Construction
I received my GEK lvl III on Monday afternoon (March 9th). Started welding that day. Finished welding and assembly by Saturday (March 14th) working evenings. I found it pretty easy and straight forward to assemble. The most difficult was the cyclone. Welding the inlet to the swirl burner was time consuming due to some mismatch (had to close up a 3/8th in. gap). Once again, the construction was not very difficult for a novice welder. I used some cheap engine paint to protect the outer surfaces from rust. To bad the paint didn't withstand the heat very well.
I installed the new tar shield and stainless steel hearth. The resulting nozzle height is higher than indicated in prior designs. I insulated with perlite purchased from a local nursery. After getting some help from BK on the proper wiring of the fan motor, I was ready to go.
First Run Set-up
My intended wood source is a local saw mill. They debark all of their lumber and chip their slabwood. I wanted to burn this slabwood. They are a softwood only mill, primarily hemlock and pine. I picked up 2 100 lb feedbags of nice, clean chips (mostly 1" square and about 3/16 " thick) with some fines mixed (sorry the picture is a little blurry). The only trouble with the chips is -- they are 36% moisture. Thanks for sending along a moisture meter with the GEK.
The next step was to load the hearth with charcoal and prepare it for lighting. Unfortunately, the bag of charcoal I purchased, though pure wood, was still not adequate for loading up the hearth; it had too much small, dusty charcoal pressed together, not the desired structure. I had some charcoal I had gathered from my woodstove, but I used all of that for the filter (still didn't have enough to fill it). I instead made a fire in the woodstove, allowed it to finish pyrolysis, and put the resulting hot coals directly into the hearth. This step greatly simplified the lighting process. I was generating gas right after I got the lid on and the fan running. It readily lit with a torch and was able to sustain the flame in the swirl burner without the continuous application of the torch.
First Run Results
Everything started out great with my first attempt to generate wood gas. I did not have the thermocouple in place at the base of the reactor -- as measured from the access port in the front of the reactor between the stainless steel gas tubes -- so I do not know exactly what was going on inside. The sun was too bright to see the flame color in the swirl burner. This did not last very long. The extremely wet woodchips very quickly became a "wet blanket" on my successes. The flame in the burner stopped sustaining itself; I needed to keep a lit torch in the base of the swirl burner. I put the thermocouple in place and measured a temperature that started at 460 C and kept falling. This location corresponds to measurement point #7 in Test Report #3 on the AllPowerLabs site.
The fan motor speed was set to medium, and then I turned it to low to hopefully increase the temperature. It only kept the temperature from falling as fast. The reactor kept burning, but the temperature dropped to below 300 C until all of the woodchips were pyrolized, then the temperature started to increase again. I took the lid off and looked inside the reactor. There was a small amount of bridging. I just needed to shake the reactor to get it to break up. I added more woodchips to see if there was any way to get the temperature up to where it needs to be but to no avail. The temperature dropped to a point where the flame in the burner could barely be sustained even with the torch.
I decided to cut up some scrap wood I had lieing around outside into small blocks with the hopes that it would be drier and a higher temperatures could be achieved. I was met with limited success. The wood blocks were at 20% moisture. They were about 1.5"x 3/4"x3/4". The temperature increased as desired, but still not to the levels suggested by Jim. I could only get it up to about 480 C, not 600-650C. I did find that once I got the temperature upto 425C, the flame in the swirl burner would sustain itself. At this point, the sun was not as bright and I could see that the flame was dominated by the orange flame and not blue. From what I understand this is the result of the gas being primarily CO and not H2. I suspect higher temperatures are needed to generate H2.
I got quite a bit of tar, everywhere. Any good suggestions on how to clean it up? I also had a significant amount of condensation, it kept flowing into the swirl burner and half filled the quart jar at the bottom of the cyclone. I added another thermocouple at the inlet to the cyclone. The highest temperatures I measured there was about 115C.
Second Run
The desire of my second run was to produce better gas (isn't that the intent of every experiment?). To prepare for this run, I spread out about 3 gallons of chips and dried them overnight to about 10-15% moisture. I agree with your assessment on the moisture meter readings Jim. I think it maxes out at about 36% moisture. I suspect my initial moisture content was greater than 35%. I lit the reactor in the same manner as in the first run; I stole some coals from my woodstove.
I dumped in all of my dry chips and within about 3 minutes, the reactor was generating gas of sufficient quality to sustain its own combustion. This cumbustion was dominated by an orange flame. I am assuming this is due to the lower temperatures at start-up in the reactor and from the need to pyrolize a significant amount of fuel. One interesting observations was the low temperatures coming out of the bottom of the reactor. I was only measuring temps on the order of 250-300C. I would attribute some of this to the reactor being cold, but as I would find out later, the temperature I measure at the bottom of the reactor has little to do with the quality of gas produced.
After running for about 10 minutes, the gas quality improved significantly. It had about equal amounts of blue and orange flame. I have noticed that the orange flame tends to not burn as quickly as the blue. It typically leaps out of the swirl burner, where the blue flame burns very quickly within the swirl burner. I was getting very little condensation burning these drier wood chips. Everything was operating great so far in my opinion.
The first problem then reared its head, bridging. I think bridging can be identified by an increase in the gas temperature at the bottom of the reactor and a decrease in the amount of orange flame in the swirl burner. Shaking the reactor didn't seem to be suffucient, I had to poke it down. To access the access the bridge required removing the lid from the bulk fuel barrel. I was a bit surprised at the amount of smoke filling the barrel. I pushed the rest of my dry chips around the top of the hearth and decided to add 2 buckets of my wet chips. My hope was that since things were running so well, the system may be robust enough at this point to handle the wet fuel. I was wrong.
The whole time, while burning the dry chips, the reactor exit temperature never got much over 400C even though it appeared to be generating lots of good gas. Once the reactor started to burn the wet chips, the reactor gas temperature did not decrease, but the quality of the gas did become quite poor. After about 15 minutes, it would no longer support it's own flame. I needed to put the torch back in to clean up the exhaust. I was also getting significant amounts of condensation.
I knew the water content in the chips was being a drag on the overall process. I thoguht that maybe if I could increase the temperature of the exhaust gas, I might be able to get through the wet chips. I started with trying a few differnt fans speeds. My 12v supply is a battery charger that permits me to set the voltage. The low setting tends to produce about 8v under load and the med setting about 11v under load. This provides me with a "6-speed" fan instead of just 3. It seems the best speed for me to operatre was the med fan setting and the low charger setting. I was finally able to eek out about 590C for a few minutes fairly early during the burning of the wet chips. This higher temperature still did not generate a gas capable of sustaining it's own cumbustion. The gas being combusted was well below 100C and as a result contained a great deal of water vapor. There was some CO in the gas, but the water vapor was too much of a thermal load on the combustion process for sustained burning. Looking into the swirl burner, I could see a nice orange flame once the blue propane flame impinged on the gas stream. I could barely turn the torch high enough to burn up all of the gas, yuck.
I had a difficult time keeping the temperature up; it kept dropping. Sometimes this was due to bridging. I repeatedly had to push the chips into the reactor. I noticed that once I did this, the temperature would increase. I found this a bit odd since I would have expected the temperature to decrease becuase of the increased moisture being drawn into the reaction from the wet chips as they were dried before pyrolysis. I determined later that it was due to my opening the top of the reactor. If I cracked the lid on the top of the draw a little, my temps would increase, but at the expense of gas quality. I think I was providing enough air in the reaction to get nearly complete combustion within the reactor leaving only CO2 and H2O (as well as lots of smoke/tar). I found I could not "burn" the exhaust clean when the lid was open.
Third Run
I worked on drying my chips as best I could this week and got some more. Still only have it down to about 25%-30%. Much better than before, but, as I found out, still too wet for my GEK. I can get to to light easily and generate great gas with the dry charcoal, but things go downhill fast once the wet chips get going. I was measuring low temperatures through the measurement port at the bottom of the reactor. Early, when producing good, self sustaining cumbustion gas, things were quite cold, but I suspect that is only because the reactor is cold. Later, during the wet chip burning phase, the maximum temperature I could read was about 420C, still not nearly what it should be. I must use a torch all the time in the swirl burner to keep the smoke down when burnign the wet chips. My run was cut short when my gfi circuit cut out on my battery charger I use to run my blower motor. The chips I have keep bridging, even with vigorous shaking of the reactor. I think they may be a little too long and flat to flow effectively.
Fourth Run
My chips are finally down to the 10-15% range. I haven't used the microwave test to confirm the accuracy of the moisture meter. I emptied out my previously unburnt chips, lit the torch in the reactor, started the blower, dumped some diesel fuel in the reactor and lit it. Started making gas in a minute or two. I dumped my dry chips (small cheer) in the fuel hopper. The swirl burner kept going without needing to use the torch to keep it lit.
I did not have nearly as much trouble with bridging when using dry chips. Things kept running quite nicely for over an hour until I allowed it to run out of fuel. During the run I opened up the top and added more chips 3 times to the hopper. I would shake the ash grate every 10 minutes or so. The temperature from the front port would get up to about 470C. It was a very satisfying run until -- the reactor "puffed", banging the lid quite loudly and scaring most people nearby. It is a good thing the lid is attached with springs to release the pressure or it may have gotten ugly. In hindsight, I now realize that I observed smaller versions of this puffing at times when replacing the solid end cap with one that has a small hole in it to permit the insertion of the temperature probe.
When I open the lid during a run, the fuel hopper is full of very dense smoke; smoke which I presume has a high concentration of CO. I suspect the cause of the "rapid combusion" was the fuel level dropped down to below the air inlets and enough oxygen was permitted to fill the fuel hopper and mix with the CO until a nice mixture existed to be ignited by the hot coals in the bottom of the reactor.
It is now time to clean out the ashes in the bottom of the reactor. I can hardly move the ash grate anymore.
Tear Down
I decided to tear the system down tonight. I had noticed toward the end of the last run that my pressure drop to the measurement port at the bottom of the reactor had increased from the typical 1/4" to 1", a sizeable increase. I think I now know why. Internal tar buildup. Take a look at the following 2 pictures and decide for yourself.
It looks like I have made a royal mess of things. Most of this tar is the hard kind and is not loose. I will now need to clean up from my mistakes. I assume most of this tar is the result of burning wet, wet chips. One last picture I took was of the hearth. Notice that there is no tar build-up near the hearth, but the is some above it on the walls of the reactor.
Run #5
I got everything cleaned up and am ready to run again. I am no longer using the fuel storage hopper during my experimentation stage. I dumped some of my chips in and lit the reactor. I then dumped may recently created "blocky" fuel on top of the small amount of wood chips. I let it run, shaking and stirring away. I could not get the temperature up to desired operating temperaures. I drilled a small hole into the 1.5 in plug on the reactor lid to allow insertion of the 24" temperature probe and took some temp measurements through the reactor. I had a hard time sustaining any combustion with my new fuel. The moisture content is below 15%. The issue now is its size. It is too big.
My "blocky" fuel consists of branches from 3/4 to 1 1/2 in diam cut into 3/4 to 1" lengths. In a reactor the size of the GEK, this size of fuel will bridge even more than the wood chips I had previously used. The bridging is also more difficult too break down. I think it also has too large of a void fraction, and is too easy to draw too much air through it relative to the reacting curface area. After trying to make it work for 40 minutes, I stopped the run by turning off the fan and closing all of the ports. I now know another thing not to do. I will return to my original wood chips.
Run #6
I removed the wood blocks and went to just chips in the afternoon. It burns better on just chips than on chunk fuel. I still run into issues with bridging. I think the issue may be my chips are too big for the throat diameter. I kept measuring the temperature throughout the reactor to get a better understanding of where pyrolisis and reduction was occuring. Due to the bridging issue, my pyrolisis kept dropping inside of the hearth rather than 3 in above it in front of the nozzles. I then would remove the plug and smash the chips back down. The pyrolisis zone would then slowly move back up to the desired position and then drop back into the hearth as the chips above would bridge again.
At this point I am not sure if it would be better to put in a stirring mechanism or to increase the throat diameter.
Tar Fence Modification
Upon the recommendation of Jim, I modified the tar fence by cutting down from the top to the middle of each hole. I then cut across from each hole, leaving about 1.5 in of metal. I bent these "petals" out to match the inner diameter of the reactor wall. The final result, installed in the reactor is shown below.
I am still having some troubles with running the GEK. I think some of my issues may be with the fuel I am using. The chips I have, were wet. I dried them. They also contain a significant amount of fines. This was really brought to my attention when I cleaned out my unburnt chips and found the following in the reactor. I was surprised about the small size of the char. I took my existing chips and screend out the fines using about a 3/8 in mesh. This ended up removing nearly half of the weight of the chips. There was much more fine material in the chips than I thought there was.
I lit the GEK with the screened chips and had much more success. It still seems that I am having issues with bridging.
My fix to bridging is found here: Jay Martin: Bridging Fix
What's Next
I am now considering making a new hearth. I want to make an hour glass shape, but I also want to make it out of cast refractory instead of stainless steel. I am also considering air passages in the crucible and setting it on the 1/2 in ells in the bottom of the reactor, doing away entirely with the tubes and nozzles. It should be possible to get a decent seal between the refractory and the 1/2 in ell. Rutland makes a decent portland based refractory that is good for over 2000F. Trying to decide if it should be cast as a solid block or as multiple pieces, possibly 5. Hopefully this wil help get the temps up higher.