Gaining efficiencies for commercial real estate HVAC systems:
- Reducing HVAC energy consumption by up to 25%.
- Reducing GHG emissions between 20-40%. Improving equipment useful life by up to 50%.
- Improving Occupant Comfort by up to 60%.
Overview of start-up
BrainBox AI is an automated intelligence-powered HVAC brain that makes buildings way more energy efficient. BrainBox AI connects to your existing HVAC system in under three hours. Within two to four months (depending on building size), the AI engine learns your building’s thermal behaviours and maps your system’s point names. It tests all algorithms virtually before deploying them in your building.
Once BrainBox AI knows your building, it feeds external information, such as weather forecasts and utility tariffs, to the AI engine. BrainBox AI’s deep neural networks can now predict the future state of each zone in a building with up to 99.6% accuracy.
Using these predictions, the AI identifies optimisation strategies and autonomously sends commands to individual pieces of HVAC equipment. That’s how it enhances your system to always use the least energy while ensuring the most comfortable temperatures for your occupants. Building efficiency is achieved.
BrainBox AI continues to learn, act, and modify, becoming more efficient every day. As your AI engine collects more data, it becomes better at understanding your building’s unique trends and how each zone reacts to specific changes. This helps increase energy efficiency and savings, decreasing GHG emissions over time as your AI grows smarter.
What makes the start-up innovative
The solution leverages deep learning neural networks to predict a building’s thermal behaviour with 99.6% accuracy based on both existing internal building data and third-party data (weather, occupancy, pollution, utility tariff structures, etc). The AI then uses these predictions to create a curated set of optimization strategies to improve the building’s overall energy consumption. This shift from reactive to pre-emptive HVAC system management is applied individually to each of the building’s distinct zones, allowing for a very granular control of the system. Next generation product focused on Smart Grad Interactivity & AI for Demand-Response (part of Canadian Technology Accelerator in the UK). The first PoC is being developed in the UK.
How the start-up has been designed to scale up quickly
No initial CAPEX investment is required. It requires 1-3 hours to install a connection to the building’s Building Automation System (the BMS). Initial Mapping and Learning period takes ~3 months in order to cleanse the data (part manual / part AI). However this data cleansing and data standardisation (to Haystack Standards) will provide benefits for the building in the long run and future technology integrations.