Key technologies defining smarter factories: digital twinning

As briefly alluded to in the previous blog in this six-part series, a digital twin basically provides a software-generated representation of a physical operation. It can be created in order to emulate a specific asset within a facility (such as an item of machinery) or potentially the entire facility itself. It may also be used to emulate individual products after they have left the factory.

In the integrated systems of sensors, cloud storage and AI algorithms, digital twins can collect, integrate and analyze data, then make predictions (commonly known as inferences). By enabling the visualization of data in a way that it is less abstract and overwhelming, digital twins can help stakeholders collaborate and reach actionable steps.

Use case in product design/testing

Through simulating how a component within the product (or an entire product) operates under varying conditions, digital twins can identify problems that usually only emerge after prototyping has been completed or during final testing. As a result of spotting problems early, digital twins can make initial product development work more effective and minimize the number of test cycles required. This accelerates the pace of product introductions, reduces overall time to market and curbs design costs. 

By way of an example, during the design of the Ghibli line of vehicles, Maserati created a digital twin of the car. The digital twin helped the company solve design issues with fewer wind tunnel tests and test-driving sessions, all the while reducing the number of prototypes needed before product finalization.

In addition, Maserati used the digital twin to plan the production of vehicle components and the automation of the manufacturing process. Lastly, digital twinning helped with optimization of the supply chain. In the end, Maserati was able to reduce the time to market by almost 50%, cutting it to a mere 16 months (which is far shorter than normally expected within the automotive industry, especially at the high end).

As consumers continue to demand customized products, digital twins are also going to play a product redesign role. Previously, incorporating customer input into the manufacturing process often meant time-consuming delays for assembly-line adjustments. However, by running the customization in the digital twin, engineers can understand how it will affect the production process and how they can accommodate such customization without delaying the assembly line and incurring extra costs.

After the product is in the field, the corresponding digital twin is able to help the manufacturer understand how different product configurations will affect its performance. From this, the manufacturer can make better recommendations to the customer on reconfiguration. 

Use case in production efficiency

The data visualization that a digital twin enables has a tremendous impact on how teams communicate and collaborate. Even before production begins, a digital twin can help simulate and predict a product’s manufacturing process using historical and comparative data – so that engineers can identify areas of improvement. During the production process, alerts will reach management immediately, enabling rapid identification and resolution of points of failure.

Stakeholders will better understand how different production processes influence each other, so problematic scenarios can be uncovered and subsequently resolved with minimal delay. As sensors installed on each machine continuously collect data, workers gain a stronger understanding of the relationship between production steps and product quality. For instance, using a digital twin system, the engineering team in a welding facility could analyze the tasks performed by robotic arms and the workflow to identify the optimal robotics path for weld quality and productivity.

This highlights how humans will work alongside robots in the foreseeable future – the robots (or cobots) doing the dangerous and repetitive tasks, while their human counterparts make more nuanced decisions. Lastly, digital twins will be instrumental in training workers, assisting with the determination of optimal parameters and locating where the possible pain points lie.

Use case in floor-plan efficiency

Manufacturers can create a digital twin of the factory floor. By collecting and analyzing data from the floor, the flow and movement of assembly parts and people can be examined, then potential alterations investigated. This method can optimize the number of operators needed, the steps assigned to each operator and the workload executed per station. Quality and volume can be balanced.

Use case in predictive maintenance

While machine breakdown is to some degree inevitable, it is hugely beneficial for manufacturers to be able to predict when such incidents are likely to occur. Predictions of this kind make it possible to schedule maintenance or carry out necessary part replacement beforehand – thus minimizing downtime and keeping repair costs low.

Individualized predictive maintenance is also cheaper, as a manufacturer will not be replacing parts on all machines, which can waste parts that are still in good working condition. Conversely, maintenance engineers will be made aware of components that require repair, avoiding costly downtime. Moreover, an industrial IoT digital twin also offers a hybrid perspective, by combining historical and freshly collected data. This gives managers a more granular view of each machine’s state of upkeep. 

Conclusion

As digital twins become more prevalent, their use cases will continue to proliferate. In addition to product design, production yield, floor-plan efficiency and predictive maintenance, digital twins will see greater use in helping to develop future product cycles.

Consumer insights plus customer behavior data and preferences could be gathered from sensors embedded in the product, then layered with the customers’ demographic and lifestyle data. Such data integration and analytics will help improve current products as well as creating new ones.

What will this blog series cover?

1. Key technologies defining smarter factories – connectivity
2. Key technologies defining smarter factories – sensors
3. Key technologies defining smarter factories – the rise of cobots
4. Key technologies defining smarter factories – digital twinning
5. Key technologies defining smarter factories – AI
6. Key technologies defining smarter factories – data security