Customer application 18 January 2022
Using Artificial Intelligence to Reduce Heating and Air-Conditioning Costs

Automation specialists mrm² have developed a system that employs machine learning to make heating and cooling office and industrial buildings significantly more efficient, and reduces energy costs by up to 20 %. WAGO provides the PFC200 Controller and Edge Computer, two key components in the system.

It all started a few years ago, when Marc Gruber, co-founder and Managing Director of mrm², wanted to build a house for his family. Gruber’s company, which includes around 30 employees and is headquarters in Bad Ditzenbach-Gosbach in Swabia, specializes in intelligent automation technology. So obviously, building technology from a pre-digital era was unlikely meet his standards for use in his own house. “I took a good look around at all of the products available to automatically control heating systems – and was completely underwhelmed. Everything on offer was simplistic, without a shred of intelligence,” he recalls. Only one possibility remained: “If there’s no intelligent solution on the market, we’ll just have to build one ourselves!”

No sooner said than done! The experts at mrm² developed a system to keep each room at exactly the desired temperature when occupied. The key to this is artificial intelligence (AI): Integrated machine learning methods predict when and for how long individual rooms will be used. Using this knowledge, the heating elements or heating surfaces can be controlled so to provide comfortable temperatures precisely when needed with the lowest possible energy consumption. This prevents the heating system from warming rooms up unnecessarily and continuing to run. In the summer, the system keeps the climate control systems operating efficiently on an as-needed basis. Movement sensors provide the necessary data for predicting occupancy. In offices, these are often already installed anyway, in order to regulate the lighting.

“The Optimal Interplay among the Building Systems”

“Thanks to the AI, our solution is able to take information about changes in the usage of individual rooms and very quickly incorporate it into the control process,” explains Gruber. The system can also factor weather predictions into the calculations. For example, if sunshine is expected in one room at a certain time in the winter, then the supply temperature for the heating can be set somewhat lower than on overcast days.

The positions of the blinds are also targeted to increase energy efficiency in the building. This is most important in the spring and fall, when rooms need to be heated or cooled according to outdoor temperatures and sunshine. According to Gruber, this is unique on the market. “Other solutions are isolated applications, which control only the heating or only the blinds. In contrast, we bring it all together,” he explains. And he is quick to emphasize that “WAGO’s technology allows us integrate heterogeneous building automation systems in the best possible way!”

Intelligent Control for Heat Generators

In order to exploit the full potential of the AI, the application program uses the integrated intelligence to control the heat generators itself. To do this, the system uses calculations that have been made of the heat demand of the individual rooms. In larger buildings, this technology is often complex and includes combined heat and power systems, gas-fired boilers, heat pumps and solar thermal systems. “Hybrid systems in particular often lack efficiency. That’s not surprising, since controlling these systems is no trivial matter. Our system ensures that the heat generators function at maximum energy efficiency,” says Gruber.

Overall, the system can reduce energy costs by 15 to 20 percent, as demonstrated by practical tests in reference buildings. The only other way to achieve this order of magnitude is with expensive modifications like facade insulation. In many cases, the system will pay for itself within one or two years. The mrm² head doesn’t mince words: “We reduce costs by eliminating humans as a source of errors.” No human being can control building technology as precisely and proactively as is possible using AI.

WAGO’s technology allows us integrate heterogeneous building automation systems in the best possible way!

Marc Gruber | Managing Director of mrm²

PFC200 Offers Maximum Flexibility

To achieve building control with intelligent, secure communication for interaction among different systems, mrm² relies on WAGO for key components. This company from the Swabian Jura region has many years of experience with WAGO: “We were using WAGO products when our company only had three or four employees. There are several projects that we never could have completed without WAGO,” recalls Gruber.

Besides power supplies and displays, mrm²’s newest solution uses the PFC200 Controller and the Edge Computer. “The PFC200 is very powerful and also offers a high degree of flexibility. The different input and output cards let us use both wired and wireless communication without any problems,” explains Johannes Siegle, who oversaw development of the new system at mrm². For the installations in the reference buildings, the company used KNX modules from WAGO, since an existing bus system was already installed there.

The Edge Computer is Docker^®-Capable

The PFC200 collects the data and transfers it to a database on the local Edge Computer. The machine learning model, which uses the data to predict room occupancy, is also installed there. Alternatively, it would also be possible to offload the database and AI to the cloud. The predictions are returned to the PFC200. An algorithm on the controller then derives the necessary control signals from this information, supplemented by weather predictions and current meteorological data. “Simplifying somewhat: The Edge Computer provides the intelligence, and the PFC200 organizes the communication,” says Siegle.

The database and machine learning models run in software containers on the Edge Computer. “Docker®-capable hardware is important to us, because that’s the only way for us to implement the AI that we are picturing. The predictions are also quite CPU-intensive,” explains Siegle. The user-friendliness is another plus: “As you know, automation engineers like us are more uses to PLCs. But we were still able to get used to the Edge Computer very quickly.”

First Office and Industrial Buildings, Then Residential Houses

The team from mrm² named the new system “WICKIE M” after clever little Viking from a cartoon series, whose quick mind finds solutions in even the most complicated situations. Following intensive testing in two reference buildings, the system is now for sale. The company initially has their eye on designers and operators of office and industrial buildings. “As currently configured, WICKIE M makes the most sense for buildings with more than about twenty heated/air-conditioned rooms,” says Managing Director Gruber. In the medium term, mrm² plans to further develop the system to help private households also save on heating costs – a plug-and-play WICKIE M system could then achieve greater energy efficiency in single- and multi-family housing units.

Text: Ralph Diermann – Energy Journalist

Photo: Ian Siepmann/Tremonia

Product Overview

Item no.: 752-9800

Edge Computer ; 4 x ETHERNET, 4 x USB, HDMI, DP ; 16GB RAM ; 256GB FLASH

{ "id": "752-9800", "name": "752-9800", "nameEn": "752-9800", "listPosition": "1", "listType": "product-slider" }

Item no.: 752-9400

Edge Computer ; 2 x ETHERNET, 4 x USB, HDMI, DP ; 4GB RAM ; 64GB FLASH

{ "id": "752-9400", "name": "752-9400", "nameEn": "752-9400", "listPosition": "2", "listType": "product-slider" }

Item no.: 752-9401

Edge Computer ; 2 x ETHERNET, 4 x USB, HDMI, DP ; 8GB RAM ; 64GB FLASH

{ "id": "752-9401", "name": "752-9401", "nameEn": "752-9401", "listPosition": "3", "listType": "product-slider" }

Item no.: 752-8303/8000-002

Edge Controller ; 2 x ETHERNET, 2 x USB, 1 x USB-C, HDMI, CAN, DI/DO, RS-232/485, ...

{ "id": "752-8303/8000-002", "name": "752-8303/8000-002", "nameEn": "752-8303/8000-002", "listPosition": "4", "listType": "product-slider" }

Item no.: 750-8216

Controller PFC200 ; 2nd Generation ; 2 x ETHERNET, RS-232/-485, CAN, CANopen, ...

{ "id": "750-8216", "name": "750-8216", "nameEn": "750-8216", "listPosition": "5", "listType": "product-slider" }

Item no.: 750-8216/025-000

Controller PFC200 ; 2nd Generation ; 2 x ETHERNET, RS-232/-485, CAN, CANopen, ...

{ "id": "750-8216/025-000", "name": "750-8216/025-000", "nameEn": "750-8216/025-000", "listPosition": "6", "listType": "product-slider" }

WAGO at Work

More Building Technology Applications

Customer Applications: Building Technology

From conventional renovations to smart buildings: Discover more exciting building projects and learn how our customers have made their buildings more efficient, comfortable and future-proof with WAGO solutions.

Customer application 18 January 2022
Using Artificial Intelligence to Reduce Heating and Air-Conditioning Costs