how to evaluate fault detection technology
Hi, my name is Felix Lipov, Lead Software Engineer at Enertiv. Today we’re going to talk about how to evaluate fault detection technology.
There are a lot of options out there and it’s tough sometimes to separate the noise from the good stuff. We’re going to look at the criteria necessary to figure out what might work best in your particular building.
First things first, you need to understand the critical pieces of equipment that make your building run, such as the pumps, motors, boilers, chillers, elevators, etc.
If any one of those pieces of equipment fails, breaks down, or has any issues, it’s going to impact you financially as well as in terms of tenant comfort. We don’t want that.
To solve this, there are a number of solutions out there that focus on anomaly detection. What does that mean? With anomaly detection, data coming into a system suddenly breaks a threshold, i.e. energy has spiked, and you receive an email or other alert.
But what does that really mean? You’ll get a message saying that you’ve broken some threshold that was built by a human. But, it doesn’t tell you the actual problem you’re trying to solve.
This is where you go from anomalies to fault detection.
Fault detection involves a lot more than just breaking a threshold. It’s about having a dynamic understanding of the environment, contextualizing the problem, not just saying that you’ve had a spike in energy use, but the system has identified that your fan belt slipped, motor failed, are experiencing short cycling, or your pumps are not feeding the boiler and you’re going to run out of hot water.
Those are the kinds of things you want to understand to bring the problem back to reality.
Fundamentally, we need a system that is machine enhanced. Now what does that mean?
Imagine you have a feedback loop, where when something goes wrong, you get your best engineers to investigate it. They can look at the data, look at the system, figure out the problem, fix it, and put the problem to bed. Fault detection technology bakes that knowledge into algorithms and this situation is reproduced hundreds, thousands, and millions of times, so next time it happens, we know exactly what’s wrong.
So, what are the criteria you need to consider when evaluating these kinds of solutions?
1. Equipment-level Data
You need very granular data on every critical piece of equipment we mentioned to make sure your building is running 100%.
2. Vertical Focus
You want to make sure that the solution you’re looking at focuses on your industry, whether that’s industrial buildings, commercial office buildings, residential buildings. Because the faults that you can expect will be different.
3. Technology + Humans
As compared to just anomalies that tell you that something is wrong, you need to get that human connection, you need to contextualize it back to reality. That’s very important to understand how to resolve the problem for good.
4. Fault Library
You want to make sure that they understand the kind of issues you have in your building and they have not just one, two, or three, but dozens of types of different scenarios that help you optimize operations and avoid potential issues.
5. Granular Data
Certain types of faults will only come up, not in data readings every 15 minute, but minutely, sometimes even secondly data. This also leads to predictive analytics.
6. Multiple Data Types
You want to make sure that you’re not just capturing power, but also gas, water, environmental conditions, noise, vibration. All of these types of data can help support, and contextualize understanding that will support fault detection to make sure that when something goes wrong, you know exactly what the problem is, how to resolve it, and how to keep your building running at 100%.
I hope this was helpful, thank you very much.
If you would like to demo Enertiv's technology based on this criteria, you can schedule one here.