PEO – Intelligent Decision at Each Timepoint – BuildingIQ


Living, Breathing Buildings: Applying Medical Expertise to Building Energy Management
April 4, 2017
Intelligent Energy Management 101: Benefits for Educational Campuses
April 25, 2017
Show all

PEO – Intelligent Decision at Each Timepoint

BuildingIQ is at the forefront of applying artificial intelligence/machine learning/optimization algorithms to HVAC systems. Our patented Predictive Energy Optimization (PEO) is the crown jewel among all intelligent algorithms. I’d like to take this opportunity to demystify our approach a bit more. Essentially, the algorithm makes an intelligent decision at each point in time to conserve energy and maintain room comfort. Imagine this simple case — at each point in time, the algorithm can choose between three options:

  • Option 1 — Maintain the same zone temperature. This option is what most BMS can do —i.e. a fixed zone temperature setpoint.
  • Option 2 — Increase the zone temperature. On a summer day, compared to option 1, this translates to immediate energy savings at the current timepoint.
  • Option 3 — Decrease the zone temperature. On a summer day, compared to option 1, this uses more energy at the current timepoint in the hope of saving energy in the long run.

Intelligent Decision at Each Timepoint

The ability to choose between these three options allows us to optimize the HVAC system in the following ways:

  1. Set the room temperature to follow the natural response of the building. In a mild climate, this minimizes the time chillers and boilers are operating.
  2. Cool and store this cooling energy in the building’s construction material and air, when it’s cheaper to cool, and then release this cooling energy later when needed. We are essentially treating the entire building as a thermal battery!
  3. Maximize room comfort by constantly adjusting the room temperature.

In reality, PEO can choose to increase the room temperature a little or a lot. This is not an easy decision to make as it requires an automated understanding of the thermal properties of the building. We are continuously improving this decision process to allow PEO to make better choices at each timepoint, using some of the following parameters:

  • Better verification on whether the algorithm has made the right choice at each point in time.
  • More control output —supply air pressure/temperature, battery charging/discharging, chiller on/off and staging.
  • Incorporation of comfort models that are fitted to a specific building room and its occupants.


Rui (Ray) Xu is Data Scientist at BuildingIQ. During his career, he has encountered many challenges in the transfer of human knowledge into machines, the interpretation of results of algorithms into human intuition, and the verification of evolving strategies. This experience has helped him to solve some of the most difficult and interesting problems in the industry. He describes himself as a bit quiet but also very energetic and assertive when there is a mathematical, simulated and/or experimental proof behind the scenes.