Water Heater Analysis - Part 3

Hopefully the saga with the water heater is coming to an end soon as the thermosyphon loop that was detailed here has been fixed.  After several months of sporadic discussions with my hvac contractor (who has supposedly been in contact with Water Furnace) about how to fix the problem, how much it was going to cost to fix, and whether I should bear the cost of fixing it, I now have an electronic valve installed on the outlet side of the desuperheater.  The valve only opens when the desuperheater comes on and should completely stop the thermosyphon loop that was sucking heat out of my tank.

 It's supposed to cool off a lot tomorrow and I turned on the hot water heater for the first time since the end of June yesterday.  When the desuperheaters stop supplying all of my hot water this weekend, I'll be able to get some temperature readings from the tank to see if this new valve really fixed the problem.

A Flatlined Hot Water Tank

One of the projects that I've been considering implementing for the house is a water based solar space heater similar to the $2K System that Gary over at BuildItSolar installed.  The basic idea is simple, gather heat from the sun during the day, store that heat in a large water tank, and distribute that heat into the house when needed. One of the issues that needs to be resolved before I jump in and start construction is what to do with all that heat during the summer.   From what I've seen on BuildItSolar and on the SimplySolar yahoo group is that most people use these system for both space heating and as a pre-heater for their hot water.  I would like to use all of the heat during the winter for space heating, but during the summer when I could make hot water, I wasn't sure that I needed it.  I suspected that my ground-source heat pumps with their desuperheaters provided all (or nearly all) of my hot water during the summer.  If that's the case, then my ROI would be significantly lower as the system would only be used during the heating season and would be stagnant for all of the cooling season.

Testing the desuperheaters

To find out if the desuperheaters could handle the entire DHW load I ran a couple of tests.  The first one was simple... turn off the hot water heater at the breaker.   After a short prayer service, the switch was flipped and the hot water heater flatlined at noon on May 24.

The graph above shows that the hot water heater really has been off for the last month.  The tank temperature seemed to be ok as all of the showers continued to be hot.  But to be sure, the second part of the test was to install a temperature sensor on the hot water tank.   The sensor is a DS18B20 and is installed against the inside tank wall underneath the insulation near the upper thermostat.  The sensor readings were spot on with the temperature of the water that I was drawing out of the tank, so now it's just a matter of logging the data every minute and graphing it.

The manual that came with the Water Furnace heat pumps says that the desuperheater circuity kicks out at 130F so I should never see tank temperatures above that.  The 5 day graph above shows that the tank temperature slowly drops during the early morning when the heat pumps run infrequently, but the temperature recovers nicely during the day and if it gets above 90 for any length of time during the day, there's a good chance that I will max out the tank temperature.  The big dip during the day on the 26th was a shower immediately followed by a bath just to see what would happen.

What I've found from this is that my desuperheaters can provide all of my hot water during the summer months as long as some thought is given to when things happen.  So, that brings me back to my original problem.  If I build a water based solar collector, what do I do with it during the summer months?

Hmm...

Water Heater Analysis - Part 2

After some discussions with Gary over at BuildItSolar I've decided that the hot water heater was coming on much more often than it should for its warming cycles when the desuperheater was turned off for the winter.  Over the past month it has averaged coming on for 5-6 minutes about every 90 minutes.   This results in about 90 minutes of "on" time per day just to keep the tank warm.  The graph below shows the warming cycles of a typical day.

90 minutes a day @ 5500 watts = 8.25kwh per day = $.90/day.  Since I keep the desuperheater off from Nov-March, that's about $140 per year in tank warming cycles.  The question is why is it coming on so often.  Gary suggested that I might have a thermosyphon loop occurring and after some investigation I believe he is correct.  The downstairs heatpump sits right next to the water heater and the plumbing between the two units looks like this:

When the valve on the cold water line at the top of the water heater is open (as it has to be when the desuperheater is on), there is an instant flow of hot water out of the top of the water heater, down through the heat pump and back into the base of the water heater.  Shutting the valve stops the flow and the lines stay cool.  Even though all the piping is insulated, shutting the valve had an immediate affect on the number of tank warming cycles that occur during the day as shown below.

The tank now cycles on for 7 minutes every 6 hours.  The "on" time for warming cycles has been reduced from 90 minutes to 30 minutes a day.  For my 5500 watt water heater, that's 825 kwh per winter that I'm not using and 982 pounds of CO2 that's not being pumped into the air.  In addition I'm also saving $.60 per day or $90 a year.

Water Heater Analysis

Since I've been able to work around the load profile problems with the TED 5000, I've started collecting enough data that I can start to change some things and see what affect it has on my hot water heater. 

This is a typical load profile for my hot water heater for a single day.  The heater was on for about 125 minutes of which 34 minutes was due to the shower at 8am.  The other 90 minutes of run time during the day was mostly just the tank keeping the hot water that I'm not using... hot.

I then turned on the desuperheater for the heat pump.  I usually keep this turned off during the winter and really had no way in the past of telling how much heat it supplied to the hot water tank.  Tank run time for this day was 68 minutes for the two showers.  Note that at no time during the day did the hot water heater come on to keep the tank warm.

Is this good?  I'm not sure.  During the summer months the desuperheater is always on as any heat that I can pull off the coil and put in the hot water tank is heat that doesn't get pumped out to the ground loops.  But during the winter, any heat that I pull off the coil and put in the tank is heat that's not going into the house.  So either the hot water tank comes on more often, or the heat pump runs longer.  I'm not sure which option is better.

Some of the issues/thoughts with using the desuperheater during the winter that I see are:

1. If the desuperheater is off, the hot water heater will only come on when the temperature in the tank drops below a certain level and the water actually needs to be heated.
2. If the desuperheater is on and it's a nice day outside such that the heat pump doesn't come on, it's the same as having the desuperheater off and the hot water heater supplies all of the hot water.
3. If the desuperheater is on and it's cold outside, whenever the heat pump comes on, heat is pulled off of the coil and pumped into the tank even if it doesn't need it.

This last thought is the one that I have the most trouble with as it seems that I'm wasting money.

TED 5000 Load Profiling

The HA software has been logging "seconds" data from the TED for a couple of days and I've started to do a little bit of post processing on the data.  For the most part detecting when the hot water heater kicks on is easy.  Below is a snapshot of data that I'm pulling down from the TED and saving.  The log has date/time and usage in watts.

03/05/2011 14:34:21, 1362
03/05/2011 14:34:22, 1362 
03/05/2011 14:34:23, 1362  
03/05/2011 14:34:24, 6897  <-- DHW ON.    Load change of ~ 5500 watts
03/05/2011 14:34:25, 6897

This is easy to parse and detect that the water heater came on.  But it's not always this simple to detect.  Yesterday the water heater came on over a few seconds...

03/09/2011 00:49:59, 2796
03/09/2011 00:50:00, 2796
03/09/2011 00:50:01, 4109  <-- started
03/09/2011 00:50:02, 4109
03/09/2011 00:50:03, 4109
03/09/2011 00:50:04, 4109
03/09/2011 00:50:05, 4109
03/09/2011 00:50:06, 8104  <-- full load

And later in the day it came on while the heat pump decided to turn off.  This is a more difficult situation to detect as I have no data to indicate a load change of ~5000 watts.  Best I can do is detect the off event and back track and try to guess where the on event occurred.

03/09/2011 07:13:53, 2694
03/09/2011 07:13:54, 5697
03/09/2011 07:13:55, 5697
03/09/2011 07:13:56, 5697
03/09/2011 07:13:57, 5697
03/09/2011 07:13:58, 5697
03/09/2011 07:13:59, 6153  <-- DHW ON, heat pump is shutting off.

I can set up the HA software to detect most of these different "on" events because I know what's on each panel and I know that there isn't another load like the water heater on panel 2.   These multi-second on events is why I think that the load profile software within the TED has difficulties and sometimes misses either the on or off event.  The profile that it learned was for 5500 watts almost instantly and for most cases this is how the water heater comes on.   If the device doesn't come on like that every time, then the TED is going to have issues detecting it.

Detecting multistage loads while dealing with other loads going on/off will be challenging.

TED 5000 Load Profiling

I finally got the load profiling data for my hot water heater that I could not obtain directly from the TED.   I went through a couple of iterations of retrieving data from the TED before deciding on a workable method.  I tried retrieving the historical minute data but that didn't give the resolution that I wanted.  So I switched to retrieving the historical second data.  It's a lot of data and the TED is really slow at delivering the data via XML.  I found in the API where I can get the raw second data and this downloads much faster.  The TED gateway only stores about an hour of second data so I have my HA software downloading the second data every 30 minutes and saving it off to disk for later processing.  This will be a workable solution.

The graph above is from the first full day of pulling second data from the TED.  It was a very low usage day for DHW which is what I wanted so that I could get a feel for how often the water heater comes on and for how long.  Total run time for the hot water heater for 3/6/11 was 83.95 min @ 5500 watts for a total of 7.7 KWH.  Almost all of this is the heater coming on just to keep the tank hot.

Fighting the TED

I'm not sure what's going on with the TED 5000.  Soon after I installed it I tried setting up the load profiling and noticed that for multi-stage devices like my heat pumps it has a very difficult time determining when the device is on/off.  So I gave up on the heatpumps and decided to just look at the water heater.  I figured the easiest thing to do would be to manually enter the load data for it since I know what it's going to be (5500).  After doing that, the Load Profile tab would correctly show if the water heater was on or off, however, the historical data portion of that page always showed "No Data".   I thought that maybe manually entering the load was causing a problem so I turned off the water heater, took a quick shower, started the load profiling, and turned the water heater back on.  It took about 10 secs for the load to be "learned".  After a day of running with the learned profile, I still have no historical data.

I called Energy, Inc since their forums are offline (it appears they didn't renew their domain name) and managed to get a support person that has "never played with" the load profiling. After chatting with him and one other support person, it was determined that they've never heard of this issue before and suggested that I upgrade to the latest firmware.  Since that will wipe all monthly historical data, I'll have to wait until first thing in the morning on the day my billing cycle starts.

I need to get this figured out and collect some data now so that I can see how much the water heater comes on (or doesn't come on) when the ground source heat pumps are heating the tank.

Heating the septic tank

The TED 5002-G that I installed in early Jan has got me thinking about a lot of things.  The other day when taking a shower it dawned on me that all that money I spent heating hot water really did go right down the drain.  It takes < 5 secs for the water to leave the shower head and enter the drain at which point all it does is heat the septic tank.  There must be a better way!   If only I could recover that heat to preheat the cold water going into the water heater.

I got online later that day and found a graywater/drainwater heat recovery product.  The GFX website is web design hell, but it or a product like it should be required by code for all new building installations.  Lord knows the builders aren't going to install it just because it should be.

Green Gate Guest House has a really good article on the one they installed.

The Energy Detective

This year for Christmas I bought myself a TED 5002-G, handed the box to my wife who put a bow on it and handed it back.  How she always seems to know what I want for Christmas is beyond me.  What a woman!

For me, this thing is awesome.  Mainly because it makes me aware of what I'm doing.  The footprints software can be a little buggy at times especially if you swap back and forth between views.  The load profiling seems to have some issues.  I manually entered load values for some devices (like the hot water heater), and when I go to the load profile page, it can correctly tell me what the current state of the hot water heater is, but it can never find any historical data for it.

Some interesting things I've noticed so far... The house hums along at about 1.5 kwh of background usage.   Is that a lot?  Some of the things that are "always on" are:

• 1 desktop system with external drive
• 2 laptops.  Even when they sleep the power supply is always plugged in.
• My electronics prototyping board
• Router, cable modem, wireless router, 2x switches, 2x powered cable splitters
• DVR, AV system, 5x TVs in standby mode, 3x DVD, Cordless phone base plus 2 charging stands
• 6 lights (4 CFL, 2 Halogen)

This seems like a LOT and I'm sure I'm not listing several things, but it's better than it was!  It's taking some time to convert some of the other residents of the house to my way of thinking.

The hot water heater kicks on more than I thought it would just to keep the tank hot with no usage.  It's an 80 gallon, 5500 watt tank. The energy usage sticker on the side estimates 5200 kwh per year at a cost of $440.  Based on the data that I see from my TED unit, the water heater comes on about once an hour for about 4-5 min (these are estimates since the load profiling is buggy).  That's 9-11kwh every day, roughly 4000kwh per year, just to keep the tank at temp.  With my electricity rates, that's $400+ every year if I never use a drop.

Heating and cooling is by far the biggest use of electricity in my house.