Page 1 of 2 12 LastLast
Results 1 to 13 of 19

Thread: Analytics

  1. #1
    Join Date
    Jun 2006
    Posts
    172
    Post Likes

    Analytics

    Can someone make the case for analytics showing a real payback in the commercial HVAC controls market? I'm not sure I believe the hype.

  2. #2
    Join Date
    Nov 2015
    Posts
    280
    Post Likes
    Quote Originally Posted by theoldman View Post
    Can someone make the case for analytics showing a real payback in the commercial HVAC controls market? I'm not sure I believe the hype.
    Great question. In the past the golden example for analytics was looking for mis-operational economizer dampers which is perfectly discovered using logic in a controller. Same type of thing for valves.

    However, the big thing is that facility management departments don't even have a proper budget for maintenance, so they aren't going to fix things anyway in a reasonable time frame.

  3. #3
    Join Date
    Feb 2008
    Location
    Purcellville Va.
    Posts
    912
    Post Likes
    Quote Originally Posted by theoldman View Post
    Can someone make the case for analytics showing a real payback in the commercial HVAC controls market? I'm not sure I believe the hype.
    I think you need to be a little more specific.

    one building or multiple buildings?
    up to 50k sqft
    50 to 250k sqft
    over 250k sqft

    and as better duck eluded too, I'm guessing analytics only makes sense if issues identified are dealt with.

    Can it work...I'd say yes. Is it worth it? Depends....

    I bet enrdapt probably has lots of data to the affirmamtive

  4. #4
    Join Date
    Jul 2004
    Location
    Kingston, Ny
    Posts
    641
    Post Likes
    Analytics are only as good as the implementation. I've been working in Analytics for the last 4 years.
    1. It is expensive, so you'd better have a lot of really good screwed up systems.
    2. What is your (customers) goal? Energy, operational, insight, capital, ?!!?
    3. Moving data _will_ get expensive over the life of the service (albeit it is getting cheaper), choose big data (lots of points polling lots of time all day), local (nested analytics, building or campus) or edge analytics (in the field, on the equipment itself). The costs between the choices are significant. i.e. a cellular or business connection alone can cost $700-$1200/yr + data traffic per site.
    4. Big Data (in-house or SaaS), Analytics Platoform (iConics, Splunk, Panoptix, SkySpark, Predix, Pivotal) servers, staff, etc is an ongoing cost. Equipment once fixed dries up the Analytics $$ well. Then it becomes an expense and can ( and has) become a giant Monstrous expensive thing that once saved you money. Don't forget, most larger systems out there take your data, and then sell it back to you. What you do with it is not their problem.
    5. OpSec. What security posture are you going to employ. You can shodan, wire shark, monitor and scalp nearly every system out there and most, if not all are vulnerable to attack. How many systems do you know actually use the TLS 3rd party signed certs, dns names, or some mechanism for a level of NIDS? Not many.
    6. Analytics can be done and it can be done well, but you need a plan, it can work, it can be a sustainable operation, a lot of thought needs to go into every step. Most co.'s just run in and buy the html'est flashyest report producing and gee whiz bang zoom systems that a HiPPO (Highest Paid Person Opinion) gets sold on, and you end up working for the system that was intended to work for you, but you're saddled with it so they don't look bad.
    -Good Luck!
    Happiness is the only good. The time to be happy is now. The place to be happy is here. The way to be happy is to make others so.
    -Robert Green Ingersoll

  5. Likes alper liked this post
  6. #5
    Join Date
    Mar 2006
    Location
    Florida, N.Y.C.
    Posts
    54
    Post Likes
    As a distributor of analytics software for the last 7 years, I would like to share my insight.

    Let's look at it from problem domain.

    We humans are really good finding problem correlating it a solution but only have certain attention span. Meaning we can't scale it or can focus long terms on the other hand computers with instructions given can replicate our initial findings without getting tired.

    Within building small or large. #1 They generate a lot of data #2 They run 24/7 where we can't analyze manually all issues that can happen with these moving systems.

    This is where analytics come in handy. We are good at finding correlations in our systems once and we can replicate that with analytics software consistently results are fantastic. Basically we repeating our past knowledge and applying all the time. When your past knowledge on equipment is replicated to multiple assets within your portfolio, it generates a lot of tasks.

    These results are the key to lengthening the life-time of your equipment and reducing energy usage. There is no question about that.

    The question rises from changing business model. Analytics do generate some noise and many very important issues that needs to fixed right away. Usually what happens is that companies spend money on analytics and the findings overwhelms them..

    Corporations if they are applying analytics on business assets shall start small. If 50% of budgets is on analytics the other 50% should be on operational repair costs. This will enable the company save money on the energy usage and the following year they would have bigger budget for analytics and operational repair costs.

    Companies that do not foresee this, and they focus on analytics results only but don't allocate business processes and finances to do anything about it after findings.

    Analogy to this would installing car oil check analytics to your car. Light comes up and you have no process or resources to change oil. Your engine burns up and you blame the analytics. Problem is not the light switch but the matter of not changing your oil.

    On the other hand, there are companies look at the findings and implement the process to repair their systems and assets, resulting in large savings and allowing them to have better budgets for the next round of continuous commissioning.

    So in short, start small have budget and process to react and repeat larger scale with new resources.

    Cheers.

  7. #6
    Join Date
    Aug 2016
    Location
    Sacramento, CA
    Posts
    148
    Post Likes
    What do mean by "analytics?"

    The term analytics has been drastically overused in our industry - The same as "machine learning" or "IoT." The reality is that most analytic-driven results are a better mousetrap than the traditional alternative of a human engineer. An engineer can correlate data (i.e. supply air temperatures and valve positions) to spot anomalies the same way a computer can.

    Why most analytics don't have a real payback is there is no true, actionable intelligence. Fixing twenty issues in a building may show up on your bill, but how do you know which repairs were actually effectual? What repairs should be replicated? Are the savings you found on your energy bill accurate, or was it simply outside factors acting on your building?

    We have some exciting technology built into our analytics and optimization platform that fills a lot of gaps in today's offerings and would be happy to talk about it offline.

    The reality is, though, that most end users fail to establish goals with new analytic platforms and therefore fall short when trying to discover actual payback for their investment. Finding problems in how your HVAC operates is great, but is only half the battle.

  8. #7
    Join Date
    Nov 2015
    Posts
    280
    Post Likes
    With all the poorly first cost selected equipment and systems analytics are verifying what somebody already tried to explain would happen.

    Select your equipment and systems better and provide diagnostics and repairs and the need for analytics diminishes dramatically

  9. #8
    Join Date
    Jun 2009
    Location
    Portland, Oregon
    Posts
    33
    Post Likes
    I agree with alper. We started doing analytics about Five years ago. I find it very interesting. As we go around and introduce this product to our customers, I can see in their faces that they just don't get it but these are the same customers that really don't do maintenance anyway. For these customers, I really push them away from it. I don't want to sell them something they will not use.

    Like alper said, if you don't have a budget to fix it then why get it. Also, setup is critical. Crap info in, crap info out and this system is all about info (data data data)

    for customers with a real energy manager, then it is a great tool. With our system, we can put a Dollar value to how much a leaky damper or valve is costing you or a air handler that is running out of it's schedule. This way you can prioritize your repairs.

    we have seen the system give paybacks as short as 2 months so the investment does have the ability to pay for itself rather quickly.

    the other way I explain it is that it is a completely different way to look at your building. Most operators just look at it from a comfort standpoint. "If I don't get any complaints then everything is working great"...now you can put a dollar value to every piece of equipment running and see how much it is costing you. in the past (without some real fancy programming) you really could not do that. putting that air handler in MANUAL 24/7 so you don't get complaints is costing you $XXXX and you better address the real issue instead of band aiding the problem.

    the other fact is the controls industry did not come up with this. Other very large companies have used this to get a grasp on processes and it has worked. We just adopted it and made it our own. IBM, Dell, Microsoft, ect.. all used it way before us.

    in short, if you don't do repairs now then don't get it. If you do repairs ( or have a budget for repairs) and need to get a grasp on your energy bills as well as find hidden problems then get it. Make sure what you are getting is 100% compatible with what you got and make sure the data you have is real. Learn to use the system and if you are a power user, find a system that gives you access to tweak it. you can change the parameters of what it looks for as you get a handle on your worst performers and get them under control.
    one more reality, I am sure they ALL have monthly fees, another reason not to get it if you don't plan on making the most of it. just more money you are throwing away along with your energy bills.

  10. Likes alper liked this post
  11. #9
    Join Date
    Nov 2015
    Posts
    280
    Post Likes
    If you do not understand how to save energy on the consumption side I can also see why analytics hold more appeal.

  12. #10
    Join Date
    Jan 2003
    Location
    USA
    Posts
    3,292
    Post Likes
    I'll agree you need to have the expectation and budget to repair equipment. If not, tossing software at (*#@$& equipment will not do anything. Start small and if it proves itself move deeper.

    There are large customers that acquired all sorts or nonsense over the years and analytics can provide benefit. IMO today analytics are the wild west. I suspect we will see them cooked into the every controller to some degree over the next decade. Proven, published results are pretty hard to come by and likely for good reason.

    Enerdapt seem have a good model that doesn't evolve ankle grabbing + humans filtering out crap & tuning alerts before it hits the customer. Its a cloud service so sites sensitive to remote connections, not an option. Look around and see whats out there if you have the budget. If the customer thinks software is going to fix equipment/design issues.....dream on.
    Propagating the formula. http://www.noagendashow.com/

  13. Likes alper liked this post
  14. #11
    Join Date
    Jan 2003
    Location
    USA
    Posts
    3,292
    Post Likes
    full disclosure...

    I spent 0$ with enerdapt, have not talked with alper, but do deal with other packages...some of which I'm not very impressed with. Look around, see what's out there.
    Propagating the formula. http://www.noagendashow.com/

  15. Likes bigguy158, alper liked this post
  16. #12
    Join Date
    Nov 2015
    Posts
    280
    Post Likes
    Quote Originally Posted by orion242 View Post
    ... IMO today analytics are the wild west.
    In the US market it seems people like to go after specific vendor solutions rather than standards. I see the EU latching on to ISO 50001. We will see where that goes. And, out of Germany they have had this Identification Key method to identify what is where for a long time now. Makes analyzation easier and more open to everyone.

    More recently we see a bunch of US "farmers" pushing their method.

  17. #13
    Join Date
    Aug 2016
    Location
    Sacramento, CA
    Posts
    148
    Post Likes
    We developed our analytics/auditing software as a basis for our energy business. We wanted to know where, why and how if something was costing us money as we do a shared savings model with customers. Every dime we save customers we make ourselves - It's a true value sale!

    What we've found is the "analytics and IoT" markets are dominated by software companies who are educating themselves about HVAC - Rather than HVAC engineering firms (Ourselves) that develop a platform. The problem with these companies is they're spamming the market with poor sales literature and diluting the efficacy of energy savings and diagnostics. Most are more interested in fiddling with a thermostat rather than digging into how primary/secondary HVAC systems operate to do actual analytics and optimization.

    One marketing campaign we recently competed against was "heirarchically correlated alarms." Basically the system would recognize that a VAV alarm correlated to an air handler alarm which then correlated to a chiller plant alarm. In theory, this sounds like a major breakthrough. Yet, when you look at a basic HVAC system you realize that a primary/secondary airside system is intrinsically heirarchal. Meaning your system should have caught the chiller alarm long before it hit the VAV or zone level. Even if you don't have monitoring on the chiller your air handler should still detect the SAT deviation from CHW Valve position long before it affects tenants.

  18. Likes fisha liked this post
Page 1 of 2 12 LastLast

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •  

Related Forums

Plumbing Talks | Contractor MagazineThe place where Electrical professionals meet.