Industrial Analytics

We optimize your production processes ensuring you are more able to make efficient use of machines, avoid production downtime and increase the quality of your products. To do this we use the most advanced Big Data and AI technology available. We use your company’s machine and production data as a database for our analyses which we can also access and process for you if needed. We also build on your own experts’ knowledge and experience and incorporate it into our analyses.  

This is how we guide you on your tailored path to Industry 4.0 and the “digital factory”. Within the “digital factory” products or parts including their components, processes and data are called “digital twins” and are typically made available on IoT platforms. Cloud-based solutions such as these cannot, however, be used universally: In application scenarios requiring high latency and data protection it is often more advantageous to use decentralized data processing and distributed learning. This is where “Edge Computing” provides computational power where needed. Our software architecture and decentralized learning experts help you develop the most suitable solution for your individual circumstances.

If Machine Learning is to be applied to production, we develop a systematic procedure model by integrating relevant domain and specialist knowledge into an automated form. This is then applied to the data analysis by adapting learning procedures to typical production issues. Our long-term experience creating time series analyses and using technology building blocks for image and audio analytics combined with our intensive decentralized learning research gives us a technological advantage when production-ready solutions need to be developed quickly.

Thanks to our excellent network within the Fraunhofer-Gesellschaft as well as in regional and national centers of excellence we are able, if required, to put together project teams made up of experts from many different sectors.

Portfolio of services

Our services are tailored specifically to applications involving mechanical engineering, manufacturing, chemical science and the pharmaceutical industry. We structure and implement analysis services customized to your individual circumstances which we then adapt to be used with your specific machine and/or production data. We build these systems from scratch or incorporate them in existing infrastructures for scalable Big Data Processing or Edge Computing.

Increase in profits based on prospective analysis

Our prospective analysis of your process data will provide you with information about how to reduce rejects and optimize processes.

Recommendation systems

We use your machine data to train AI systems to use prospective analyses and to extrapolate from that recommendations for machine and plant operators

Wear and tear prognosis in production

With Condition Monitoring and prospective analysis applied to your tool and machine data you can plan maintenance intervals more precisely and avoid production standstills.

Material property prognosis

We can use Big Data analyses and the visualization of results to help you if you have any questions about material property research and development.

Machine as a service

Working in cooperation with you we can develop new digital utilization models which will allow you to hire out your machines and to bill your customers based on their usage.

Production Quality Control

Improve the quality of your products and processes and increase your production yield with our powerful software.

AI based Design of Experiments

With our AI-based DoE methodology, we optimize test planning and execution processes.

 

Industrial Analytics flyer

You can also find a compact overview of our solutions, cooperation opportunities and best practices in our flyer.

Highlights

Case study for machine operators

Root Cause Analysis

On the instruction of a German machine builder we developed a scalable Big Data architecture and integrated it into the customer’s own infrastructure. By using data hosting during the project stage the customer benefited from saving the initial hardware investment costs. The objective was to use the Root Cause Analysis to recognize particular flaws in operating procedures and to generate automated rules describing the root causes of the flaws. The customer currently has a system in place at their premises that produces regular reports detailing the sources of all errors that have been recognized and attributed. A review process makes it possible to check the rules discovered for each machine and to identify and record the sources of error.

Fraunhofer key project

"Machine Learning in production"

The Fraunhofer-Gesellschaft is keen to play its part in Germany and Europe’s technological pre-eminence by creating a systematic procedure model for the application of Machine Learning to production and to develop the required tool suite. In order to do this it is necessary to integrate all relevant domain and specialist knowledge into an automated format within the data analysis, to adapt machine-based learning procedures to recognize typical issues and to make available the necessary tools and process patterns. We developed a software architecture for industrial operations and implemented it as a prototype. Significant interim results can be seen in a review article on Machine Learning in production.

Publication

Machine Learning for Optimization of Production Processes

Weichert, D.; Link, P.; Stoll, A.; Rüping, S.; Ihlenfeldt, S.; Wrobel, S.:
A review of machine learning for the optimization of production processes
The International Journal of Advanced Manufacturing Technology 104 (2019), Nr.5-8, S.1889-1902, Springer-Verlag London Ltd.

We research and develop for