Optimized HVAC management to benefit climate
By Brainbox AI | Published on March 11, 2022
Problem and objectives
Problem
- Conventional HVAC systems account for an average of 40% of the industry’s greenhouse gases (GHGs), or 5% of total global GHG emissions;
- High costs for energy consumption from all sources (e.g. electricity, natural gas, fuel oil);
- Temperature adjustments made with traditional systems can cause discomfort to people inside the building, as they are not always suited to the occupants’ needs.
Objectives
- Reduce GHG production in the real estate sector;
- Enable simple and efficient HVAC management;
- Maintain the level of comfort of commercial and institutional building users.
Proposed solution
Solution
BrainBox AI is a Montréal-based company specialized in HVAC systems. The AI algorithms that were developed optimize energy savings and reduce costs for organizations with commercial and institutional buildings.
Principle
By collecting internal and external building data, BrainBox AI trained its models to automate HVAC system management. The company’s algorithms enable a 25% reduction in the buildings’ energy consumption and a 60% reduction in their carbon footprint.
User experience
BrainBox AI’s HVAC management solutions automate control of management commands. The model performs updates every five minutes over a six-hour period. Use of this technology results in a significant reduction in the quantity of energy used without compromising comfort.
Role of AI
BrainBox AI’s algorithms use artificial intelligence (AI) to predict internal and external building conditions and optimize HVAC use. For instance, the company uses:
- AutoML;
- Predictive models;
- Analytic visualization platforms.
The AI techniques that are used include:
- Machine learning;
- Linear regression;
- Principal component analysis (PCA);
- Deep learning, including:
- Deep neural networks;
- Convolutional neural networks;
- Graph convolutional networks;
- Recurrent neural networks;
- Encoders and decoders;
- Reinforcement learning;
- Simulation;
- Regression;
- Clustering.
Impacts
Main outcomes
- Reduced HVAC costs for buildings;
- Reduction in greenhouse gas production;
- Optimization of the HVAC management process and saved time.
Generation and dissemination of new knowledge
- Patent pending;
- Collaboration with research institutes and universities.
Conditions for success
Team mobilised
BrainBox AI’s AI team has over 80 employees, including:
- Data analysts;
- Solution architects;
- Data scientists;
- Data engineers;
- Software developers.
Collaborations
BrainBox AI works with a variety of strategic partners. These include:
- SEG Ingenieria;
- ForTerra;
- Green Bee Energy;
- Sunland Cleantech.
BrainBox AI collaborates with several national and international research institutes and academic networks such as:
- Institut de valorisation des données (IVADO);
- National Renewable Energy Laboratory (U.S.);
- McGill University;
- Polytechnique Montréal;
- Monash University (Australia).
Context of the solution
Use cases
Example of an application for a commercial building:
- System automation;
- Using data such as temperature and degree of pollution to adjust heating or air conditioning;
- Maintaining the level of comfort of employees and visitors;
- Encrypted data to ensure their security.
Customers
The algorithms developed by BrainBox AI are aimed at commercial building owners. The target clientele includes:
- Companies;
- Governments;
- Educational institutions.