Frozen Solar Panel Incident

What We Know

• Date: 1/3/2011
• Time: Approximately 10:55am situation was detected
• Found using the Green Garage HVAC website
  • Solar Panel Air Temp = 195F (sure)
  • T-out for the panels = 64F <needs to be confirmed...very close to correct
  • Pump 9 = On
  • Tank Temp = 97F <needs to be confirmed...very close to correct
  • The evidence indicated that the panels were very hot but no water was running through the panels as T-out was in the 60's (not it's normal 110F).
• Ambient weather
  • Outdoor temp was 18F
  • Sunny with intermittent clouds
  • Wind was moderate (10 mph range) with gusts up to 26 mph <confirmed via Weather Underground History
  • Wind chill that morning was -2F (from Tom who road his bike!)
• Joe G. took pictures of the panels at about XX:XXpm (< please advise)
  • It showed water leakages inside three panels....Panels #6, Panel #7 and Panel #10.

• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 

What is the sequence for the panel heating?

• Valves and pumps involved in solar heating: Valve 1, Valve 2, Valve 4 and Pump 9
  • Default position
    • Valve 1- Open
    • Valve 2- Open
    • Valve 4- To Array <<< Open or Closed in the unpowered situation? Non-spring, stays in last position. Shouldn't be a problem, located after v1.
    • Pump 9 _ Off
  • On panel start
    • Valve 1- Close
    • Valve 2- Open
    • Valve 4- To Array
    • Pump 9 _ On

What is the location of the failure in the panels?

• It occurred in three panels ... #6, #7 and #10 (last).
• Appears to be a fracture in the copper tubing in the thermal transfer body (under the glass) of the panel.
  • Need to confirm location of the failure occurred in the panel system.
• Key Questions

1. Where is the failure location in panels #6, #7 and #10. ANSWER: In the bottom 18" in the vertical `1/2" copper tube risers.
2. Are they the same? ANSWER: Yes
3. Does it look like a burst tubing? ANSWER: Yes...looks just like frozen pipe bursting.
4. What is the current state of the panels? Frozen water in them? Are they drained? ANSWER: They were empty on 01/03/12...the water was shooting out of the holes.

When did the failure occur?

• Option 1 - It occurred on 1/3/2011 <<< NOT TRUE <<<
• Option 2 - The leak has been occurring for some time <<< TRUE
  • It appears this has been happening for a while and it just got cold enough to freeze.
• Key Questions

1. When was the last time the panels ran before 1/3/2011? ANSWER: 12/28/2011, about 1 week prior.
2. What is the tank level? If it's been leaking for a while, the tanks would be low. ANSWER: Tank alarm indicates normal level.
3. Any evidence of routine leaking up on the roof? Water...deposits showing steam of water? ANSWER: NO

What caused the panel failure?

• Potential reason for the leakages:
  • Potential Cause #1 - Panel manufacturing failure <<< NOT TRUE
    • Not likely since the panels had been running without leakages (need to be confirmed)
    • Two panels failed in the same spot at the same time
    • Panels were never tested for water flow before installation. <<<< Appears to be the issue.
  • Potential Cause #2 - Water froze in the copper tubes causing them to burst <<< TRUE
    • The fractures in the pipe are exactly the same as occurs when a pipe fails.
    • Note that the leaks in the two panels were at the same level above the bottom of the panel.
    • Note that the these were two panels later in the series...water is cooler.
  • Potential Cause #3 - Thermal shock in the copper tubes causing them to fail <<< NOT TRUE
    • Presume that panels are designed for drain-back system to withstand thermal shock of introducing cooler water back into panels when started up, but may need to know to what extent panels can tolerate shock (temperature difference and number of cycles of shock until fatigue) and make sure pump control set up within that tolerance level. ANSWER: This is not likely the answer as the panels are designed for this situation. Alan Rushforth uses these in Philadelphia all the time.
    • FOLLOW-UP QUESTION: What differential does Alan use to activate his system? How much different is the water temp introduced into his panels vs. the temp of the panels?
• Key Questions

1. What is the best way to determine the cause of the failure? Wait for warm weather and run water through the system. Currently looks like Friday is a good day.
2. What does Alan Rushforth have to say about this? ANSWER: Longer answer...but Alan feels it is in the complexity of our drainback design. He never has valves in his return. Straight up and straight down. His pump allows water to return through it. This could just be his bias....but there is a simplicity to his answer. He said his only failure was when the return pipe got bent and was in the water, instead of in the air above the water level in the tanks
3. What is the temperature of the pipe / tube at the bottom of the panel that fills first?
4. To what level of thermal shock are the panels designed to withstand (stagnation temperature tolerance with drain-back system water introduced at much lower temp)? Is pump differential temperature control set up to operate system within this tolerance level?

Why did the "cause of the failure" occur?

• IF it was caused by frozen pipes (pure conjecture at this point 1/4/12), why did they freeze?
  • Potential Reason #1 - The water froze because the panels were too cold on the initial filling <<< NOT TRUE
    • What temperature was the panel air at the time of filling? ANSWER: Approximately 194 deg F.
    • Could the water at 97F(Tank Temp) lose enough heat as it was going through the system to freeze? ANSWER: Not likely.
    • Could all the panels not be the same temp at the time of filling? ANSWER: Check for shading/shadows per previous observations by Joe for chimney impact. Otherwise appears that it would be uniform, but can be verified by thermal imaging possibly.
    • Could the temp gauge at the top of the panels be inaccurate? ANSWER: Could be. Was checked for calibration prior to being relocated but has not be rechecked since moved. Can be re-calibrated.
    • Could there be a big difference in the temp at the top of the panel and the temp at the bottom of the panel? ANSWER: Check with thermal imaging possibly.
    • Could the programming have an error in it and pump #9 came on too early or incorrectly? ANSWER: History trend does not appear to indicate this, unless there is potential that copper pipes were overheated prior to introducing water into them, causing shock?
  • Potential Reason #2 - Panels didn't drain properly this time or Panels never drained properly <<< TRUE
    • Could design of drainback have a flaw i.e. the panels don't fully drain?
      • If the leaks are actually 2/3 up the panel this seems unlikely.
      • We have had colder nights and we didn't have a problem...freezing
    • Could valve V1 for drainback not open correctly the last time the panels needed to drain? ANSWER: Not likely
    • Could the water in the panels have frozen before it fully drained? Yes! See Current Theory below.
    • NEW THEORY: Air lock in the system preventing the water from draining. Could there have been air lock in the system where it couldn't drain well. Besemer and Brennan ADDITIONAL COMMENT: Catey thought the same thing and wondered if the air vent might have failed, then researched and found that Rushforth's design does not include air vent. Is this normal for drain-back system? Do we have adequate air gap where tubes enter tanks?
    • CURRENT THEORY
      • We think the problem was they didn't drain fully because the solar panel return line was in the water. The design called for it to be above the water line to let air in the return line allowing it to drain.
      • The Labs used a camera probe (video borescope) to see that the return pipe was only 0.5 - 1.0 inch above the water line. Given that we lose some water (tiny tank leak) and we lost water during the panel leak, it is highly probable that in the near past that pipe was in the water. (Interestingly enough, this is the same way that Alan Rushforth had a failure.)
      • With the return line in the water, the water on the solar panel supply side drained to the point where the suction (vacuum) from above was as strong and the weight of the returning water (the head). This caused the panels to drain to about 2 feet from the bottom of the panel...which is where the panel tubing fractures occurred. As Joe says, the problem is as if you put your finger on the top of the straw and then lifted it out of your drink ... all the water stays in the straw...the vacuum holds it up.

How do we recover?

• Option #1 - Repair the panels
  • Manufacturer says that this would be difficult...they are no help. ANSWER: All panels were repaired and returned to their location on 1/4/2011...thanks Joe, Eduardo, Heladio, Jason and Expert (brazer).
  • ADDITIONAL MODIFICATION: Removed a small section of Tank Town "roof" to access solar panel return line where it enters top of tank, and cut a segment off at the bottom so that the tubing is now just inside the tank, approximately 6" above the current water line, instead of down into the water. Joe will continue to check that the return line is drained each night to verify that this has resolved the issue. We will be cognizant of this condition also any time we need to add water to the tanks.
• Option #2 - Replace the panels <<< FIXED already
  • Could do this first and then repair panels (Option #1) later.
  • Would need to make sure that cause is accurately determined and corrected.
  • Could only run panels when outside temp is above 20F (plus low winds)
• Option #3 - Wait until spring (consistently warmer weather) <<< FIXED already
  • Better for testing and working out all possible issues
• Option # 4 - Reconfigure from 10 panels to 8 panels. <<< FIXED already
  • Costly in labor and then need to change back.

Next Steps

1/4/12

• Check on availability of replacement panels and mounting materials (Joe)
• Get input from others...email, phone (Joe, Tom)
• Complete recording of what "we know" section on wiki
  • Research using data from Niagara System
  • Answer ... When did the failure occur?
  • Plan test for Friday with Labs
  • UPDATE:
    • Found third panel with hole.
    • Repaired all three panels. Took them down. Had Expert repair person braze the holes shut. Retested. Put back up.
    • Tested the panels for drainback and they worked. Using the system as designed.
    • NEW THEORY: Air lock in the system preventing the water from draining. Besemer and Brennan

1/5/2011

• Focus on the drainback.
  • Check status of Valve 1 and 2 Done. Verified valve 2 in open position. Removed motor. Removed motor on valve 1 and verified open positon. Replaced motor and started panel sequence, pump9 command 'on', (had power switch off), valve 1 closed (takes approx 1 minute to close. Took away panel call for heat, valve 1 opened fully in 10 seconds. JoeG
    • Check value status for Dec 29 - Jan 3
    • Check that they are installed correctly physically
  • Design error? Do a physical walkthrough.
    • Tested drainback and it works (1/4/2011)
  • Programming error. Find the possible scenario not tested.
  • Could unplug Valve 2 for the winter (Always open....no risk of closing)
    • Identify and test all "legs" of the logic ... including emergency...non-powered.
  • One time problem when Fish was reprogramming everything on 12/29/2011. Accident.
    • Talk to Fish about what he tried to do...ever turned on Pump #9?
    • Could the "default" (unpowered) setting for the valve mistakenly be closing the drainback.
• Send links to this page and GG HVAC page to Alan Rushforth.