All right, the next piece needed for the combustor was the flame tube (AKA combustion liner or flame holder). The flame tube dimensions were calculated earlier using the GR-1 Combustor Formula. According to the formula, I would need a flame tube with a length of 28.875 cm (11.37”) and an ID of 97.5 mm (3.84”). The 4” OD X 0.080” wall 304 stainless tubing that I had purchased earlier was a perfect match for the ID dimension.
I started out by forming the end of the 4” pipe to match the inside contour of the tapered neck of the combustion chamber. This will help the flame tube fit farther into the neck of the combustor thus allowing it to have a greater overall length. The matching contours will also help create a uniform “bypass” area in where the some of combustor’s air charge is allowed to pass around the end of the flame tube. Before I could move on and drill the flame tube holes I would need to cut the flame tube down to it’s final length. In order to do this I would need to position the flame tube in the combustor as it will mount so I can measure how far it will extend into the tapered neck of the combustion chamber. The bypass area would need to be set up correctly before I would know where the flame tube would actually rest. This is because the bypass area is adjusted by moving the flame tube in and out of the tapered combustion chamber’s end. The Combustor Formula suggested a bypass area of 27 sq. cm which can be measured as an 8 mm border around the end of the 4” flame tube. To simulate this dimension I stuffed rags between the combustor and flame tube to hold the tube in the center of the combustor. I then carefully adjusted the distance of the flame tube to achieve an 8 mm (average) border around the flame tube. This allowed me to measure where the flame tube would be located in relation the the combustor end plate. I marked where the tube lined up with the end plate flange and moved on. When I designed the GR-7’s combustor I had made allowances for the length of the injector base before it was built. This is because the injector base will displace volume from the flame tube and it needed to be compensated for. The flame tube was designed to slip over the injector base just far enough to meet the mounting holes prepared for it in the base. This would be about 1” from where the flame tube will meet the injector base top plate. Considering these measurements I calculated an overall flame tube length of 12.5 inches which left about 11.5 inches of useable flame tube volume. This is very close to the value I estimated when I designed the combustion chamber taper a while back. Needless to say I was fairly happy that my math did not let me down :0) I could now cut the flame tube and prepare it for the flame tube hole layout. Once again, my $150 band saw was earning it’s keep.
Once the flame tube was cut I proceeded to lay out where the mounting holes would need to be drilled. After some thought I decided to cut slots instead of holes for the flame tube mounting hardware. This will allow me to fine tune the bypass area later on. I used my Dremel rotary tool to cut the slots out of the sides of the flame tube. Notice that the air plenum port is partially covered by the flame tube (below). This will help tune the evap tube combustion air flow. The flame tube can be moved back and forth to reduce or increase evap tube combustion air pressure/flow. Sounds like a good plan doesn’t it?
The combustor formula suggested a flame tube hole area of about 30 sq. cm. distributed among 90 flame tube holes. It was my job to decide how to layout these holes and what size they would be drilled. On my GR-1 engine I learned that the holes drilled in the combustion liner may need to be different sizes to allow for a uniform burn within the combustion chamber. Most likely the holes on the end of the flame tube would need to be a little larger to help cool and control the tail of the flame. To allow for these larger holes I divided my proposed hole pattern into two groups. The primary zone and the secondary zone. The primary zone will consists of 84 holes @ 7/32” diameter and the secondary zone will have 28 holes @ 1/4” diameter. The 1/4” holes will serve as the secondary zone which will hopefully help produce a cool controlled flame. The total number of holes in the layout comes to 112 with a total area of 29.282 sq. cm. This is 22 holes more than the suggested number but was required to adjust for the drill sizes I wanted to use. To space out the holes I followed a “diamond” pattern which is the same I used on the GR-1 engine. Using a marker I layed out the 16 rows of holes on the flame tube, seven holes in each row.
Using my mill I drilled 1/8” pilot holes at all of the hole locations, all 112 of them :0P I then drilled out the final hole sizes and deburred them with a countersink tool. To deburr the inside of the flame tube I used a die grinder.
I could now install the flame tube to the injector base and assemble the combustor for testing. Notice the “bypass” area around the flame tube below. This is, in my opinion what makes the tapered neck combustor a success. Many combustor designs do not have a bypass and only rely on flame tube holes for cooling the combustion gasses. These holes can produce turbulent mixing where the bypass system creates a relatively stable “envelope” of cool air around the hot gasses entering the turbine volute. This envelope gives the gasses a bit more time to stabilize and cool before they touch the turbine wheel.
Finally the combustor is ready for some bench testing. I built a test fixture to hold the GR-7’s combustor and mounted it to the side of the engine frame. I installed a couple of 1/8” NPT needle valves for fuel control. One on the evap nozzle and one on the pilot flame nozzle. Both fuel valves were piped to the fuel pump and tested for flow with diesel fuel.
To simulate the inflow of combustion air I used a 1 HP dust collector blower. 4” aluminum dryer vent hose was used to couple the dust collector blower to the inlet combustor pipe.
With 5 gallons of #2 diesel and my trusty ABC rated fire extinguisher I started testing the new GR-7 combustor. I set my fuel pressure regulator (integral to the Webster fuel pump) to 80 PSI and cracked the pilot flame valve on the back of the combustor. I ignited the pilot flame with a propane torch and turned on the blower motor. Ahhhh the distinct smell of diesel burning, a close relative to Jet-A :0) The pilot flame nozzle was producing a nice “full” flame. It seemed to be covering the evap tube assembly quite well.
I let the combustor acclimate a bit and then slowly cracked open the evap tube injector. The evap tubes started to work and the combustor assumed the familiar howling and growling noises. I shut off the pilot flame nozzle and proceeded to test the ability of the evap tubes to process more fuel. The fuel valve was very sensitive so I had to be careful not to flame throw my truck at the end of my driveway :0)
I immediately noticed that the flame was being pushed away from the flame tube wall that is on the opposing side of the inlet diffuser. If I can recall, the same thing happened on my GR-1 engine which is a side entry combustor as well. My theory is that when the inlet air flows around the flame tube it converges together again on the back side of the flame tube causing a higher static pressure. This higher static pressure causes the flame to be forced away from the back side of the flame tube wall causing a “cold spot”. To solve this problem I decided to narrow a few holes on that side of the flame tube. I removed the flame tube and TIG welded the holes (7 in total) closed and then re-drilled them with an 1/8” drill bit. I remounted the flame tube and resumed testing. The flame was now fairly centered in the flame tube which solved the cold spot problem. I could now test the ability of the combustor to burn fuel at higher rates. I opened the valve a bit more and the flame extended about 3’ out of the combustor. This had me a little worried as the flame will have to be contained within the combustor or a turbine meltdown will undoubtedly ensue. Although the flame may act much differently when it is under the higher atmospheric pressure of normal engine operation. Unfortunately there is no easy way to simulate normal engine operating conditions without actually running the turbine on the combustor.
At this point all seemed well with the exception of my flame length concerns. The flame was burning evenly along the flame tube and was fairly responsive to fuel flow settings. I could “gun” the fuel valve and see the flame respond without much hesitation. This is important as some evap systems can cause a lag in burner response which in turn can cause throttle delay. Satisfied with most of the test results I shut down the test rig. I decided to tear down the combustor to inspect the flame tube and evaporator. Once apart, I looked over the flame tube for any erosion or pitting and did not see any. Although the evaporator tubes looked like they were subject to high heat towards the end of the “branch” tubes. This may be caused by the tubes being too long which can prevent them from being sufficiently cooled by the fuel/air flow through them. If the problem persists I will have to shorten the tubes.
Before I went any further with testing I wanted to install a spark plug system that could be used to ignite the pilot nozzle. To do this I planned on using a modified NGK B8ES spark plug mounted in the side of the combustor. The spark plug will be mounted in a “pedestal” that will allow it to reach the side of the flame tube. The center electrode will slightly protrude into the flame tube allowing it to spark to the tube itself. I was not sure if the idea would work but I was willing to give it a go. I started out by rounding up some stainless scraps for the spark plug pedestal. I had a 3/8” thick bar scrap that could be used as the threaded hub and a 1-1/2” stainless tube for the pedestal neck. I had purchased a special drill bit and a M14 X 1.25 tap specifically to build this piece, $32 later :0{
I drilled the 3/8” piece with a 1/4” pilot drill and then followed it with a 12.8 mm drill bit in preparation for the tap. I then carefully tapped out the hole using plenty of cutting oil.
I was pleasantly surprised at how well the American made tap cut through the stainless compared to the “imported” taps I have used before. I guess you get what you pay for ;0) I needed to turn the threaded piece on the lathe to prepare it for mounting. Using a slow head speed and a lot of cutting oil I was able to shape the spark plug hub. I used the spark plug itself as a spindle to hold the hub to the lathe chuck.
I carefully welded the hub to the pedestal neck. Notice the “stepped” edge of the spark plug hub above. This is to reduce the chance of “burn through” during the welding process.
I located a spot for the spark plug on the combustor that would be easily accessible from outside the engine frame. I then hole-sawed a hole for the pedestal to fit into.
The B8ES plug is a common 14 mm motorcycle plug which features extra long threads. This makes the B8ES a great plug to use for this application as it can be modified on a lathe to expose the center electrode. By removing some of the threads you can “raise” the center electrode and allow it to have a larger spark gap. A large spark gap will have a better chance of igniting the diesel fuel spray. I proceeded to use my lathe to remove about a 1/4” from the treads of the plug.
The pedestal was ready for installation so I installed the flame tube/end plate assembly into the combustor. I had to figure out how deep the pedestal needed to recess into the combustor wall. I wanted the electrode to protrude into the flame tube about 3/32” so it would be in the fuel spray pattern. Once I figured out the depth I needed I welded the pedestal in place. I then cut off the remainder of the pedestal neck and ground the weld smooth.
I now had to drill the hole in the flame tube that would accommodate the electrode. I marked where the plug would penetrate the flame tube and drilled it with a 7/16” drill bit. This hole should be large enough to allow cooling air to envelope the electrode and keep it from overheating. Unfortunately this hole may need to be modified or widened if the flame tube requires any back and forth adjustment.
I needed to wire up the ignition coil to test the spark plug so I dug up my old motorized points setup. This is the same setup I used on the GR-1 engine. I wired it into the coil and connected the plug with a 7 mm silicone plug wire.
I fired up the coil and was delighted to see a robust spark at the electrode. A very loud buzz echoed in the combustor and I knew I had a good chance of getting reliable ignition from this system.
I rolled out the test rig and set up a video camera to capture the first ignition test. With the camera rolling I opened the pilot flame valve and recorded these images (in order).
Repetitive tests were done to measure the reliability of the ignition system and all were successful. This was very encouraging as this system will eliminate the need for a propane starting system. I am now in a good position to finish up the combustor and bolt it to the turbine for a test run. It wont be much longer folks so check back for another update!!!
Till then...........
Don Giandomenico Chief Engineer of DRG Engineering, Gas Turbine Division :-)
