Select your location
Austria

Austria

Czech Republic

Czech Republic

Germany

Germany

Italy

Italy

The Netherlands

The Netherlands

Romania

Romania

How to optimize the production process with a digital twin

Use case for how to become a data company,

Manufacturing scrap is expensive. With a digital twin, it’s possible to use sensor and setpoint values to predict and automatically optimize production quality.


Manufacturing processes by default come with a certain amount of waste. End products need to answer to a certain quality standard specified in the SLA, so everything that is subpar is wasteful. And if there’s one thing manufacturing companies want to minimize, it’s scrap. With a digital twin, it’s possible to predict the quality outcome and significantly increase the Overall Equipment Effectiveness (OEE).

Increase Overall Equipment Effectiveness in manufacturing

Operational efficiency
In any production environment, Overall Equipment Effectiveness (OEE) is always top priority. Delays in production, machine errors and suboptimal settings are time consuming and expensive. Human intervention – such as an engineer applying a trial-and-error approach to achieve optimal quality – is often very inefficient. When a production process is not optimized, it creates scrap, which can be costly even in small percentages.

Production complexity
Manufacturing is an incredibly complex process, with thousands of parameters and factors influencing the quality of the end result. Huge quantities of production data are available, but it is impossible for humans to unify and process this data at any speed at all. Leveraging this data to improve the production process requires a data-driven approach.

Leveraging huge quantities of production data to improve the production process requires a data-driven approach

Quality optimization with a digital twin

Model production conditions
A digital twin is an AI simulation that simulates the production process in a virtual environment. In the digital twin, you can adjust setpoints of the production process and predict what the outcome will be. Consider, for example, identifying the ideal temperature in which to prepare a tortilla. With a manual trial-and-error approach, you would generate huge amounts of waste and lose time and money in the process. With a digital twin, you can change the virtual temperature settings and predict what the effect of the change will be in your production process.

Create a digital copy of your production line
To build a digital twin, it’s first necessary to collect all sensor and setpoint values (every setting an engineer could adjust across the production process). Each of these settings has an impact on the end product, but finding the ideal balance is a complex undertaking. However, with a digital twin, you can undertake this process in a fraction of the time, and without investing any raw material. If the virtual outcome is suboptimal, simply tweak production conditions in the digital twin to define the ideal setpoints for higher quality.

Calculate the ROI
To justify building a digital twin, it is crucial to first calculate the scrap cost. If the cost of the digital twin is lower than the potential scrap cost, the model offers higher ROI and is well worth the investment.

Less scrap, greater cost efficiency

Less scrap
For many manufacturing companies, even a small percentage of scrap is highly cost inefficient. With a digital twin, it’s possible to avoid trial and error and gut-based decision-making while reducing the volume of scrap to an absolute minimum.

Real-time quality assurance
Processing leads to a delay between the adjustment of a parameter and the quality of the end product in a real-life production environment. When manually adapting setting, it can take up to hours to see the end result. With a digital twin you can model the outcome in realtime.

Automated settings
Since the digital twin is a prediction model, it can use available sensors and setpoint values to automate quality control. With the right technology, setpoints can be automatically adjusted in response to modeled quality outcomes and real-life observations, relieving engineers from production line monitoring so they can focus on added value tasks.

Overall Equipment Effectiveness (OEE)
Real-time and automated quality control not only take repetitive manual tasks from line operators and engineers, but they also provide data about the status and effectiveness of production machines. If a machine is too error prone, it can be replaced to increase operational efficiency even further.

Building a cloud data platform for future projects

Additional sensors
To increase efficiency of the digital twin, additional sensors can be used to measure the quality of the end product. Using the tortilla example, cameras can be added to the production line to observe the diameter of the tortilla. Through computer vision, any deviation from the diameter specifications is immediately and automatically detected, and will in turn inform the digital twin. This results in the use of richer, more detailed and more reliable data to validate the prediction model.

Predictive maintenance
The data platform used to build the digital twin offers possibilities for future improvements. Sensor data can be used to build a predictive maintenance model that is capable of forecasting when a machine will require maintenance or a part needs replacement.

Cegeka’s expertise in IoT and Azure solutions ensures an end-to-end approach that links business with IT. We can combine optical sensors with a digital twin to further increase Overall Equipment Efficienty (OEE). Even more, based on our extensive data expertise, we are able to allign business, IT, OT and the data scientists.

Microsoft Azure

As a Microsoft Gold Partner with +600 certified experts, Cegeka is the perfect partner for your Microsoft Azure journey.

More on our Microsoft Azure-based solutions and platforms

Any type of data can be collected, centralized, analyzed and transformed via a cloud-based data platform. We apply our proven Cegeka reference architecture to build a data platform that meets the needs of your enterprise.

Cloud Data Platform with Cegeka Reference Architecture

New call-to-action

Challenges

  • Increase Overall Equipment Effectiveness (OEE)
  • Minimize scrap
  • Unify and process production data
  • Implement a data-driven approach to optimize production

Approach

  • Collect and unify all sensor and setpoint values
  • Calculate scrap cost
  • Implement a digital twin of the production environment

Solutions


Results

  • Less scrap
  • Real-time quality assurance
  • Automated quality control settings
  • Increased Overall Equipment Effectiveness (OEE)

Perhaps we’ve inspired you to work out your own use case?

  • Kristel Demotte

    Kristel Demotte
    Global VP Data Solutions

Then we will discover together how we can help you to become a data company.