The End of Building Energy Modeling Part Three: Micro-Interval Data Delivers
Promise for Quantifying Efficiency Opportunities in Commercial Buildings
A common refrain in college engineering programs is the importance of utilizing accepted standards and developing best practices to avoid making—and repeating—errors.
After completing 50 audits for a Department of Energy-sponsored building asset rating (BAR) program in 2014, I turned to standards and best practices as a way to identify solutions to the modeling problems observed in the resulting final reports.
Determined that modeling had to work, I spent months driving friends and fellow energy geeks crazy, developing matrices and charts to see where the reports deviated from the smart utility meter interval data. My goal was to outline a new set of best practices that would help energy engineers avoid building auditing and modeling mistakes in the future.
Building energy use evaluation, and the software that drives it, needs to be as resilient as self-driving vehicle technology.
One day while watching a TED talk about the self-driving car program at Google I had an epiphany: Energy use in buildings is a continuous stream of variables, not unlike traffic on city streets.
Just as you can end up sitting in traffic on the freeway because a road is closed on the opposite side of the city, building energy variables are all interconnected. They have the capacity to have dramatic effects on each other, in ways that are not always immediately apparent.
I realized that building energy modeling does not work in its current form because it cannot adequately account for the staggering number of energy drivers in real-world scenarios. For example: in one building, the configuration of elevator shafts results in noticeable drafts; in the next, variable work schedules result in continual changes in occupancy; in a third, measurable air leaks in aging HVAC ducts, hidden above the ceiling, result in constant cycling of the cooling equipment.
For energy analysis software to accurately quantify building energy use it must rely on active analysis. That’s the only way to understand how systems actually work as opposed to how they are supposed to work.
Smart Meter Data is the Answer, But What’s the Right Question?
To better understand the best prospects for energy analysis going forward, it’s necessary to first grasp existing smart utility, meter-based analytics capabilities.
Most software platforms for crunching meter data are designed to:
a) Identify building schedules (or too often, the lack of schedules)
b) Account for energy consumption associated with seasonal cooling/heating
c) Determine how the energy use in a set building compares to similar peer facilities.
Most energy analytics software platforms sift through utility meter data that has been sampled every 15 minutes or even every five minutes. They then create trends of typical energy profiles for buildings during set seasons or weather patterns.
Unfortunately, existing platforms struggle with many of the same weaknesses and deficiencies that plagued the modeling-based BAR reports. For instance, it can be helpful to know that a building turns on cooling equipment at 4AM each morning during the summer. However, all too often, building staff starts equipment during the middle of the night, well before buildings are occupied out of necessity.
What really needs to be ferreted out is the underlying mechanical deficiencies that cause spaces in the building to overheat if more optimal schedules are followed. At the end of the day, we need to be able to properly quantifying the savings potential for fixing deficiencies so that upgrades can be implemented and financed at no net cost.
The painful truths about existing meter-based analytics platforms
1. The software is adept at identifying the least efficient 10 buildings in a selection of 100: Too often, those facilities have no capital budgets, are unlikely to have invested in energy efficiency historically and are the least likely to implement upgrades.
2. Most existing energy analytics software struggle to quantify end-use disaggregation: quantifying energy used by things like lighting, air conditioning equipment or building pumps.
The Key Driver Needed to Fund Efficiency Improvements
Realizing implemented energy savings hinges on developing reliable projections for economic/financial analysis of specific upgrades. Until we can accurately predict the payback of upgrading a specific type of florescent light fixtures to LEDs, or how much energy will be saved by installing premium efficiency roof top HVAC equipment versus merely code compliant units, we’re all just flapping in the wind. Simply put, when projects can be financed so that the savings from upgrades pay for the expense, upgrades will happen.
The Next Generation: Smart Meter Interval Data Analytics
The next generation of smart meter interval data analytics shows promise, especially when it comes to properly identifying and quantifying how energy is consumed in commercial buildings.
Micro-interval analytics take a dramatically different approach to analyzing the interval data from smart meters. The software looks for discrete indicators or events in building consumption data that are proving to be reliable metrics of the efficiency of specific pieces of equipment or systems. The consumption might be one day’s worth of data or even the information from as little as a four-hour period. A number of specific examples demonstrate how the analysis works and why the approach is so valuable.
Micro-Interval Strength #1: Quantifying Outdoor Air
Bringing fresh air into buildings is essential to healthy indoor environments. However, for most facilities, ventilating buildings with outdoor air comes at a considerable energy cost.
When the weather is hot, large amounts of energy are needed to cool the air that fan systems push into buildings. Similarly, during winter months, heating the air for ventilation is costly.
Given that heating or cooling outdoor air tends to be one of the largest energy uses in most buildings, accurately quantifying the total volume of air introduced into a building is critically important. In the Building Assets Rating program reports, errors in quantifying outdoor air proved to be one of the largest sources of cascading engineering errors.
Pinpointing the volume of outdoor air being ventilated into buildings is notoriously difficult. Due to engineering peculiarities, it is nearly impossible to estimate airflow through the percent open readings on HVAC dampers (the automatic doors or louvers on the side of buildings that regulate airflow).
Airflow through dampers is not linear. Gauges or readouts indicating 5% open may result in dampers that are introducing 40% or more of the total potential airflow. At the same time, a percent open indicator reading 10% can be introducing 90% of the potential air volume.
Similarly, even the best air balance data is often incorrect. Test measurements listed in balance reports are heavily dependent on how the outdoor weather conditions impact building pressurization (think of buildings where a wind tunnel forms when the front doors are cracked open). Other variables such as how many exhaust fans were running at the time of the test, how clean the HVAC equipment is and even the exact location where airflow measurements were taken are also important.
Outdoor Air Flow via Micro Interval
Three engineering standards enable quantifying building outdoor air in real time:
1. Most HVAC coils used to cool building air operate at temperatures that are close to 50 degree
2. The Psychometric Chart
3. Energy consumption in most buildings responds quickly to dramatic changes in outdoor air humidity
By comparing the energy consumption of buildings based on outdoor humidity it is possible to reliably infer the amount of air being ventilated into buildings. In many climates, rapid weather changes and thunderstorms frequently take place in afternoon hours. By tracking building energy consumption starting during periods of stable temperature with low humidity, followed by stable temperature and high humidity, out door airflow can be calculated.
How It Works
Micro-Interval software searches local weather data for a specific building. This “Indicator Weather” could be an extended period of stable morning outdoor temperatures—say three hours between 75 and 80 degrees. Once Indicator Weather days are identified, the software searches for dramatic deviations in the weather data from the latter part of the identified days. For example; if the stable temperatures continue, but there is a rapid increase in outdoor humidity.
When we know the temperature and humidity level of outdoor air, the Psychometric Chart along with the standard 50 degree HVAC coil temperature make it possible to calculate the amount of energy needed to condition fresh air. Working backwards from the observed increase in interval energy consumption we can calculate the number of units or volume of outdoor air.
Micro-Interval Strength #2: The Demand Importance of Insulation on Flat Roofs
Heat transfer from roofs is an especially important energy efficiency and sustainability consideration. View any heavily developed area of the United States from the air and you’ll see thousands of what is often referred to as “low-slung” buildings. These are office buildings and warehouses with only one or two levels, but large footprints and expansive roof areas.
The total roof area of these buildings is disproportionate to their relatively short wall heights. As a result, the materials used on the roofs of these facilities have outsized impacts on their overall energy efficiency.
This is especially true for air conditioning-related energy consumption. Too often, large roofs act as solar collectors. Inadequate or degraded roof insulation allows energy and heat from sunlight to be absorbed by the roofing materials during daylight hours. Over time, this energy is transferred into the interior spaces of the building through conduction.
When heat from the roof makes its way into occupied spaces, it must be removed with air conditioning during the early-to-late afternoon. That’s means that its impact on building energy consumption is typically greatest when the electric grid is experiencing peak demand and efficiency measures produce the greatest benefit.
Roofing insulation improvements can benefit both individual buildings and, when installed at scale, reduce the peak load on electrical grids—a win-win. Yet two interrelated problems persist: identifying exactly which buildings would benefit most from roofing upgrades and reliably quantifying the economics on a project-by-project basis.
An Innovative Approach to Using Interval Data to Identify Which Buildings Benefit Most from Roof Insulation Upgrades
Used together, utility smart meter data, historical weather files, measured solar data and satellite-based measurement software can deliver unprecedented precision benefit assessments for potential roof insulation projects.
Here’s how:
Calculating baseline energy consumption: A no sun/sun independent consumption baseline is established by recording interval data readings on overcast days when the outdoor temperature and humidity maintains a stable value for the majority of the building’s morning hours.
Calculating Solar Influence: Next, the software searches weather and solar data for periods of continued stable temperatures when clouds have disappeared and the sun is shining on the roof. The ideal scenario is large amounts of unobstructed sun at high angles that directly result in substantial amounts of solar roof heating or solar gain.
Factoring Insulation Values: Especially for buildings with poor roof insulation, the no-sun/sun influence on electricity consumption is immediately apparent. When the solar irradiance data is multiplied by the observed area of the roof (measured via satellite mapping software), the performance of the roofing materials can be fully factored.
Further, calculations allow for real world thermal performance of the roofing materials. The thermal performance data is important for calculating potential savings from seasonal temperature differences independent of the solar irradiance.
In this way, both heating and cooling savings can be fully calculated (efforts are underway to use interval indicators to validate these calculations).
Due to high costs and escalating building codes, it is often difficult to achieve fast paybacks for roof insulation projects. However, as mentioned earlier, roof insulation is a demand reduction champion. There is a strong argument for heavily incentivizing roof insulation upgrade projects based on reliable reduction in peak demand load. Using Micro-Interval data offers promise for quantifying roof-by-roof savings potential and the combined benefit to the electrical grid during peak hours.
Building Energy Analytics 2.0
To date, 11 Micro-Interval indicators have been identified for quantifying the savings potential of standard efficiency measures. The work is extremely challenging. Identifying each new indictor has required at least two highly collaborative, top-level engineers, each of whom must understand equipment design, real world systems operation and have a working knowledge of building controls.
Progress to date has required a daunting number of Thomas Edison lightbulb-type trials and do not conform to many of the software algorithms that were integral to the development of the first-round energy interval analytics.
It is important to point out that an added benefit of the Micro-Interval approach relates to Measurement and Verification (M+V). Revisiting the same indicators used to quantify efficiency measures post upgrade streamlines the process of verifying that efficiency projects are installed correctly and that actual savings are meeting financial projections.
A Modeling Renaissance
Energy modeling has proved to be an important tool for developing and advancing the efficiency industry’s knowledge of how buildings use energy. Reaching critical energy and carbon reduction goals will require the active involvement of the finance industry in building energy improvements. Modeling today does not and will likely not ever offer the level of certainty and predictably that is essential to engaging financial markets to drive project funding to scale.
Just as cruise control is starting to look like a distant relic in the age of self driving cars, it is time for all of us to take deep breath and acknowledge that modeling as we know it is done. It’s time to embrace new, more promising approaches that will enable implementing efficiency improvements at the scale that is so critical to climate stability.
Read Part 1 – Published the week of (2/26/18) - Moving forward when an engineering gold standard falls apart
Read Part 2 – Published the week of (3/5/18) - Why the modeling deficiencies should be no surprise