Using a Building Energy Management System for Energy Control
Commercial buildings account for nearly half of total energy consumption in the U.S. at a cost of ...
How do you solve complex building issues like energy efficiency without constant human intervention? How can you better understand building energy consumption patterns and gain the insight you need to address inefficiencies? Machine learning may be the answer.
Machine learning (ML) analytics are gaining ground in the commercial buildings industry due to its cutting-edge ability to uncover patterns, produce accurate predictions, and automatically respond to those predictions. According to the U.S. Department of Energy, “as much as 30% of building energy consumption can be eliminated through more effective use of existing controls, and deployment of advanced controls,” which can be greatly enhanced by using machine learning.
Integration of an intelligent analytics platform with advanced machine learning capabilities can help you identify energy saving opportunities that would otherwise not be apparent, while minimizing reliance on human intervention to identify and resolve operational problems.
Using machine learning to improve building energy efficiency begins with data analytics. ML algorithms constantly incorporate and analyze data from a variety of sources (such as equipment, sensors, and devices) to refine an internal model that can be mined for trends and identify anomalies. Over time, it allows an analytics system to learn how a building operates and detect excess energy usage, identify operating expenditure (OPEX) savings opportunities, and recommend solutions to complex issues.
As the analytics system develops a deeper understanding of the building, advanced algorithms can perform increasingly complex learning processes and use data patterns to automatically adjust operational setpoints, initiate actions, and make changes to building systems and device behavior. And the more data obtained about a system or piece of equipment, the better your system becomes.
In a complex network of building equipment and automation systems, machine learning is invaluable for improving operations. In particular, it can be a powerful tool for reducing energy use and improving overall energy management.
Four key ways machine learning can optimize building energy efficiency are:
One of the major applications of ML is energy consumption forecasting. ML analytics can use a building’s historical energy consumption data to reveal trends and predict future energy use. When actual energy consumption is higher than predicted, it could point to inefficiencies.
A building’s interconnected network of equipment, sensors, and devices can generate an unwieldy volume of data and trigger multiple alarms when equipment malfunctions. Advanced analytics organizes, analyzes, and prioritizes this data to produce meaningful insights and isolate points of vulnerability and failure. Significantly, ML can go beyond traditional fault detection and preemptively alert you to system and equipment failures before they occur by flagging early and related deviations from historic trends. This can be critical for avoiding catastrophic failures, preventing energy waste, and minimizing downtime.
Predictive analysis is particularly valuable for high-consumption targets like HVAC equipment, which accounts for approximately 40% of commercial facility energy consumption in the United States. Even small inefficiencies can add up. Traditional models and rule-based strategies typically do not provide opportunities for preemptive intervention, potentially resulting in significant energy waste before faults are identified.
Occupant needs and optimal building conditions change throughout the year, opening the door for seasonal inefficiencies. Seasonality modelling involves correlating set points to seasonal conditions to account for these changes. Continuous analysis with machine learning capabilities can ensure efficiency over time without manual intervention.
Machine learning can create a thermal model of your building based on historical HVAC and temperature sensor data, taking into account weather and occupancy. With this model, automated action controls can be set and automatically adjusted to preemptively heat or cool your building according to predicted near-term conditions. For example, facilities can be automatically pre-cooled in advance of an anticipated heatwave or pre-heated based on upcoming occupancy changes.
The longer an ML-driven analytics system is gathering data from a set of equipment, the more historical data it has on how, when, and why that equipment operates, allowing it to make more accurate predictions and better adjustments. However, the value of machine learning can become evident in just a few weeks after the analytics system is deployed.
By deploying an analytics platform like onPoint Analytics, you can harness the power of machine learning to achieve your energy efficiency goals. onPoint is an intelligent analytics platform that combines machine learning with deep domain knowledge to provide meaningful insights and can create pragmatic efficiency strategies for day-to-day building operations. With cutting-edge analytics, onPoint can automatically streamline operations and reduce energy use across your facilities.
Jon Schoenfeld, PE is Buildings IOT's Vice President of Energy & Building Technology. He's been developing advanced algorithms for building automation applications for more than a decade and he applies his tremendous building expertise as he oversees the team of building scientists creating the onPoint platform.