Allow a resolution of 1 minute or 5 minutes. Limit the total export date range of smaller increment export if that helps with performance/capacity issues.
Several people have requested this on the original Export Released Features topic.
18 Likes
I would also like this, maybe even at 1 second interval. With that, I could use Excel spreadsheet macros to look for my 3800 Watt A/C or 318 Watt refrig going away but 3800 Watts or 318 Watt sudden increase in Other to recategorize those to A/C or refrig. Your ML may not allow you to say 3800 Watts suddenly moving to Other category is still the A/C, but with several second resolution in exported data WE COULD. !!! GREATLY compensating for Sense lousy device detection. Thanks
1 Like
There is an ulgly but workable method to get even sub-minute data from the Power Meter.
Without going too crazy, you can get around 2s resolution from a 1hr screen grab using WebPlotDigitizer.
It looks like the dev on WPD for R is on hold for last few years but the current web version is functional.
Here’s a DATA (= red, shifted 20px in Y-axis for clarity) vs SCREEN GRAB (gray) 1hr plot of 2670 datapoints of what would ideally be 3600. My screengrab dimension was 2456 x 1090. Not trying too hard. You could of course display a smaller time chunk and get higher resolution. Keying in the timescale in the web UI helps keep things mapable.
There are a couple of anomalies I didn’t bother fixing in the .csv export.
3 Likes
I would like one second interval total power data to be available for download so that I can better understand the instantaneous load on my electric panel, and how it relates to NEC calculations required for sizing electric services to residential installations.
Thanks for liking the request… The good news for you is that you can partially get what you want out of the Power Meter which shows 1/2 second interval RMS “instantaneous” power. I put the “instantaneous” in quotes because true AC instantaneous power goes up and down 120 times per second.
The 1/2 second interval spike are likely to be somewhat higher than 1 second intervals, but close enough for you to make NEC sizing calls.
Thanks! It’s not obvious how to download the information from Power Meter - please give me a hint!
You can’t download the Power Meter output but you can look at your long term history and adjust your vertical scale by zooming out horizontally and vertically so that you can gauge your max spike.
I can tell from this view that my max power is in the neighborhood of 28kW.
I am a new user of the Sense App and device. I have been debating installing a Schneider PM or something less robust. Full disclosure - I am a Schneider Electric employee.
The desire to monitor and log detailed power data for my home is to build a demand profile. This demand profile is the basis for sizing and designing a pico-grid for my own use. Standby generator, PV, and BESS sizing decisions require much higher resolution for proper sizing, and load management decisions.
I am fully aware of the data bandwidth ramp that builds with increased resolution. If we can pick a resolution AND a time period, so as to limit the data bloat, that would be fantastic. The Sense Power Meter gives me a near RT understanding of when I would want to pin the logging function window.
Hopefully this will be available soon. Even 15 minute interval data, while better than 1 hour, is not sufficient resolution.
My ugly hack above is eclipsed by @kevin1’s in-depth work here:
Sounds to me like you might want to follow that thread and could well contribute some insight.
BESS & pico & “load management decisions”: excellent!
This was a nice and concise blog you may appreciate around these issues (“load shifting”) if you didn’t run across it already:
@kevin1 has also gone well down the rabbit hole on TOU and Tariffs … I’m guessing that will be of concern to you as well.
1 Like
Thanks - I can do the screen grabs, and then manually diddle with the data, but this tool has the potential to really leverage a user’s load profile for good!
I’m probably in the minority of users, but data is only as good as it’s application.
No TOU tariff’s out here in PP&L country for small resi customers. And when I get my PV + BESS system up, I won’t really care!
I am a power control solution architect for microgrids in North America. I see this data dilemma all the time. From customers with 500kW peaks to 15 MW peaks. Ignorance is NOT bliss!
I hear you. There’s the post-usage analysis you can do with screengrabs and so on but the RT actions is where the “money” is.
In the Sense world I would argue that a “peak” (as you highlight) is a complex beast. Thinking out loud and in RT here: there are peak watts and peak complexity. In many ways this is what residential usage (and supply) vs commercial usage (and supply) presents. In the residential setting, peak watts is generally much more controllable than peak complexity. In the commercial world you establish peak watts more reliably as a consequence of the larger scale, for obvious reasons, and care less about “complexity” because, traditionally at least, usage is less complex if only because it’s better analyzed and controlled and/or the scale of operation simply smooths things out.
To that end, an actionable complex Sense’d event (be it manually configured or an automated) is going to be more beneficial than a “peak watt” action.
Can Sense help answer key questions like: Are BESS at the edge better than grid-scale storage systems of some form?
It depends on how you define “controllable”. Concurrent peaks, which is what I think you are saying is “complexity”, is absolutely where the resolution is required that we are asking about. In-rush currents, and concurrent load events, are what determine peak “watts” and is part of the decision making process when sizing DER, and applying load management logic.
Anecdotally, the “skinny” peaks are starting/in-rush current peaks in kW demand. The stepped/stacked peaks are concurrent running demand. A stiff grid is what we depend on to deliver these overlapping demand periods and unpredictable (without control) spikes.
1 Like
Even though I’m the one who originally requested, I can see why this would be challenging to add. Today, the hourly and daily export piggybacks on Sense’s Trends data aggregation - the data is already living in the core database. And I’m betting Sense has designed the hourly and daily aggregations to be hyper-efficient (incremental and on a per demand basis), so as to be inexpensive in the cloud.
Adding a new smaller-resolution export needs a new aggregation function. It might be cheaper to have a different “export Power Meter” function that dumps the raw data from a selected Power Meter for a day or a week.
1 Like
With respect to BESS as a DER vs “grid scale” (a nebulous term, to be sure) it all “depends”. My goal is a reasonable level of energy autonomy. Grid-scale energy storage is fundamental to meeting grid stability requirements (NERC-CIP) while accommodating the destabilizing impacts of variable renewables like wind and PV.
2 Likes
Exactly - like wave form captures in a PQ meter, you have a running log function at the higher sample rate, but for a limited window period. Turn it on, and it only logs for X hours/days/weeks. FI/FO data flow.
Quotes above sum up, I think, the current problem. So to speak.
Ignoring those with welders and plasma cutters, existing and near-future devices (and so homes) inevitably move away from “current rushes”.
PV + EV + BESS is easy to define in this sense: Solar energy is typically on a daily bell curve; fast moving clouds are easy enough to deal with (sudden slow charging!); people move fairly slowly and predictably, even if they want to go 0-to-60 in 3s at 8:59am.
What’s more “complex” is when you consider, lets say, showers & laundry.
In a typical resistive element HW supply there is that sudden current demand; in an HWHP system the current demand is smoothed out and the peaks are lowered. So the definition of “shower time” from a home energy perspective, while equally predictable, becomes something that is more easily shiftable in terms of when the energy should be used to heat water. Add in the ability of PV or thermal solar to trickle-charge a tank, and the ability of a tank to act as a battery, and production vs use is further smoothed out.
Meanwhile, unlike showers, laundry is something that is often easily shiftable. A washer/dryer can sit waiting for activation by Sense. It could even be paused if the sun is obscured. Your laundry time can also be shifted to accommodate supply and demand fluctuations well beyond your house.
This by way of saying (asking), from a grid perspective, the combination of wind and PV and BESS and control is a stabilizing force. No?
The grid system can loosen up as it shrinks, de-centralizes and appears more like the Internet every day.
I think @kevin1’s point about the challenge of increasing resolution is spot-on. I would add that part of the Sense equation really deprecates the notion of high resolution export … the core value of Sense is in the analysis of that data and most of the juicy stuff is happening at high sample rates at the DSP level. I too would love to get all the detail but I zoom out from my own Sense and home situation and think “What does high resolution even mean for thousands/millions of households?”
I have used a Sense as a poor-man’s Fluke 438.
MUCH cheaper and a lot more memory.
A little tricky on some of those really fat wires but we won’t go there.
1 Like
ixu, I suspect you and are work in similar fields. 
I think any reactive load, even at the resi level, is impactful if it shows a kW spike of more than 25% of baseload. They are like vectors, their magnitude and duration can destabilize a small islanded system. They are irrelevant for a full grid-tied system. I have just started using this device, and look forward to summer and seeing what the starting loads for my central A/C unit are.
Careful 'round those fat wires. They bite. 
You’ll probably get a rush on this thread then 
https://community.sense.com/t/graphs-of-utility-voltage-for-detecting-issues/3698/44?u=ixu
I remain frustrated that high resolution data is not available. I’m bumping this because I think high resolution data is imperative to a product such as this. Sure, there are “minimal use case” scenarios, but by the time you’ve purchased a power monitoring tool and installed it in your own breaker box… you’re already part of the minimal use case.
PLEASE, add high resolution data output. If it is a processing burden of some sort, you can even do it like NOAA does for large data sets: you request the data, and they email you with a link once it has been processed and is ready. This can take several hours, but at the end of the day, you have your data.