By Rami Khushaba, Sr. Data Scientist, Ph.D., M.Sc., B.Sc.
April 4, 2017
If someone had asked me a year ago if I saw a future for myself in the energy management space, I most likely wouldn’t have taken the question seriously. The focus of my career has, until recently, been within the medical field. The work that I was conducting when earning my Ph.D. was centered around neural signal processing for artificial arms. During my post doctorate work, I ventured into neural marketing — examining eye movement and monitoring what neurons are firing when a subject watches an advertisement. Eventually, I found a position with a well-known company in sleep apnea devices that utilizes signal processing and machine learning. After nearly five years focused on sleep apnea, I was eager for a new challenge. Almost by chance, I came across BuildingIQ and the more I learned about the methodology that the company uses for energy management, the more I realized how many parallels could be drawn to sleep apnea technology and my past experiences with signal processing.
At a high-level BuildingIQ’s approach to managing energy is the same approach that is used for diagnosing sleep apnea —a disorder in which a person’s breathing is being interrupted while they sleep, which can lead to a lack of oxygen to the brain and the rest of the body. Depending on the type of test being conducted, sleep apnea may require the monitoring of a patient’s breathing patterns, heart rate, blood pressure, blood oxygen levels, brain activity, limb movements, eye movements, snoring levels, and chest movements. This is similar to how an energy management system monitors buildings. In each instance, machine learning and advanced modelling techniques are used to create a model that shows what is happening and predicts if the event will happen again.
For another example, take the comparison between a human breathing and the oscillating power consumption within a building. Taking air into the lungs can be compared to peaks in the consumption of power and exhaling is equivalent to the time of low power consumption. If you look at this graphed side-by-side it is quite clear. The only difference is that a severe reduction in power might indicate a change in power schedule or occupancy pattern, while in terms of breathing, it is quite problematic for a person to have reduced air supply to the lungs. The same way unexpected fast breathing might indicate that something is wrong with a human, abnormally large power drawn by a building may indicate a problem as well.
When monitoring energy management within a building we look at what is happening, why is it happening, and what should be done? —just like with a patient. To monitor the “breathing” of a building, variables such as oxygen levels and heart rate, are replaced with air pressure, temperature, weather conditions, occupant patterns, etc. Through BuildingIQ’s platform a building’s vitals are used as data streams to feed and an advanced thermal model in near real-time. This model is used to visualize energy usage, identify anomalies in consumption and, with the application of predictive analytics, forecast future energy usage.
What is truly exciting about BuildingIQ’s technology is the real-time aspect. Within the medical device field, I always saw data after-the-fact —much like traditional building energy management solutions. With BuildingIQ’s focus on data capture and analysis, advanced modeling, and machine learning, it’s been easy to feel right at home.
Dr. Kushaba is one of the newest additions to the BuildingIQ team, based out of the Sydney office. His unique experience and fresh perspective will be key to the future development of BuildingIQ’s platform. Although he has jumped into his new position feet first, he’s been kind enough to share his insights on what intrigued him to join BuildingIQ.