therefore, it should account for the plant’s FMECA (Failure Mode Effects and Criticality Analysis) knowledge.
• Degradation patterns and their risks: Digital Twins should also capture and manage information about the asset’s possible degradation patterns, but also information about their likelihood and impact. Hence, the Digital Twin design should include risk-related information for each one of the possible failure patterns of each asset.
• Statistical information: Failure modes and degradation patterns will occur based on varying probabilities, which are likely to follow different distributions. Hence, a Digital Twin should typically comprise statistical models (e.g., distribution functions) and mathematical concepts that will drive the simulation and what-if analysis processes.
• Information about maintenance and business goals: Any digital simulations of assets and processes are typically driven by the strategic goals of a company regarding enterprise maintenance, including targets and KPIs associated with maintenance costs, inventory levels for spare parts, OEE targets, risks and business impacts of unscheduled maintenance, downtime implications and more. Thus, the digital modelling process should reflect the business priorities of the enterprise that will be deploying and using the Digital Twin(s).

It’s no question: it will be a struggle to get the best out of a Digital Twin deployment. For some, this argues well for a phased approach, which begins with simple models of assets and their failure modes, and gradually expands to more sophisticated ones. Such an approach can allow plant operators, equipment vendors, and solution integrators time to gain confidence in their Digital Twin deployments. It can also allow them to minimize relevant development and deployment risks.

In the wake of digital disruption, maintenance professionals everywhere are reflecting on their current enterprise maintenance systems and processes and analyzing the challenges – and possibilities – of embracing new innovations. Are you already collecting and leveraging digital information about your assets and field service processes? If yes, you can start considering the development of a Digital Twin model, along with related applications. This can help you take the right decisions at the right time, not to mention the potential to increase efficiency and yield sky-high ROI.