It’s a decisive moment for the future of the world’s climate. Every day more and more people are becoming aware of the challenging goals to cut CO2 by at least 55% by 2030, and reach zero carbon emissions in 2050.


It’s a decisive moment for the future of the world’s climate. Every day more and more people are becoming aware of the challenging goals to cut CO2 by at least 55% by 2030, and reach zero carbon emissions in 2050.

The time to act, is now!

In recent years, giant steps have been taken in the evolution of new technologies: IoT and AI have been playing a role in our homes for some time now. While this rapid development might suggest this technology is widely used, the building HVAC regulating sector still has a long way to go, and this includes all types of buildings, tertiary, industrial buildings and residential community housing such as apartments and blocks of flats. Enerbrain’s R&D team collaborates in close contact with European research centres to find timely solutions for a wide variety of challenges in the European community.

The Enerbrain R&D team is working continuously to integrate technologies, on the one hand to maximize energy efficiency, and on the other to find more scalable Plug & Play solutions.
The R&D team is made up of mathematicians, physicists, electronic engineers, energy and IT experts, and more people join the team every year. The aim is to create an innovative team that can provide energy efficient solutions that are easy to install, and have an immediate impact.

The innovation of HVAC systems using AI

Enerbrain research focuses in particular of regulating HVAC systems, because the energy consumption of these systems is still a major factor, especially in the non-residential sector, despite the significant improvements in terms of energy efficiency in recent years.
One of the main challenges when trying to control HVAC systems for optimal energy efficiency is how to manage non-linear control variables such as: occupants’ behaviour and weather conditions.

Therefore, researchers have concentrated their efforts on innovative control strategies for HVAC systems that can maintain indoor thermal comfort while simultaneously reducing energy consumption. These strategies must also be able to manage signals from the electricity network to meet power requirements while improving network reliability and stability.

These opportunities derive from the introduction of ASO (Automated System Optimization) instruments based on artificial intelligence (AI), that can make a significant difference to the energy flexibility of a building. Energy flexibility is an essential feature of intelligent buildings and defines their capacity to adapt and respond to the requirements of the network, weather conditions and user needs.

Intelligent predictions

Traditional controllers like ON/OFF controls or proportional-integrative-derivative (PID) controls, may not go far enough as they cannot predict dynamic changes that affect the performance of energy systems. An ON/OFF control regulates a process operating within certain limits, while a PID needs control parameters to be set before implementation. While the performance of the latter is better than an ON/OFF control, it may still be insufficient, as PID performance drops notably in different operating conditions to the tuning conditions in which the constants that regulate the control were set.


Internet of Things (IoT) is a technological evolution used today to connect an increasing number of devices and appliances of almost any kind.
MQTT is a standard IoT messaging protocol devices use to communicate with the network, and messages can be sent to one or more receivers so applications can be unpaired when necessary.

The evolution of control algorithms

Today, PID controllers are the most commonly used lower level control systems, while Rule Based Controllers (RBC) are considered the standard for optimising the high level control of HVAC systems, but with limited energy savings. This is because RBCs use set strategies rather than ones that are scalable for various weather conditions or building specifications, and they can’t predict changes in the HVAC systems.
In recent years control strategies based on models have been tested to overcome these limits. In particular, the Model Predicted Control (MPC) has become the dominant control strategy for intelligent building control.
However, while MPC proved excellent for improving thermal comfort and reducing energy consumption, in some cases the fact that the control is based on models can represent a problem. One significant problem with MPC is that it depends on how well-suited the model is for optimizing the control strategy, and it can often be difficult to do so in a precise way. Furthermore, it’s difficult to generalise its use for various types of systems and buildings. As a consequence, MPC controllers haven’t been used as might be expected in the building sector, despite the promising results.
Research has focused on the development of model-free controllers that use Machine Learning (ML) and these have shown great potential for improving building performance, in particular when based on Reinforcement Learning (RL). There’s growing interest in control strategies based on RL, and Enerbrain is making its contribution to the research done in this field.

Research projects we’re involved in


Help2Grow-Innovative technologies for cybernetic agricultural:

Help2Grow intends to introduce cybernetic agriculture for the farm of the future (Farm 4.0), proposing new technologies and renewable resources managed by AI algorithms to make farms more energy efficient and grow crops using fewer phytosanitary products.

Period: ongoing, March 2020-March 2022

Financing: PRISM-E, F.E.S.R. tender 2014/2020, Piedmont Regional Authority

Industrial partners: ATM SERVICE S.R.L.

Research institutes: Turin University – AGROINNOVA Centre; Turin Polytechnic – Department of Applied Science and Technology; University of Milano-Bicocca – Department of Materials Science