Putting Failure into Context with the Digital Twin

Most assets are built once and maintained over time; but this doesn’t tell an asset’s whole story. Maybe they are built once, but they get upgraded, are retrofitted for a new task, and go through so much change, that they may look completely different than when they first started operating. This is especially true of complex, high value assets such as aircraft, ships, railroad locomotives, and wind turbine’s that all have long asset life. This, coupled with federal regulatory and compliance requirements, means that maintenance providers are scrambling to find ways to make maintenance decision-making more efficient. 

Organizations are constantly looking for ways to improve maintenance effectiveness. For example, by performing work according to a defined plan in order to avoid or mitigate the consequences of failure thereby doing that work in the most cost-effective manner. Often, unscheduled maintenance is one of the key reasons for increased maintenance cost and delays.

Lately, there is a trend on improving reliability and reducing unscheduled maintenance through predictive analysis―anticipating possible failures before they happen. While this has come with varying degrees of success, predicting the unpredictable is now seen by many organizations as the answer to solving longstanding issues with asset failures.

For predictive analysis to gain a seat at the table, and generate the value everyone perceives, it will require a partner―one that creates the individual context of each asset―the Digital Twin.

The Digital Twin for Context

Digital twin, by definition, is an exact replica of a physical asset. This means you need to capture every single asset’s configuration and then manage it as it changes over time. This creates the context required for accurate predictive analysis. The Digital Twin will tell you what it is now, as well as its history, using the Digital Thread to connect relevant information such as computer-aided design (CAD), simulation models, IoT and time series data, and maintenance records. This context and history create a story of each individual asset and its operating condition.

In a perfect world, the Digital Twin is first created when the physical asset is completed along with its manufacturing processes and its serialized information is recorded. If it is not captured there, there are still plenty of other chances to capture and manage the changes of a Digital Twin during the commissioning and operation of an asset. The key is to capture and manage it, this is the context you need for effective predictive analytics implementations.

Digital Models are not Digital Twins

There is an emerging trend to use simulation models, or CAD models created during the engineering phase of the product lifecycle, as the Digital Twin of an asset. The concept is that comparing these digital models with operational data, while running different simulation models, may result in the identification of failures. After all, the simulation models are tested for many types of possible operational scenarios and the related failures that would occur if they persisted. At first look, this is a vast, rich resource of information that can be used by maintenance to monitor the signals for failure in the field.

The problem is that these models may not necessarily reflect the final as-built configuration of the asset that went to the customer. As assets go through manufacturing, much can change. For example, suppliers change, modifications are incorporated, and defects are rectified. The original models will not reflect these differences and the actual performance profile will differ from what was predicted. Fast forward to the asset operating in the field that, over a few years, has undergone maintenance and upgrades to the point that its configuration is now significantly different from that assumed in the original simulations.

Maintaining Digital Twin Context

Promising technologies such as, IoT, predictive analytics, and simulations all have value, but only when used in context of the individual asset. This means building and maintaining a digital record of the configuration of products as they are manufactured, maintained and upgraded. This is the key to keeping the asset in the field and its Digital Twin synchronized.
The Digital Twin is subsequently updated whenever a significant change happens to the asset. For example, if electric motor serial number #001 is replaced with electric motor serial number #002, the corresponding change is made to the Digital Twin configuration. The electric motor, serial number #001 becomes part of the Digital Twin’s history and is forever linked with the Digital Thread.

Now, with a constantly up to date Digital Twin configuration, simulation models can be built to the specific characteristics of a particular asset and coupled with Time Series and IoT data generated from the physical asset to predict potential failures. Context is the partner that predictive maintenance needs, for example, two engines operating for the same number of hours might need different maintenance cycles if one accrues these hours at cruising altitude versus the other performing high-G maneuvers.

Predictive Analysis with Context

Predictive maintenance is the ability to determine when maintenance should be performed based on the actual conditions of the operating asset in the field. Predictive Maintenance being an end result, cannot be looked in isolation and is dependent on sending real-time data about the performance of the physical asset in the field, using IoT sensors, to the Digital Twin configuration of the physical asset, analyzing this data against OEM specifications, applying multi-physics simulation models to the Digital Twin, interpreting the results to predict component failures, and scheduling maintenance proactively.

Technology advancements in sensors, analytics, and simulation can be the solution to predicting maintenance problems, but only when a Digital Twin configuration approach is considered. An individual asset’s context must be used as a baseline to predict failure with data generated from IoT sensors and validated with purpose-built simulation models. The key is to develop corresponding business processes to support asset configuration changes, the ability to collect IoT data specific to individual configurations, and the use of the power of simulation to build a digital model of the specific asset configuration. Predictive maintenance using the above framework will yield results that could be of significant value to organizations maintenance effectiveness.

Anonymous