IoT 4 mins read

The energy-flexible factory of the future

With the advent of renewable energy, industrial companies have to deliver energy-efficient production while contributing to energy grid stability. How?

Gabriele Strobel Gabriele Strobel

With renewable energy sources for power generation gaining ground, German utilities have to ensure that they can still generate and deliver reliable, high quality power to customers.

Germany is undergoing a transformation in its energy sector. The number of traditional large-scale power plants is on the decline, while, at the same time, the percentage of regenerative energy suppliers and decentralized cogeneration power plants is growing.

On any given day, power companies need to ensure they have enough generating capacity (operating reserves), that enough power gets to where it is used (reactive power), and that there are no faults in the delivery of that power to the customer. Only then can they guarantee the stable operation of electricity grids and deliver high supply reliability and quality with regard to voltage and frequency.

With the advent of renewable energy, industrial companies must be able to deliver energy-efficient production while contributing to energy grid stability. So, how does using a high percentage of renewable energy sources impact this goal?

The PHI Factory project was designed to provide the answer. In December 2016, the German Federal Ministry for Economic Affairs and Energy launched the project. The central idea was to design a digitalized and energy-flexible factory of the future, and to demonstrate that a factory can actively self-regulate its energy usage/costs, while also providing system services for the public power grid.

The research consortium comprised of a coalition of research institutes in the mechanical engineering, mechatronics, and electrical engineering fields – along with selected industry partners that participated in both development activities and as users of the research results. The solutions the consortium developed were integrated and tested in the ETA research factory at the Lichtwiese campus at Darmstadt University of Technology—and met with great success. The ETA factory was expanded into fully digitalized, energy-flexible model factory.

A lean, real-time platform

Because the comprehensive data being captured in the context of flexible power grid management is so vast, PHI Factory needed capabilities in new big data processes.

This is where Software AG came in. We developed one of the core innovations of the project: a big data platform connected to the data sources that delivers real-time data analysis and control. External software systems like power monitoring, forecasting models and production planning are integrated in the platform via suitable interfaces and calculated control variables are transmitted to the machine and system controls.

Software AG’s platform, based on its Apama streaming analytics, meets state-of-the-art requirements for efficiency, scalability, security and ease of maintenance without generating high costs.

Here is a use case

Let’s take a look at one of various use cases: Load forecasts predict future energy needs. A forecasting model developed by one of Darmstadt University of Technology’s project partners was integrated via the platform in Apama. The platform was supplied with data every second to predict energy demand with a short-term horizon of the next 100 seconds. The load forecast was sent directly to the EnEffCo® energy efficiency management system from consortium partner ÖKOTEC (energy consumption consultants) and then visualized in a dashboard.

Key innovations in the model factory

After three and a half years of research, the results from the PHI Factory are impressive:  With the help of full monitoring of energy and process data, future energy flows can be anticipated, and the power load profile adjusted subject to weather and market data.

The project implemented methods from machine learning for industrial energy systems, among other things, and developed an artificial intelligence that independently learns optimized operational behavior. In addition, the consortium deployed and tested a high-efficiency hybrid energy storage system comprised of flywheel storage and lithium ion batteries. The energy storage system integrates the load strategies for factory-associated e-mobility. This enables the factory to operate autonomously for up to two hours with a small supply from renewable power sources. Up to 100% of the factory’s electricity load can be made flexible this way and postponed by several hours. This way the grid supply can be shifted to less expensive times with a high availability of renewable energy.

These findings of the project will be evaluated in a follow-up project, “KI4ETA” (Artificial Intelligence for Energy Technologies and Applications in Production) and researched further. Software AG is participating in this follow-up project and will contribute its expertise as well as its Cumulocity IoT Cloud, Edge, and Apama technologies.