The BPM5 engine is a 5 kN bi-propellant engine which is developed as an experimental engine prior to undertaking the design of the much larger 100 kN BPM100 engine. The BPM5 engine will serve as a small scale test bed for various technologies before they are implemented into the BPM100 design. It runs on liquid oxygen (LOX) and an ethanol/water mixture.
The BPM5 engine consists of three main components
- A regeneratively cooled combustion chamber
- An injector
- A LOX dome
The top of the engine, the LOX dome, is milled in aluminum and serves as the LOX inlet and mounting point of the assembly to a thrust structure in a rocket or for mounting onto a test cell. It is fitted with a 3/4″ thread LOX inlet which distributes the LOX through seven smaller ports to the injector. The fuel is fed through a pipe that connects to a manifold at the nozzle end of the engine. From here the fuel enters into a 2 mm thick cooling channel that runs the entire length and circumference of the combustion chamber. The fuel enters the injector radially and is then injected into the combustion chamber.
The injector is milled in aluminum and has 120 0.9 mm LOX injection holes and 136 0.8 mm fuel injection holes. The injection channels are paired in a like-on-like configuration with a fuel rich circumference and a LOX rich core. The engine has a nominal operating pressure of 15 bar at which it will generate 5000 N of thrust and consume 1.46 kg LOX per second and 1.12 kg fuel per second.
The design phase started in late 2014 and it has been a long way to get to a working engine. In the video below you can get a peek at the initial planning and design of the BPM5 and the test stand.
Machining the parts
Most of the components are milled in house at CS on either 3 axis milling machines or lathes. To produce the injector we however had to team up with Protobuild who generously fabricated several units for us on their 5 axis CNC milling machine. The tricky part here is of course the 256 injection channels as most of them are machine at a 20 degree angle to create the like-on-like impinging flow pattern.
The engine tubes, both the outer and inner parts, are made by metal spinning. Metal spinning is done by machining a solid steel shape that has the contour of the engine and then press or spin a cylinder onto that solid shape. We got a helping hand from Olsen Metal on this process, see how it was done in the video below.
Putting it all together
The cavity betweeen the inner and outer liner acts like a cooling sleeve in the BPM-5 rocket engine, where the fuel is used as s source of immense cooling, preventing inner liner from melting due to the heat of the combustion process. If not for this cooling, the inner liner would melt and burn through in seconds, as the temperature of the combustion process is twice the melting point of the inner liner. Uniform cooling of the inner liner demands a uniform fuel flow, which in turn can only be realised with cooling channels of identical height. The practical problem is that several materials in the BPM-5 engine really wants to bend, warp, expand and elongate due to sudden temperature changes in the various materials of several hundred degrees.
In order to create this uniform cooling channel height of just 2mm, we fused a number lengthwise equally spaced copper wires to the inner liner in a soldering process. These highly heat conductive copper wires prevent each cooling cavity from collapsing due to heat and pressure, thus ensuring uniform cooling flow. Keeping these copper strands in place, equally spaced, proved to be somewhat of a challenge.
For a deeper insight into the silver soldering experiments check out the in depth 50 minute video below.
Also, you can watch Jonas tell more details on assembly of the BPM5 engine below.
During assembly of the first BPM5 engine we ran into a few issues. The angle of the diverging part of the nozzle did not match between the combustion chamber tube and the cooling jacket tube. Thus, we cut open the outer tube and used a bit of brute force to get it to line up with the combustion chamber shape. The result is an engine with a nozzle that is slightly shorter than intended. Also, the nozzle end has a rather distinct look from being fitted with hammer and anvil. This has earned BPM5-001 its nickname “Franken5” after Mary Shelleys “Frankenstein”.
Since the engine is made from ordinary carbon steel it will corrode and rust easily unless we do something to prevent it. It is especially important to prevent rust forming in the cooling channel as there is no easy way to get rid of it again and small rust particles might be able to clog holes in the injector. To prevent this we nickel plate the entire engine with a layer of about 10 µm of nickel. Nickel plating is done at the Technical University of Denmark where we borrowed a electrochemistry lab for an afternoon.
For details on the nickel plating process please take a look at the video below.
After nickel plating the fuel manifold, the combustion chamber, the cooling jacket and the flange are all welded together. Jacob, one of our skilled welders assembled and welded “Franken5”, check out the video below to see how it went.
Before the engine is test fired it is important to characterize the flow through the injector. It is necessary to measure the propellant flow rate as a function of feed pressure to hit the correct operating point and oxidizer to fuel ratio. The flow tests are performed with water instead of ethanol and LOX,
For this purpose we build a dedicated setup where we reused the TM65 LOX-tank as a big water tank where we can safely run tests at up to 20 bar if necessary. Usually it is only necessary to map the flow rate at pressures between 3 and 5 bar though. The LOX and fuel injection channels are characterized separately as seen in the GIF-animation below showing the first 30 miliseconds after the flow starts.
Get more details on the injector flow test from Jonas in the two videos below.
Final pressure and water flow test
Prior to performing the first static test the engine is subjected to a pressure test. When the main valves are opened the cooling jacket will be subjected to the full pressure of the fuel feed system, this is around 20-21 bar in most cases. Thus, the engine must be able to withstand a static pressure in the cooling channel of 21 bar.
At the same time we do a final water flow test of the assembled system to map out pressure drops in the propellant feed system.
In order to subject the engine to static tests a new test stand was build. In the videos below you can see how the propellant tanks are made and how the test stand is constructed.
On previous engines we have designed and tested we have used a small set of valves to run the engine in pre stage. In pre stage the engine runs on a low propellant flow for a short period simply to make sure the engine is lit before turning the larger main valves and go to main stage burn. With the BPM5 we are doing this a bit different in order to save mass on the Nexø rockets and only want to have a single valve for each propellant component. Thus we have implemented Vari-V valves sponsored by Gosco Valves. This is a special type of valve where the ball itself is cut into a V rather than a circular hole, this allows us to establish a small flow with a single valve.
In order to utilize the fine control the Vari-V valves offer we must of course be able to control them accurately. This is done with a couple of servos from JVL. For the fuel side we use a MAC050 and for the LOX side we use a MAC095 since it has a bit more torque. These servos are extremely precise and really allow us to fine tune the pre stage conditions. JVL were very nice to us an offered us a great price on these babies!
First round of static tests
May 10th 2015 Franken5 was finally subjected to its first three static test runs. The first run was a short 6-7 second burn done in order to validate the operation of the engine but it was not long enough to reach thermal equilibrium. Afterwards the engine was disassembled and carefully inspected for any signs of problems. No problems were observed and a second and significantly longer burn was carried out. Again a careful inspection revealed that the engine was still in great shape. The datalogger crashed during the second burn due to the rather incredible sound pressure generated and thus we had to carry out a the second burn once again and thus a total of three burns were completed.
The three burns are all shown in this technical video.
For more details on the May 10th static test watch the video below where Jacob and Jonas explains more about how the test day went.
At this point we are still processing the data collected, we will update with more information once it becomes available.
Second round of static tests
The second round of static tests where conducted on May 31st 2015. The objectives of the day where to
- Carry out the first hot fire with injector-002
- Verify operating point with the new injector
- Thermal test of jet vanes
- First test of automatic pressurization system
Injector-002 differs from the first injector in the size of the LOX-channels as they have been increased from a 0.8 mm diameter to 0.9 mm. Thus it is necessary to verify the operating point with the new injector aiming for a combustion chamber pressure of 15 bar.
For the flight of the Nexø I rocket we will utilize jet vanes submerged in the exhaust for thrust vector control and we must find a suitable material to make the jet vanes from. Several missile systems use graphite for this purpose and thus we also tested a graphite jet vane. Since we have previously used copper jet vanes on the Sapphire rocket, we also tested a copper jet vane.
The test stand had been upgraded with a new automatic pressurization system for this test. The system consists of a 24 liter high pressure tank that can be filled with helium up to 200 bar. This is connected to a PWK06020W proportional valve from HYDAC, generously sponsored by HYDAC. This valve is actually designed to be used in hydraulic systems but we use it to pressurize our propellant tanks. The system proved to work excellent, the HYDAC valve and the control loop we had programmed hit the target pressure of 20.6 bar on the fuel tank spot on. For the next test round it will be fully implemented to the LOX tank as well. With this in place we expect to be able to pressurize the tanks in as little as 30 seconds. And most importantly, it is done automatically without human intervention, this is a must for the upcoming launch of Nexø I where the pressurization prior to flight has to happen immediately before lift off from Sputnik.
During the second test day we ended up performing four hot fires although we had errors on two of them.
The first run was intended to verify the operating point with the new injector and to subject the jet vanes to a relatively short burn of 7 seconds to investigate their thermal response. Sadly, we had a misoperation on the main valves which such that they never went to their fully open position after pre stage. Thus the chamber pressure peaked at around 8 bar since the valves where stuck at an unknown opening angle. It later turned out to be due to a bad wire connection.
Although the engine was only running at about half of its intended performance due to the valve misoperation the thermal response from the copper jet vane was still quite impressive… It quite simply started to burn away in a spectacular green flame.
Watch the full burn below and notice how the graphite vanes glows while the copper vane partially burns away. This is exactly why we need to test these things. Copper is now ruled out as an option and we will continue to experiment with graphite.
As the valves had gotten stuck on the first burn we repeated it after troubleshooting the issue. This second time around the test stand and engine performed excellently with a nominal burn. The combustion chamber pressure peaked at 14.5 bar, just 0.5 bar below the target operating point.
During this burn we measured the temperature of the graphite vane. It was measured in three locations and the temperature peaked at 750C with a maximum gradient during main stage of 180C/s. This looks quite promising as a jet vane candidate for the Nexø I rocket but it will take more experimentation in the coming weeks.
For the third burn of this test day we tried to shorten the pre stage time by implementing an automatic pre stage detection. So far the pre stage to main stage transition had been running on a timer, for this test the Engine Control Unit was programmed to switch to main stage when a pressure of 2.3 bar had been sustained for more than 1000 ms. For some unknown reason this failed and the engine shut down without switching to main stage even though post data analysis show that the criteria of 2.3 bar for more than 1000 ms was met. We are currently investigating this issue.
For the fourth burn we increased the feed pressure in the propellant tanks in order to hit the desired operating point of 15 bar chamber pressure. And we hit it! The engine peaked at 15.2 bar and 5.4 kN, very satisfactory indeed and a great achievement!
The graphite vane was still mounted in front of the engine as it had been throughout all the day. In total it lost about 10% of its mass, which is quite satisfactory and it holds great promise that this particular graphite can sustain a full duration burn of 50 seconds required during the launch of Nexø I. During upcoming tests we will investigate this much more.