JMP & Energy grid + IoT? Where can JMP products add value? Help needed!

georgeplatings

Dear SE Community,

 

I've come across a company that is searching to monitise it's data providing from IoT. And I'm comming to the SE community to pick your brains and find out if we can solve their challenge which can be a great customer story!

 

The Data they collect from their propriatary devices (smart meter), installed in domestic and professional settings, can report/collect data every 10 sec to 1 month (depending on user agreement). This Data can go quite deep by using home appliance "signatures", meaning they can tell what model of fridge you use and if it is functioning correctly.

 

They are sitting on this goldmine of Data and are brainstorming hard to find a way to make a commercial model out of the exploitation of that Data. At the moment this company makes profits from the sale of their smart meters which collect the Data (IP protected) and are backed by National Grid to increase efficiency through Volt/VAR Optimization.

 

This is a US HQ company with users in UK & Ireland (which explains my personal interest as it is my territory, and I'm keeping their identity quite for the sake of the exercise). My contact assured me that they are "Cash Happy" and I'm turnning toward the Employee Community and SE's for your breadth of knowledge and expertise to try and get some great ideas to discuss with them and more importantly how any JMP product can fit into their journey toward a viable commecial model.

 

Their main clients (revenue streams) come from energy providers and I can see an interesting level of synergy between them and appliance manufactures e.g. Bosch/Samsung/LG...

 

Feel free to answer me on the Employee Community and/or via email - george.platings@jmp.com

 

Also, I'll be happy to share the companies identity if there is no conflict of interest.

Daniel_Valente

From an analysis side, I think FDE has a very interesting play in this space. There are a lot of commerical (and consumer-focused) products that aim to collect the continuous waveform of the energy usage from a meter, and then upload to a cloud for processing.

 

Two examples are Neurio and Sense. The "special sauce" in these platforms is the automatic identification of appliances. A simple use-case would be to alarm ths user if they have left the house (i.e. all devices are away and out of geo-fenced area and the stove is on). But how do you segment a stove from the overall electrical signal coming through the meter (e.g the washing machine might be on, maybe some other devices)? So being able to look at a signal over time and extract the envelope for appliance identification is something FDE ought to help with (Indeed a lot of their platform work is in waveform/envelopment development with the help of end-user appliance training).

 

The other use-case is when you have many of the same devices in a fleet (think big things that spin or respond to wind) and you want to make sure all of them have the same startup envelope in terms of amperage draw, ramp to operating temp/voltage etc. And you want to be able to respond when something doesn't look right, so you can take the asset down, perform some maintenance on it and prevent a more costly failure or unplanned outage.

 

This model-building activity is "Functional Machine Learning" that Chris Gotwalt talks about. To be clear, JMP Pro is not well-suited for the real-time processing of these data (and shouldn't be positioned that way), but can be quite effective in looking at historical data, building models and with the Formula Depot, generating score code to help put those models into production (by converting the models to SAS, C, SQL, Python or JavaScript).