The evolution of technology is taking artificial intelligence (AI) to the fore in nearly every industry. As AI gradually becomes mature, it is being applied in the energy management sector. A number of Internet of Things (IoT) companies are using AI to help businesses reduce energy consumption and expenses.
U.S.-based BuildingIQ is one of these companies that aim to improve energy efficiency in large, complex building structures. BuildingIQ’s Predictive Energy Optimization (PEO) service uses cloud-based software to calculate heating, ventilation and air conditioning (HVAC) related utility expenses.
The company uses a sophisticated AI engine that learns a building’s thermos characters and energy use patterns in order to establish the optimum operation mode. The objective is strike a balance between user comfort, utility costs and energy efficiency, the company says.
BuildingIQ will collect various data relating to weather forecast, occupancy levels in the building, energy price, taxation and user feedback, before carrying out thousands of simulations to come up with the most efficient HVAC operating strategy in the next 24 hours.
The optimization is conducted in three steps. First, the system will learn how temperatures in a building change in response to a variety of conditions – weather, humidity and user occupancy. Then the system will conduct its own predictions of how temperatures change in the building. It takes four to six weeks to master the prediction. “Accurate prediction of the building’s thermal behavior, in aggregate and zone by zone, under changing conditions is the precursor to optimization,” BuildingIQ says.
The third and final optimization step is about adjusting the HVAC to ensure tenant comfort while keeping the cost down. “The building might be cooled several degrees below midlevel comfort during the morning in order to let the temperature drift upward, while backing off on power demand during peak pricing on a hot afternoon,” says the company. Temperature can and does vary in a given space by one, two, perhaps even three degrees without noticeable discomfort. Power prices in the marketplace, which rise and fall during the day and by season, are also factored in.
BuildingIQ’s software is connected to the building management system (BMS), a computer-based control system that controls and monitors a building’s mechanical and electrical equipment. Since the BMS controls the air handling unit (AHU), temperature settings can be fine-tuned directly throughout the day.
Verdigris Pinpoint Problems at Device Level
Verdigris Technologies is another energy management provider serving hotels and corporate offices, among other business clients. Its system has been deployed around the world, including the U.S., the European Union, Brazil, Egypt, China and India.
Verdigris’ smart sensors clamp onto electrical circuits to track energy consumption of individual devices. The raw data is collected in real time and then beamed to the cloud where Verdigris AI algorithms take over.
According to the company, the sensors take substantially more data points than a typical smart meter. It tracks energy output down to the microsecond. “The best way to improve plug-load is to start quantifying and tracking their usage,” said Verdigris co-founder Mark Chung.
More importantly, this enables Verdigris’ AI and algorithm to “learn” a building’s equipment over time, and to identify normal and abnormal energy behavior. As such, facilities managers can isolate devices and to address poor-performing equipment, wasteful behavior and even malfunctioning or failing equipment by studying anomalies in their power signatures.
Verdigris even sends alerts when demand forecasts are expected to exceed typical thresholds or peak-demand levels.
Besides reading a building’s energy usage in real time, Verdigris’ AI factors in the weather and building occupancy to come up with an estimated power consumption. The company’s analytics then present insight into which circuits are using more or less energy than they should be, what equipment has suffered a failure, what equipment needs maintenance immediately or in the near future, etc.
With the analytics, users can also benchmark energy use with national CBECS standards, identify devices that are consuming excessive power and verify if supposedly energy-efficiency HVAC equipment really live up to their promises claimed by vendors.
All the computation is conducted on the cloud. “Cloud service is important because cloud technology moves much faster than edge. The edge needs to be deployed for very long periods of time with little turn over,” Chung added.
“Our technology is built to support a SaaS model, where the payback is instant,” Chung said. “After the initial learning and optimization period, meaningful recurring savings should occur to offset any costs and provide much needed infrastructure for layering additional technology.”
Data is Key to Insight
Another energy service company that employs AI is FirstFuel Software, which helps utility providers better engage with their business customers – schools, retailers, government and hospitals. The company collects a variety of data to conduct energy analytics in its cloud. Using machine learning and AI, FirstFuel obtains insight on how energy is used in a building and provides energy-saving recommendations.
By combining data science, building science and software, FirstFuel’s SaaS platform derives intelligence from over six million business customer meters. The company looks at data from electricity meters, weather stations, business profile, commercial real estate databases (including building size, type, fuel type, etc.) and figures out which pieces of data give the most insight into what’s happening inside a building.
Energy consumption of various equipment types – HVAC, lighting, motor and electrical outlet – is compared with that in similar buildings. As such, the company can determine a reasonable range of utility consumption.
With visual analytics of energy usage by the hour, FirstFuel can recommend energy-efficient practices to its clients. It will advise how to benchmark energy efficiency, what actions can be taken, and whether energy efficiency has been achieved after suggested actions are taken.
“Data gives us a sense of how people are using the building and how we can optimize energy use,” said Austin Whitman, VP of Energy Markets at FirstFuel. “It’s not just about reducing energy consumption; it’s about changing how we use it, when we use it and where we use it.” With fine-grained data and information about energy use, businesses can be more optimized in their energy use. For example, they may use a little more energy during certain time of the day when the fuel price is lower.
FirstFuel may, for example, find out that a school is using too much energy when students are not in the school, or when the school is not in session; or it may discover that a car dealership is very energy-efficient except for the lights. In this case, a utility company can cross-sell energy-efficient light bulbs.
Another use case is that a restaurant owner may have three premises, so he can start to analyze all three premises separately and start to see which one is underperforming and which ones are doing well and then he can drill down on the one that is underperforming and make needed adjustment, like a temperature control or a lighting replacement project.
Analytics is what brings data into the realm of human knowledge. Analytics is a about bringing massive amount of data that only computers understand and translate it into something that building operators can act on. In other words, AI is using advanced computational and statistical methodologies to translate data into insight that building operators can use.