Injector Project
Harbard Liquid Pintle Injector Assembly
Project Harbard ERFSEDS
Project Harbard is a undergraduate level project at Embry-Riddle Aeronautical University (where I attended undergraduate studies). The project is funded under one of the school's rocketry clubs, Embry-Riddle Future Space Explorers and Developers Society (ERFSEDS). You can learn more about the club at their website erfsedsrocketry.org. Project Harbard is focused on developing a flight capable liquid propelled rocket. The engine will run on liquid oxygen/rp-1 and produce 1000 pounds of thrust.
My role on this project was to be the lead designer of the engines injector. For this task a non-movable pintle injector type was selected. Details of the design are listed below but will not mention any specific numbers or dimensions. Additionally, in the MATLAB section of the website will provide a redacted version of the code. The redaction may not be necessary, but there is a sensitivity that surrounds the design of rocket engine injectors and I will always chose to err on the side of caution.
Design Process
This is a flow chart I made to describe the design process I used to develop the injector for this small rocket engine
The injector design process had many steps and a lot of research went into its development. The first step had to do with getting enough background knowledge in order to begin the design. I had never worked on an injector so I had to research the types of injector designs out there and weigh the pros and cons of each design. It was eventually determined that a non moveable post pintle was to be the injector used as it was a single element design and had a much lower manufacturing complexity. As we were ordering custom machined parts, keeping the parts cheap and easily modifiable was very important. With these things in mind I then had to begin putting math behind the ideas.
I started out in MATLAB. I began writing in every formula that I found from the many papers I read. There were a lot of numbers that I had to "guess" on and would refine through iterative test and adjustment. Eventually I began to get reasonable numbers from the output and decided to move on to the next step.
CAD models like the one shown at the top of the page were created. Over time the design evolved and changed. Part of what allowed us to make these design decisions was by creating a crude injector test stand out of our shower, some pressure dials, and a flowrate sensor. Even though this was a simple and "resourceful" test stand design, the shower water came out at 40 psi and about 2 GPM. This was decided to be decent enough to look at flow patterns and estimating a discharge coefficient at full operating pressures.
From these 3D printed parts I eventually landed on a design that I were able to put into ANSYS Fluent. I am not yet very confident in my CFD skills and am working on getting better at it. However, for this incompressible low speed flow, I was pleased to see that my simulation came out similar to what was found on the test stand.
Design Iterations
For the most part, the top part of the assembly stayed the same. Small adjustments were made on the sizing of the manifolds and a little bit of changes were made on the plate. The part that changed the most was the tip as this where the direction and interaction of the flow was most influenced. In order to keep manufacturing and design cost down it was decided to make the tip interchangeable. That way when the final aluminum piece was ordered, if a mistake was found it would not cost a large amount of money to replace all of the hardware.
Tip Iteration 1
The first design called for a 45 degree tip angle. This was to try and narrow the spray angle of the mixing propellants. The reason for this design decision had more to do with a lack of knowledge of acceptable spray angles than it did with trying hit a specific number. Because this was not a moveable post injector, I had to get creative with how the geometry accomplished these angles. I figured it was more important to experiment with these spray angles and be able to have that flexibility than it was to just assume a 90 degree angle from the beginning.
The hole size was determined through the oxidizer orifice area. For the parameters of this engine and from a manufacturing perspective it made sense to not go smaller than a 2mm diameter hole. This meant that I could not achieve having oxygen flow from all 360 degrees of the post. It is possible if a unusual or difficult manufacturing process was used. For this iteration, a larger hole size was used.
This tip was 3D printed and mounted on the test stand. It was found that in the gaps between the post jets the annular sheet would grip to the tip surface, curve to 45 degree incline, and condense in the center on the hex. Although the adhesive properties are different from water to liquid oxygen, I changed the design to better prevent the possibility of a hard start.
Tip Iteration 2
For this iteration, the hole sizes were lowered to the minimum 2mm, more holes were added, the flow was made to come out at a right angle, and the torqueing method was changed from a hex wrench to a standard wrench size.
When this was put on the test stand the issue of the annular sheet slipping past the post jets was lessened to a large degree. However, due to the differences in geometry across the surface along the tip some of the annular sheet would still grip and bend differently relative to other parts of the flow.
This was felt to be a step towards a better direction and a 90 degree jet angle was found to be acceptable with a spray angle of around 75 degrees from the center line.
It was decided to attempt a different design in order to see if the issues can be improved.
Tip Iteration 3
The only major thing being tested with design is the two layers of post holes. Another detail is a slight change in the post diameter. This was shrunk in order to make the annular gap a little larger as there was problems with keeping the parts concentric as well as meeting such a high tolerance.
As the hole size remained the same and the circumference of the post only changed slightly, there is still spaces between each jet. The only purpose of trying this was to see how the resulting sheets changed.
It was found that a strange periodic pattern of sheets was formed. It did not seem to be better than having the holes in a straight line and I was unsure of how the new pattern would effect mixing efficiency and combustion stability. Therefore, I stuck with the in line design.
Final Tip Iteration
This was the final design iteration. There are 14 2mm holes around the tip and a hexagon of a standard size as the torqueing method.
Now that the tip was axially symmetric, the flow was coming out even and looking good.
To address the annular sheet that slips past the jets. It is uncertain what problems this may cause until more cold flow test are preformed, with water and with propellant, and a hot fire test. However, I believe it to not be a huge issue during steady state operation. Some designs for similar injectors had small holes in the tip of the pintle such that there was fuel rich combustion near the tip. This kept the tip relatively cool. Because the annular sheet will be fuel, the excess should vaporize and cause a fuel rich combustion near the tip.
There is a concern about avoiding a hard start and not allowing excess fuel to pile up inside the combustion chamber during the startup. This could possibly be helped by sequencing the startup valves precisely enough, as well as the fact that the ignition source will most likely be a solid propellant rocket motor. If the chamber is oriented properly and the solid rocket motor is positioned correctly, the chamber could be filled with hot enough gases to prevent a hard start, or at lease have the excess fall through the throat of the engine. These are only initial ideas as the design continues to evolve and more testing is preformed to ensure the safest and most thought through launch possible.
3D Printed Injectors and CFD
Below are a few images from the test stand along with a CFD simulation of the same flow. All photos are using the final tip design and are simulated with liquid water at the same flow rates and pressures.
Here is a side view of the CFD done on the injector.
It shows the interaction of the annular sheet and the 90 degree jets. It shows the part of the annular sheet that misses the jets as well as the predicted spray angle.
The image was also altered slightly to make the flow easier to see but from the results it was shown to match the real thing.
This is a close up of the tip. It is hard to see but there is an annular sheet interacting with the jets. If this was not the case there would be discrete jets coming out at right angles.
This is an image from a little further away. It can be seen that there is a sheet forming, but there are definite areas where the jet is mostly deflected downward.
A bottom view of the injector. This was taken to compare with the images outputted by the CFD. If you look carefully you can see my bare foot as I try not to get wet from the water spraying everywhere.
This is a bottom view of the 3D CFD results of this injector. It is pleasing to see that it matches somewhat closely. The results from the simulation also matched what was found on the test stand. The contrast was altered on this image to make the pattern easier to see.
Injector References
These are a list of references I have used to learn about the design of different injector types. Not all give methods on how to design your own injector for your engine, but they are a good start. All resources were found online and some may require payment/membership to the website they are published on, however if you are a student see if your school has access to these papers, either through their library website or with your school email associated with each of these websites.
Impinging Injectors
This NASA SP8089 gives details on all sorts of injectors and even some equations on how to design and quantify them:
https://ntrs.nasa.gov/search.jsp?R=19760023196
A book by Huang and Huzel entitled "MODERN ENGINEERING FOR DESIGN OF LIQUID-PROPELLANT ROCKET ENGINES" is a great resource for learning the design of all injectors, especially impinging. Unfortunately, I cannot link a pdf to it as it is a textbook. Check your school library or online to see if you can get a copy if it interests you sufficiently.
Pintle Injectors
This is the injector that I have the most information about and it is proportionally complicated. There are different types of this injector and likewise becomes hard to quantify for each specific situation. These are not all the resources I used for my injector but this is a good start with a good amount of information.
This reference is good for planning a design procedure for a gas-liquid continuous injector:
https://arc.aiaa.org/doi/abs/10.2514/1.B36301
This reference gives an overview of how to test a single discrete pintle injector element as well as methods for the calculation and comparison of discharge coefficient:
https://www.sciencedirect.com/science/article/abs/pii/S0094576518309883
This gives a good history of the injector and a lot of insight into the qualitative aspects of a pintle injector:
http://www.rocket-propulsion.info/resources/articles/TRW_PINTLE_ENGINE.pdf
This paper gives a good scope on the dimensions used in a gas-liquid pintle as well as CFD simulation of the injector spray. This is mainly good for visualization:
https://www.sciencedirect.com/science/article/abs/pii/S0094576516314035
This paper is good for a liquid-liquid pintle and comparing to experimental data.