Whether at home, in a study center, or in an office, we’re usually aware of how we feel temperature-wise. If the room temperature feels uncomfortable, we try to adjust it by changing the climate control settings. We probably don’t stop to think about whether we’re using renewable energy or what impact our demand for thermal comfort has on the grid. What’s interesting is that a building that’s part of a smart grid can handle this management automatically, anticipating the availability of resources and deciding how to respond to our comfort needs to provide us with a pleasant environment in which to go about our daily activities.
More than just a connected building
The transition of buildings toward digital monitoring through the installation of devices capable of recording various variables is becoming increasingly evident. This equipment provides data that helps interpret the status of different areas and forms the basis for high-level decision-making mechanisms. Sensors thus represent the first layer upon which a building’s intelligence is built, although they alone are not sufficient for the building to actively interact with the network.
To do this, we need to take it one step further: measure, anticipate, decide, and act. The building can monitor what is happening through its sensors, anticipate demand, weather conditions, or the availability of renewable energy, use that information to decide how to manage its energy resources, and, finally, take action on flexible assets such as storage, specific loads, or HVAC systems.
Decisions made at this supervisory level are typically linked to the concepts of efficiency and flexibility: making better use of available resources and deciding when, how much, and how to consume energy. But this ability to adapt truly comes into its own when it also takes into account the needs of those who use the buildings. That is where HVAC becomes one of the main points of intersection between energy management and the user.
In the context of technology development and transfer for the use of energy resources, thermal comfort offers precisely that opportunity: to explore optimization techniques while keeping the needs of the people who occupy the buildings as a guiding principle.
ADCONTROL: Prediction and Supervisory Control for Smart Energy Management
The ADCONTROL project has allowed us to apply this approach to energy management in a real-world setting. In collaboration with GEnergy, CARTIF has worked on developing a solution designed to manage a building’s thermal comfort and energy resources by combining prediction, optimization, and control.
In this context, forecasting involves estimating how energy demand and photovoltaic production will evolve, for which machine learning techniques, among other approaches, have been used. These predictions, along with variables such as environmental conditions and building temperatures, feed into an optimization-based supervisory controller. The challenge lies not only in identifying the most energy-efficient operation but also in balancing various potentially conflicting objectives, such as making better use of available resources and maintaining adequate comfort conditions for occupants. Based on this information, the system determines how to operate the energy resources and translates the result into temperature setpoints that can be implemented by the HVAC systems.
Coming up with a feasible solution is only part of the challenge. For these decisions to actually be implemented in the building, prediction and control services must be integrated with the monitoring and actuation infrastructure. The solution developed is based on a distributed architecture in which local devices collect information and execute commands, while prediction and optimization services run in the cloud.
Implementing a solution like this in a real-world setting also demonstrates that no two buildings are exactly alike and that adaptability is just as important as the algorithms themselves.
From Algorithm to Actual Building: Data, Adaptation, and Learning
ADCONTROL also allowed us to study what it takes to transition a solution from a demonstration environment to a real building. The system, initially designed for deployment in CARTIF’s Energy Living Lab, later had to be adapted for an office building. This change not only required a review of the monitoring and control infrastructure but also meant dealing with new operating conditions, data availability, and energy performance.
During the adaptation process, it was necessary to review the instrumentation requirements and make adjustments to the prediction models and the control algorithm. Among the challenges encountered were, for example, initial difficulties in accessing photovoltaic production data and obtaining disaggregated information on energy consumption.
It might seem that an energy management and thermal comfort service like the one developed at ADCONTROL could be directly applied to other spaces with similar characteristics. However, the transfer of this technology depends on the actual condition of the buildings and the extent of intervention permitted. As service developers, at CARTIF we continue to gain insights that enable us to address environments as diverse as offices, homes, daycare centers, and even historic buildings, where energy and comfort challenges are compounded by the need to preserve the heritage that adds value to our cities.
An opportunity for users, operators, and the energy system
Smarter energy management can add value at various levels. For users, it means maintaining optimal comfort levels without having to deal with the complexity behind the scenes. For managers, communities, and property owners, it means having new tools to make better use of energy resources, integrate renewable energy generation, and automate part of the decision-making process. And, on a larger scale, adaptable buildings can contribute to a more flexible relationship with the grid and a more efficient use of available resources.
For service developers, the opportunity lies in ensuring that these technologies are seamlessly integrated into spaces—in some cases, almost imperceptibly to users—while decisions consistent with their needs and available energy resources are made in the background.
At CARTIF, we continue to explore precisely this path: how to apply these solutions to real buildings, learn from each context, and adapt the technology to the needs of those who use them.
More about intelligent buildings
