A few weeks ago, March 2022, I saw a callout on Reddit for sign-ups to the NYSERDA (New York State Energy Research & Development Authority) RTEM Hackathon. As I'm writing this I'm in the middle of a career transition going from photographer to programmer. I love data, and I couldn't pass up the prospect of having access to a bunch of time-series data for a variety of buildings all over New York.
My undergrad was in Architecture so I come with a lot of passion for the built environment already! I'm using this Hackathon as an opportunity to work on a real-life project and develop my skills in data analytics and programming.
In 2016, NYSERDA launched the Real Time Energy Management Incentive Program throughout NY State, empowering building owners to collect vital data about the health and functioning of their buildings. They have provided data from over 200 buildings with an enormous collection of data points.
Ultimately, their goal is to further the attractiveness and success of their RTEM program. They are seeking submissions that propose novel use cases and push the boundaries of what we can learn and achieve using data from the built environment.
We seek submissions that propose novel use cases for the provided data set. Be creative. Ideal submissions are those use cases that positively impact or accelerate the electrification and decarbonization of New York State commercial and multi-family buildings. — RTEM Hackathon
- Use Case Description & Demonstration (10 Points)
- This score rates the novelty of your use case.
- Your submission must describe how the RTEM database is instrumental in your use case and how the growth of this database (>700 buildings) may impact your use case.
- In addition, your use case must identify the principal actors, activities, interactions, minimum data schema and necessary API functionalities associated with your use case.
- Finally, your submission must contain an illustration of your proposed use case using the provided API and existing data records.
- Market Adoption & Impact (10 Points)
- This score rates the impact of your use case if it were integrated into the current RTEM Program.
- A key focus of NYSERDA is the accelerated adoption of RTEM Systems in New York State, without dependence on significant incentives/cost-share by New York State’s investor-owned utilities.
- As an example, use cases that demonstrate a more significant impact in growing the RTEM System Market and/or increase the confidence of building owners and operators to adopt an RTEM System will receive a higher score.
- The judges value technically interesting use cases; however, a high market impact is essential to receiving a high score in this category.
- Decarbonization through Electrification (10 Points)
- This score rates your use case on its ability to promote building electrification and/or lowering a building’s carbon emission intensity.
- Decarbonization through electrification is the process of replacing fossil fuels with electricity.
- Electrification examples include switching from fossil fuel-fired sources of comfort and hot water heating (e.g. natural gas boilers), to in-building air-sourced heat pump heating, geothermal heat exchangers, or through a shared thermal network for a campus or neighborhood.
- As an example, a use case that demonstrates an acceleration of winter heating and domestic hot water in New York State would receive a high score.
- Ease of Implementation (5 Points)
- This score rates the ease of implementing your use case, if NYSERDA were to adopt your submission.
- Submissions that utilize readily available skillsets and commonly available technologies provided by generally obtainable hosted environments will receive high scores.
- Additionally, any novel partnerships that the participant proposes to implement their submission may be favorably scored.
- Conversely, use cases, which require proprietary technology or skillsets will receive a lower score.
- Use of Open Data Sources (5 Points)
- This score rates the use of open data elements to support your use case.
- Use cases that significantly incorporate the provided RTEM dataset will be ranked higher than others that do not.
- We encourage data fusion, and use cases that leverage open and publicly available datasets.
- If your use case depends on an external dataset, please include in your submission how NYSERDA might easily locate and/or access these new data elements.
- This description will assist us in evaluating how opaque, reliable and accessible these elements are in your submission.
The Hackathon provides access to the RTEM Portal which is presumably the same interface building owners and operators use in production environments. The RTEM Program is delivered through the Onboard Data Platform who work with engineers to provide access and storage of building data for analysis.
The Onboard platform provides an API, and a handy Python SDK with various utilities access and make sense of the data.
I've had a few conversations with people about what exactly I could do with all of this information. These were had before we were given access to the data though so some early ideas were not possible with what we had been given.
The first concept I explored was the idea of a physical or digital dashboard bringing awareness of consumption within the building benchmarked against other similar buildings within the RTEM program.
Using historical data, the dashboard could show how current usage compares to past figures and provide a simple indicator (red light/green light) if the building is doing better or worse than usual.
More extreme versions of the idea split the building into floors, or even units and having LED lights on your front door or letterbox. This would probably be very effective in providing public shame to particularly wasteful residents but I don't think it would be a popular program...
However, on an anonymised building-wide level, an indicator might nudge tenants to be just a little bit better in their day-to-day life.
The next idea was inspired by a video by Technology Connections on YouTube about using a building as a thermal battery by shifting heating and cooling load to times where energy is either cheap or green. By adjusting the target temperature by a couple of degrees you could potentially use the entire building as a massive battery, storing or shedding excess heat when it is cheaper and better to do so.
The US Government, and NY State Government have great access to current and historical electricity grid data including energy by source (coal, gas, solar, nuclear etc.) It would be possible to estimate the amount of greenhouse gas emissions as well as calculate cost savings.
Using Concept 2, could we set cool-room thermostats to cycle at a lower temperature while energy is clean and cheap, to reduce usage later on?
Is this even safe/legal? A fail-safe measure could make sure the cool-room never gets above a food-safe temperature.
By far the most boring, but probably most practical idea I've had is to identify opportunities to install power factor correction equipment for buildings that stand to benefit.
The TL;DR of Power Factor is that it's a ratio of the real power used, over the total power (apparent power) needed from the grid to satisfy a given load. The difference between real power and total power is known as reactive power.
Domestic users don't produce much reactive power, so they're not charged for it. But commercial users are, because at larger scales poor power factor becomes a big problem that causes stress and inefficiency in the grid.
Power Factor Correction Units are essentially large capacitor banks that compensate and improve power quality. Reducing reactive power not only reduces the electricity costs incurred by the consumer but also reduces strain on the network as a whole.
The benefit of a Power Factor Correction Unit Program would be trivially demonstrable by using the RTEM Database combined with historical price data. Particularly extreme cases could achieve an ROI within a couple of years.
However, this idea is reactive not proactive (ironically) and the viability of PFC Units does not necessarily require a sophisticated RTEM system.
- Is equipment becoming less efficient over time?
- Increasing lifetime of equipment?
- Thermal losses?
With those early ideas out of the way, let's actually start looking through the data we've got access to, to see if any insights start to appear. At least having some top-down visibility of the data will provide a good starting point.
Related: research/01_exploring_api.ipynb
| Building Types | 0 |
|---|---|
| ('Multifamily',) | 69 |
| ('Commercial Retail',) | 49 |
| ('Commercial Office',) | 46 |
| ('Healthcare',) | 16 |
| ('Food/Beverage',) | 13 |
| ('K-12 School',) | 9 |
| ('College/University',) | 7 |
| ('Hospitality',) | 7 |
| ('Not For Profit',) | 5 |
| ('',) | 3 |
| ('Chemicals',) | 1 |
| ('Manufacturing',) | 1 |