Artificial Intelligence in Energy Management Software
This is not a blog about AI, machine learning or even deep learning.
It is about a quick review on how far energy management software (EMS) has developed over the last years. And if we can see first signs that applications are now embedded in EMS which some would refer to as based on or using AI or machine learning.
Basic level EMS
There seems to be broad consensus about the basics of EMS: the software should be able to
- Collect data automatically via different interfaces and from different applications
- Process data quickly and provide it as a basis for decisions and actions
- Be integrated in exisiting systems and processes using existing bus systems, protocols and standards
- Support the rules of the overall energy management system via automated actions
- Ideally also allow for forcasting
- Send alarm signals for certain pre-defined parameters
And yes, there is also broad consensus that such a software is a prerequisite for long-term success of energy management. Although excel sheets are still used in industry. But this is another story.
Use cases: The link between energy management and IIoT
IIoT (industrial internet of things) is often referred to as the driver for the next industrial revolution. A number of use cases are defined to showcase the benefits of IIoT. Some of them are rather easy to understand, think of "predictive maintenance", some are more difficult to think about and to develop, e.g. "new business models". And energy efficiency is also mentioned as one of the obvious (and easy to understand) use cases.
But whatever IIoT use case you think about, they have one thing in common. We are always talking about big data. And big data is the foundation for any AI, machine learning etc application.
So where are big data based AI or machine learning applications when looking at energy management software?
Or to re-phrase the question: what AI/machine learning applications would be most obvious or would deliver highest benefit?
We are no AI experts, but...
We spoke with various different vendors of energy management software. And we found only one we believe touched on it: IngSoft and the "automatic pattern recognition"
According to IngSoft, this new functionality can detect irregular energy consumption data automatically, allows to identify causes immediately and to react upon. We have not spoken with IngSoft about how they do it and how deep the software integration needs to be to allow for such pattern recognition. So clearly we are also interested in feedback from companies using this new application.
In parallel, I discussed with IBM on an event in Brussels about "Energy Efficiency: the AI factor" industrial uses cases for the IBM Watson AI platform. Although impressive examples were shown, these were mostly around the management of technical information rather than focussing on energy management. But we stay in touch with IBM as the promised us to refer us to more use cases. Potentially also ones with a loink to energy management.
Another initiative we came across is the German Project KiPro, funded by the German Federal Ministry of Economics and Technology. KiPro stands for "Künstliche Intelligenz in der Produktion" (in ENG that would mean "AI in Production"). The objective of KiPro is to develop AI supported platforms to assist production to improve energy efficiency. It is based on EMS collecting the data but then applying AI applications to support "humans" in their decision making. Well, automatic pattern recognition sounds like one possible use case, right? But there are likely more.
Outlook: AI & energy management software
This is a call to open the debate. Tell us what you think will be the next big "AI EMS thing". Or just direct us to intersting companies or examples as we are always keen to present them to a broad EEIP audience.
For any question, comment or to provide us with further information. please contact me under juergen.ritzek[at]ee-ip.org
- IoT is Today!, by Levis Gandeu
- Can data analytics replace people in energy efficiency?, by Steve Baab
- Event: Energy Efficiency: AI Factor, IBM, Brussels, 19.09.2017
- IngSoft: innovation - automatic pattern recognition
- KiPro: Künstliche Intelligenz in der Produktion
About Jürgen Ritzek
Juergen Ritzek is co-founder and Business Director of EEIP. Juergen is responsible for strategy, marketing and business development of EEIP and drives the growth of EEIP towards an energy transition platform. Juergen leads EEIPs B2B communication and relations and ensures EEIP relevance for value chain challenges (inter-company) and internal decision-making processes (intra-company). Following an international career at Unilever he founded European network consultancy GBC (2009) and EEIP (2011).