Product companies are investing in digital transformation (DX) efforts. An executable digital twin is one such effort. Discussing digital twins can be tricky, as there are several similar definitions. For the purposes of this post, we will define an executable digital twin as a simulation model that mimics the performance of an operating product using sensor data.

Benefits of a  Digital Twin Implementation

Companies are eager to implement digital versions of their field-operating products. That’s not surprising, as digital twins offer many tangible benefits when they reflect the real-world behavior of the products they mirror. This virtual representation is synchronized with the physical product at regular intervals, making the model more accurate.

By tracking the product’s behavior, the digital twin provides deeper insights into the product while facilitating future product behavior and performance predictions. Companies can schedule predictive maintenance, thereby increasing the life of their products. They can also use the operating data to inform the design of future products. This post discusses the steps involved in creating and deploying a digital twin for a product using ANSYS Twin Builder.

Concept of Virtual Sensors 

ANSYS provides simulation solutions in numerous application areas. So it comes as no surprise that simulation is central to the company’s digital twin solution. But to fully appreciate how ANSYS creates a digital twin, we need to understand virtual sensors. 

Let’s start by examining physical sensors. These sensors collect data about a product’s performance and transmit it to the digital twin at regular intervals. This makes the digital twin more effective. However, sensor use in products is limited for multiple reasons. First and foremost, sensors add to the cost of the product. A product also has limits on the number and type of sensors that it can incorporate. And finally,  some products operate in conditions where sensors cannot work—like in the case of a gas turbine engine, where the temperatures are very high. 

In these situations, companies can benefit from digital twin implementation governed by a simulation mode. Simulations utilize virtual sensors to improve digital twin implementation and provide engineers with an unlimited number of sensors. These virtual sensors can operate anywhere inside the digital model—without limitations. Virtual sensors can measure any quantity of interest, acting as extensions of the physical sensors and making the digital twin very effective. 

Steps to Build and Deploy a Digital Twin

Simulation models are key for digital twin implementation. Both physical and virtual sensors play a role in developing accurate simulation models for digital twins. ANSYS Twin Builder has capabilities beyond simulation models, however. Let’s examine how users deploy a digital twin using Twin Builder. 

First, engineers make a mathematical-simulation model. Engineers create the model using components from many sources, such as validated analyses, library systems simulations, and anything else that can connect using a functional mock-up interface/functional mock-up unit (FMI/FMU). This is the most crucial step—and also the most difficult. Engineers have to write software code, manually connect the pieces, and debug the implementation during this process. Twin Builder’s low-code interface conveniently allows engineers with little to no programming background to create the digital twin.

In the next step, Twin Builder compiles the model into a reduced-order model (ROM) executable. This executable is essentially software code that can be deployed in an internet of things (IoT) platform.

In the final step, the user deploys the executable to the cloud, the edge, or some combination of the two. The ROM executable works closely with many popular IoT stacks, such as AWS, Microsoft Azure, PTC, and more. As engineers improve their simulation model, they can remotely update the software and thereby model on the IoT platform.

Combining the Power of Data and Simulation 

The data from the IoT platform serves as rich input to improve the simulation model.  At the same time, virtual sensors in the simulation model can improve the data model. ANSYS Twin Builder’s latest version combines the power of data (from the IoT platform) with the power of physics (from the simulation model) to create a hybrid digital twin. 

Hybrid digital twins allow users to feed sensor data right into the simulation model. At the same time, they can harness the simulation results to train an artificial intelligence/machine-learning algorithm. This hybrid approach improves the accuracy of the digital twin and closely mimics the product’s real-world operations. In specific customer use cases, this hybrid approach has achieved over 98% accuracy. Furthermore, this approach can be used to verify the accuracy of digital twins that have already been deployed.

Companies often implement a digital twin using a data model, only to realize that they don’t have good data to train their data models. The lack of good data limits the capabilities of digital twins. In these situations, companies can include virtual sensor data to improve their digital twin.

Many product companies have a strong simulation-driven product development culture. They also have simulation models that accurately predict and characterize the operation of their products.  ANSYS Twin Builder can benefit these companies as well.  ANSYS Twin Builder allows companies to derive more value from their simulation investment by helping them deploy a hybrid digital twin. 

All of these capabilities help companies improve the accuracy of their models and derive better insights from their digital twin deployments. 

Twin Builder Powers Companies to Build Better Products

Digital twins are an effective digital transformation initiative. ANSYS’s Twin Builder is an effective solution for companies looking to develop digital twins for the first time. It’s also helpful for experienced digital twin developers. 

Engineers can use virtual sensors to measure any and all aspects of a product. Twin Builder’s low-code platform makes this easier than ever, enabling engineers with little or no coding experience to develop digital twin models.

Companies may have to update digital twin models on the fly. With Twin Builder, users are not bound to the office. Engineers can remotely update their digital twins on the IoT platform of their choice. Twin Builder also allows companies to use both virtual sensor data simulation and data from IoT platforms to create hybrid digital twins. This hybrid approach improves digital twin accuracy, offering new product insights. 

Digital twins enable companies to improve product performance and design better products in the future.