Energy Insights From Top to Bottom, and Back Again

How does FDD energy analysis start at the meter?

The first view is the utility meter. Clockworks tracks total energy use, energy use intensity (EUI), cost, and carbon. The number that matters most, though, is how actual use compares with a model.

The model comes from a full year of energy and weather data. It can be a fixed baseline year an organization measures itself against, or a rolling baseline Clockworks builds on its own. Either way, Clockworks normalizes it for weather, usually outdoor temperature and humidity, and can add other factors like big changes in occupancy.

Watch this moment in the webinar — 09:18

The approach has a long history. Alex traced it to a technique “commercialized from Pacific Northwest National Labs called whole building energy diagnostician.”

Why isn’t meter data enough on its own?

Knowing a building used more energy than expected is only the start. Alex described the problem energy information systems have always had:

“That’s great to know that this building or this system is using more energy than it used to. Now what?”

Watch this moment in the webinar — 17:44

Do you order an audit? Send someone to investigate? Alex’s answer is FDD energy analysis: connect the top-down view from the meter with the bottom-up view from the equipment, so the team can see what happened and why it happened.

How does virtual metering work without submeters?

Putting a meter on every piece of equipment costs too much. “No one has all their air handlers metered,” Alex said, “or I shouldn’t say no one, but I don’t believe I’ve ever seen it.”

The sensors are already there, though. If Clockworks knows how big a fan is and how fast it’s running, it can calculate how much power it uses. If it knows the airflow and how much heating or cooling is happening, it can calculate the thermal load. That gives air handlers, pumps, and chillers their own virtual meters.

Watch this moment in the webinar — 14:31

From there, a team can rank its biggest energy users by building or by unit, and track efficiency numbers like kW per CFM for fans and kW per ton for chillers. The Energy Insights dashboards put those views in one place.

Missing documentation doesn’t stop the analysis. When a motor size or rated flow is unknown, Clockworks fills in a small, medium, or large placeholder based on statistics from the more than 25,000 air handlers connected to the platform. “I know the answer is not zero,” Alex said. If an issue turns out to be big, someone can check the nameplate and update the number.

How does FDD separate energy use from energy waste?

Total use and wasted energy are different numbers. Clockworks calculates the waste from real data: utility rates, equipment sizes, airflow, water flow, and temperatures. “This is not estimated,” Jerry said.

A lot of that waste can be fixed without new parts. In the demo, one cooling tower opportunity only needed a change to how the towers were staged. “You don’t have to turn wrenches,” Jerry said. “This is simply an opportunity to reprogram your sequence of operations.”

Clockworks also scores each diagnostic for energy, comfort, and maintenance, based on how severe the fault is, how long it lasts, and how often it happens. Teams can sort straight to what matters to them.

What does FDD find at the equipment level?

Jerry walked through several examples of FDD energy analysis on the demo site:

• Air handler. A preheat coil valve read closed, but hot water was leaking past it and heating the air, which then had to be cooled back down. Clockworks modeled the ideal heating rate for the conditions and put a cost on the difference. “Nothing is black box about these calculations,” Alex said. “It’s all standard engineering.”
• Chiller. Clockworks calculates kW per ton every five minutes and plots it against the manufacturer’s published curve. It also tracks approach temperature against the design value. When the chiller drifted, the report listed possible causes: scaling, corrosion, refrigerant problems, or bad data.
• Chilled water plant. A low delta T, and flow above the expected range starting around 1 a.m. “Unless somebody’s watching the building automation system, they’re not gonna see this,” Jerry said. “It’s likely not gonna have an alarm on this.”
• Ventilation. Clockworks analyzed every VAV box together with its air handler. Early one morning, all of the zones were reheating at once. Raising the air handler’s supply temperature, or adding a reset, lets the reheat valves back off.

Watch this moment in the webinar — 43:55

On large campuses, Alex said, ventilation is “up there with some of the most common significant energy findings,” and low delta T on loops is “also up there with the biggest energy hitters.”

Can FDD be trusted when BMS data is messy?

An attendee asked what happens when points are mislabeled, sensors are unreliable, or the BMS database is incomplete. Alex’s reply: “What you describe here is pretty much every building automation system, isn’t it?”

Watch this moment in the webinar — 38:52

Onboarding usually catches mislabeled points, through simple logic checks. Sensors are a longer story. “It’s not a question of if your sensors will drift and fail,” Alex said. “It’s a question of when.” Clockworks has about 40 to 50 ways to flag a bad sensor, including readings that can’t be physically true, like a mixed air temperature outside the range of return and outdoor air.

When Clockworks flags a sensor, it drops that sensor’s bad data from the graphs and leaves it out of the other diagnostics. That keeps one broken sensor from setting off false faults everywhere else.

What makes Clockworks’ energy numbers different?

Every number in this post comes from the same data foundation. Clockworks maps every point to a standardized model of equipment types, relationships, and engineering units, and runs the data through 50+ expert systems. Jerry described it this way: “we’re taking building automation data, five minute interval data, and we’re running our diagnostics that are all based off of… engineering calculations. This is an expert systems model that we’ve built to do our fault detection and diagnostics.”

Watch this moment in the webinar — 28:35

Alex pointed to the same model when he talked about false positives: “This is where the model based, physics based expert systems approach that Clockworks has is highly differentiated in terms of reducing false positives.” Read more about what actually powers building analytics.

Watch this moment in the webinar — 41:27

How does a fault show up back at the meter?

The leaking valve from the air handler example showed up a second time, on the building’s chilled water BTU meter. Clockworks runs diagnostics on meters too, and it flagged extra chilled water use on the same day. “Really cool to see it from different angles to corroborate exactly what’s going on,” Alex said.

Watch this moment in the webinar — 38:07

That’s the full loop of FDD energy analysis. Start at the meter, follow the number down to the fault, then check it back at the meter.

Watch the full webinar above, or request a demo to go deeper on your own buildings with the Clockworks team.

Frequently asked questions

What is FDD energy analysis?

It’s using fault detection and diagnostics to manage energy all year, instead of in periodic audits. The software compares each building’s energy use with a weather-normalized baseline, then traces any gap to the equipment faults causing it. Every finding comes with a calculated cost.

Do you need a submeter on every piece of equipment?

No. Clockworks calculates electrical, heating, and cooling loads from sensors the building automation system already has, like fan size and speed, airflow, and coil temperatures. That works like a virtual meter on every piece of equipment.

Where does Clockworks get its data?

About 95% or more comes from the building automation system: temperatures, pressures, flows, and set points. A software gateway reads the data behind the customer’s firewall and sends it one way, outbound, to the cloud. Utility meter data adds the whole-building view.

What happens if sensors are faulty or points are mislabeled?

Onboarding usually catches mislabeled points with simple logic checks. Clockworks flags faulty sensors about 40 to 50 different ways and leaves bad data out of graphs and other diagnostics, so one broken sensor doesn’t create false faults elsewhere.