Talk About the Latest in Home Automation/Home Electronics -
Home Automation Forum

Smarthome Forum
Insteon Home Automation
Login or Register
 
Home | Profile | Register | Active Topics | Search | FAQ | Smarthome
 All Forums
 General Discussion
 Energy Saving Solutions
 automated standby power elimination
 New Topic  Reply to Topic
 Printer Friendly
Author Previous Topic Topic Next Topic  

quartOmilk
Starting Member

USA
6 Posts

Posted - 10/25/2016 :  7:13:56 PM  Show Profile  Reply with Quote
Hey all! I am putting together an application to eliminate standby power automatically. Please note that this project is for a grad course and I am looking to see if there is any interest in this project becoming a product commercially! Don't worry, I am not planning on charging for the product itself!

This is not solicitation as I am not asking you to buy anything. I just want to know if there is any interest out there. If you are interested, hop on over to my wix page or watch the video below and then hop over. Even if you are not interested, if you could hop over and leave a comment on the form at the bottom of the page. We will have the page down in about 30 days unless I get enough feedback to make it interesting.

Thank you all for the help!

http://niyi73.wixsite.com/vampire-draw-iot
https://goanimate.com/videos/0NumQAWW2FKM

~quartOmilk

quartOmilk
Starting Member

USA
6 Posts

Posted - 10/26/2016 :  6:36:28 PM  Show Profile  Reply with Quote
Am I to assume that there is no interest in this product or am I to assume that the readers of this post do not trust the content? Either is fine, I am just trying to capture data.

If you do not trust the content I can provide a quick explanation here and you can comment here on the interest or disinterest.

The product we are trying to develop is, working with connected outlets, develop an IoT software controller that will incorporate AI and other data sources (weather, user profile, time of day for location, etc) and eliminate the electrical draw that is being wasted. I assume everyone on these sites knows what standby power is due to the fact that you are in a Early Adopter/Innovator technology forum. This is what we are attacking. It can eliminate an average of 10% of total electrical demand simply because (based on the government statistics and collected data consumption) 30-40% of all electrical demand is waste.

We are offering this application free of charge and are currently seeking out the companies who have connected outlets to take on their software aspect for adoption.

In other words, you use this application and you save money by doing nothing.

Let me know your thoughts! I figure this is a acceptable alternative to the links if you do not trust them.

Thank you everyone!
Go to Top of Page

Geo
Advanced Member

Canada
732 Posts

Posted - 10/28/2016 :  08:06:26 AM  Show Profile  Reply with Quote
I have two comments:

1) Stand-by power consumption seems to be overstated by at least an order of magnitude. To minimize my consumption of electricity all my lights are LEDs and heating and cooking is by natural gas. I measured the stand-by draw and came to approximately 1% of the total. Then what is the nature of the alleged "30% to 40% waste" reported by the government? Is it inefficiency of appliances we cannot eliminate anyway?

2) IoT has been a buzz-word for some time and now everybody wants to jump on the band-wagon. However, experts warn that most of the systems in existence and under development have inadequate, if any, security. Consequently, they warn, a "big IoT hack" is waiting to happen.

GJN
Go to Top of Page

quartOmilk
Starting Member

USA
6 Posts

Posted - 10/28/2016 :  08:16:17 AM  Show Profile  Reply with Quote
Fully agree with the IoT aspect you stated in point 2. Just the other day there was a major attack using IoT that brought many systems on the internet to their knees; Netflix and Amazon just to name a few.

As far as the standby power; yes the government is considering waste such as heat, and inefficiency, in this system. I am not even looking at that, so yes, the 30-40% (35.6% actual as of EOY 2015) is slightly misleading but works for a pitch point.

When I clocked the devices in my home using a Kill-a-watt meter I was able to identify ~12% of my electrical bill as standby power. Of that 12% I was able to, at this time without being overly aggressive, lower that down to 4%. I brute forced it based on my awake/sleep/work/home patterns.

Now, with that being said, when I did this use test I was a DINK so very predictable. I would imagine SINK home owners are just as, if not more, predictable as well. My 12% also did not account for all electrical devices I had plugged in as I didn't even measure the ones that I wasn't sure how to manage without breaking a system they were put in place for (DVR for instance) and I did not measure any of my lights in the home because I just have not gotten there and like you I have replaced all of them with LED, so waste is minimal at this point.

Thank you for reaching out, I appreciate the feedback!

Steve

Edited by - quartOmilk on 10/28/2016 11:13:09 AM
Go to Top of Page

EVIL Teken
Advanced Member

Canada
2316 Posts

Posted - 10/28/2016 :  09:07:49 AM  Show Profile  Reply with Quote
Hello Steve,

Would you be able to offer a little more insight about the *AI* aspect and how you envision this to operate? Also how would your software determine when things are supposed to be on vs off? Is the users behavior and energy use being tracked and later the software (learns) the user patterns?

I ask because there are lots of variables to be successful in energy monitoring / energy management. For your point of reference I have been heavily involved in both as a consumer and in a professional role.

Teken . . .

Want to make a real difference? Cast your vote to make the PLM Pro a reality: http://forum.insteon.com/forum/main-category/new-insteon-device-wish-list/8221-plm-pro
Go to Top of Page

quartOmilk
Starting Member

USA
6 Posts

Posted - 10/28/2016 :  09:52:45 AM  Show Profile  Reply with Quote
Sure thing Teken,

So the AI would be just as you stated. Initially it would require a lot of profile management of the user base. Based on location, and data like if they are single, married, with/without kids, pets etc it will begin to build heuristics. As the AI begins to attempt to determine when to turn things on/off if the user wants to override they are more than welcome to do so and it will record this. This is done so that when it ends up in a decision point it will make the best educated decision.

Of course there will be times where it may be wrong. For instance you decided to break your pattern and stay home from work today. The best way to combat this is as the system becomes more robust, partner with other sensor companies and allow this AI to read states. It should also have the ability to understand states of other devices and notice that there is an anomaly in the usual pattern.

I would like to also partner with a company "Sense" who can identify the electrical signature of devices in your home. Obviously there are limitations on this (light bulbs for one) but this will help make sure I do not force the user to establish an exception (no one wants their refrigerator turning off to save from a non-existent but recognized vampire draw)

The early part of this product would be heavily dependent on the community, which is why I came to the forums to ask for advise. It would be the members of the forums, who are typically early adopters and great customers, who would help build this product for the whole community. When it enters mainstream the customer base begins to get tricky but encompasses almost 83% of all potential customers so the learning from the early adopters/innovators (i.e. the members of this community) become integral to the AI being somewhat fleshed out.
Go to Top of Page

Geo
Advanced Member

Canada
732 Posts

Posted - 10/28/2016 :  4:05:53 PM  Show Profile  Reply with Quote
AI has come a long way since Dr. Marvin Minsky of MIT wrote his Society of Mind. Unfortunately, I don't think we are anywhere close to the level of artificial intelligence we are really striving for. That is one that could make up its own mind, not just execute a long if/then algorithm (I'm oversimplifying to make my point).

As to the software security (IoT), testing for it is even more difficult then for qualifying software. You broke the code, you have proven it is faulty. But how long do you test to prove the software is good with no faults? It has been mathematically proven that a 100% test coverage is impossible - but you can decide what to test and how, then, relying on statistics, take a chance. But testing for security? Even if you are an accomplished hacker there is no guarantee there is somebody better.

But back to your project. Stand-by power was a big deal with TVs and computer displays/terminals using CRTs. Not that much today. I assume that you'd be really interested in eliminating wasteful use of electricity in general. Turn off appliances when not needed, but some big users, like a fridge, you could not. Your efforts may snowball into a mountain of sensors, where, if not carefully watched, the power required for monitoring and processing could exceed the power saved.

At any rate, you should be applauded for your efforts. Just keep in mind that outside the academia, new products must give some real benefit worth their price.

Cheers.

GJN
Go to Top of Page

quartOmilk
Starting Member

USA
6 Posts

Posted - 10/28/2016 :  4:56:21 PM  Show Profile  Reply with Quote
I am fully tracking Geo, and why the more we discuss here, as well as a lot of the other forums I am in, I am seeing a potential pivot necessary to truly help the community as a whole.

Based on the data I have personally collected, and data published online in other studies (Business Insider and Neilson) I know there is a viable opportunity here, but not enough to charge money to the consumer, the app user. I would be giving that away for free because if I am saving you 15 dollars a month total (~10% average electrical bill for family of 4) there is not much room to monetize the savings; however, if you monetize the data collected there is a viable business option.

Now with that being said, this MVP has clearly answered my initial hypothesis and for that I thank everyone. I also appreciated the discussion so I absolutely will run ideas past everyone here because there is clearly a wealth of knowledge in both breadth and width!

Go to Top of Page

EVIL Teken
Advanced Member

Canada
2316 Posts

Posted - 10/28/2016 :  8:46:54 PM  Show Profile  Reply with Quote
quote:
Originally posted by quartOmilk

Sure thing Teken,

So the AI would be just as you stated. Initially it would require a lot of profile management of the user base. Based on location, and data like if they are single, married, with/without kids, pets etc it will begin to build heuristics. As the AI begins to attempt to determine when to turn things on/off if the user wants to override they are more than welcome to do so and it will record this. This is done so that when it ends up in a decision point it will make the best educated decision.

Of course there will be times where it may be wrong. For instance you decided to break your pattern and stay home from work today. The best way to combat this is as the system becomes more robust, partner with other sensor companies and allow this AI to read states. It should also have the ability to understand states of other devices and notice that there is an anomaly in the usual pattern.

I would like to also partner with a company "Sense" who can identify the electrical signature of devices in your home. Obviously there are limitations on this (light bulbs for one) but this will help make sure I do not force the user to establish an exception (no one wants their refrigerator turning off to save from a non-existent but recognized vampire draw)

The early part of this product would be heavily dependent on the community, which is why I came to the forums to ask for advise. It would be the members of the forums, who are typically early adopters and great customers, who would help build this product for the whole community. When it enters mainstream the customer base begins to get tricky but encompasses almost 83% of all potential customers so the learning from the early adopters/innovators (i.e. the members of this community) become integral to the AI being somewhat fleshed out.



Hello Steve,

I would humbly suggest if this business venture is for the long term and serious. That you avoid using any of the virtual monitors like sense and engage Brultech.

They have been making energy monitors for more than 20 years. They don't use any kind of fairy dust software guessing to determine energy consumption or what loads are active / inactive.

There are countless *Me Too* companies all offering the very same like sense. None of them including sense will be around or selling their wares in five years.

Trust me I've been in this space for more than 25 years and have seen it all and 99% have come and gone.

Teken . . .

Want to make a real difference? Cast your vote to make the PLM Pro a reality: http://forum.insteon.com/forum/main-category/new-insteon-device-wish-list/8221-plm-pro
Go to Top of Page

quartOmilk
Starting Member

USA
6 Posts

Posted - 11/03/2016 :  06:36:32 AM  Show Profile  Reply with Quote
Thank you Teken!

We will most likely pursue some version of this project. The market by itself isn't large enough to sustain a business, but with a small pivot and incorporating this project as a feature it now becomes doable and provides companies a competitive advantage that use this.
Go to Top of Page

jec6613
Starting Member

USA
5 Posts

Posted - 12/13/2016 :  7:40:38 PM  Show Profile  Reply with Quote
One other thing to consider is that an IoT device is by itself a standby user of electricity. Sure, only 1-2W, but as with everything else, that starts to add up.

Additionally, many devices require standby mode for at least part of the time. Modern TVs can take 30-60 seconds to power up if standby power is removed from them, and then immediately run a lot of power hungry checks, leaving them about neutral. Microwave ovens and similar are fine to an extent, but they must have a persistent battery to keep clock time and so on - and, yes, if you're doing this, clock time is darn important since the security required needs accurate clocks (+-5 minutes is normal for Kerberos).

Smart power strips that turn off monitors when computers go to sleep, and turn off subwoofers when the receiver is powered down, and multi-USB outlet devices which minimize standby power by only having only slightly more vapiric device instead of a bunch of small ones are all helps, but IoT brings with it other electrical costs that you may not even think of:

The need to keep an Ethernet or WiFi PHY interface in standby at least part of the time, the additional load, even though small, on the central DHCP, and with IPv6 the need to route more traffic, et cetera. Every packet transmitted uses a nonzero amount of power from the modem, router, et cetera, and it all adds up. Keeping devices relatively dumb generally reduces such loads for now.
Go to Top of Page
  Previous Topic Topic Next Topic  
 New Topic  Reply to Topic
 Printer Friendly
Jump To:
Smarthome Forum © 2000-2017 Smartlabs, Inc Go To Top Of Page
Powered By: Snitz Forums 2000 Version 3.4.07