propulsion Archives - Castle Point Rocketry

News

Testing Review

July 18, 2019July 19, 2019
by Dakota

[This is a long article, but it has some very important updates.]

Over the last ten days, Castle Point Rocketry tested our proof-of-concept rocket. We traveled from HQ in Hoboken to our test site in southern New Jersey on July 9th and returned on the 18th.

Here’s what you need to know.

Setup

The preliminary phases of our testing schedule took longer than anticipated. Though this was the team’s third visit to the site, more infrastructure needed to be laid out. We ordered chemicals, installed the truss, and established ground control. We spent Wednesday through Saturday making sure everything was perfect.

Tank Testing

As mentioned in Ready to Test, validating the liquid oxygen (LOX) tank was the first step of our testing procedure. Infinite Composites Technologies manufactured our LOX tank, and they asked that we do preliminary tests for them.

On Sunday, July 14, an advisor approved us to start testing.

Time lapse footage of our LOX tank going up.

TT.01: COPV Cryogen Validation

Composite overwrapped pressure vessels (COPV) such as our LOX tank have been proven capable of withstanding high pressures. COPVs provide a burgeoning market for lightweight tanks by eliminating the need of an internal metal liner. The manufacturer expected that their formula would withstand low temperatures, too. But putting both together… That was our job.

After retreating 100 meters to Ground Control, we opened valves in a unique sequence to begin the test. First, cryogenic liquid nitrogen (LN2) filled the tank 4/5 full. Then we squeezed pressurized gaseous nitrogen (GN2) into the space that remained, elevating the tank pressure to 500 psi. (Air pressure at sea level is ~15 psi.)

From our perch in Ground Control, we couldn’t see exactly what was happening. But we had an array of sensors and cameras on-site that processed live data back to us.

The team watches the LOX tank closely… from afar.

The LN2 in the tank caused a thin layer of water to condense — and then freeze — on the tank. This temporarily turned the tank from black to a cloudy white, then back to black when the ice melted.

After holding for several minutes with no drop in pressure, the team agreed there were no leaks present. Ben actuated the dump valve, releasing the remaining LN2 in a plume of white steam, and we approached the tank to inspect the tank. No cracks, vents, peeling, or patching were found. We concluded that the Infinite tank could withstand cryogenic temperatures at high pressures.

TT.02: COPV LOX Compatibility Validation

TT.01 took less time than expected. On the order of hours. We planned the day with 4 hours for each test, yet TT.01 only took 34 minutes from start to finish. Too easy. Upon consensus from Nathan (Team Lead), Monica (Safety Officer), and Luke (Industry Advisor), we moved swiftly into TT.02.

Following TT.01, there wasn’t much to do in terms of preparing for the next test. After all, we simply needed to swap out an LN2 cylinder (called a dewar) for a LOX dewar. This required new hoses, too, but we came prepared. Within 30 minutes, we were back at Ground Control.

A screenshot from the program recording our incoming camera feeds. CCW from top: Infrared thermal imaging, test site visual, and dump valve close-up.

We took our time with this test. Though LOX boils at a slightly higher temperature than LN2 (-297°F instead of -320°F), it is much more dangerous. When LN2 boils off, it creates GN2. GN2 makes up 70% of the air we breathe — it is stable, is non-reactive, and plays well with others. LOX, however, boils off into gaseous oxygen (GOX). GOX is incredibly reactive, as oxygen is the driving force of any combustion reaction. With the slightest disturbance, a thimbleful of LOX can create a dazzling explosion. Should either GOX or LOX chemically react with the experimental COPV, a hole would release all oxygen at once, providing the basis for a massive fireball.

Luckily, we did not deal with any such eruption. Though the tank off-gassed a lot more than expected, much more than the LN2 run, we rang in yet another success. A holding time of 10 minutes proved our tank held an adequate amount of LOX to launch with.

TT.03: COPV Pressurize LOX Validation

Though we took our sweet time to make sure all of the GOX had adequately diffused before we drove the van up, we still had a remarkable amount of time left in the day. “Why not do another test?” we thought.

This test required LOX again. Under pressure, this time. We took a collective breath and pushed “Start.”

Once again, we were surprised by the amount of gas released from the system, but we assured ourselves it was nothing to worry about. The pressure did stay constant at ~300 psi for the full test, which indicated we didn’t have a leak.

Full Stack Testing

Following our three successful Tank Tests, we went into overhaul mode. Full Stack Testing required removing the LOX tank from the gantry hoist, placing it back in the rocket, and raising the rocket on the truss. Additionally, we needed to move our ground support relay boxes and fire extinguishers. (These relay boxes are like runners in a relay race. They act as a hand-off of information between Ground Control and the valves and sensors.) This change took the team a full day to complete.

FST.01: Full Stack Pressurization Test

Now comes the part of our story that gets a little bit… sad. On Tuesday, July 16th, we launched into FST.01 with great expectations. But the pressure just wouldn’t build. FST.01 only required the use of GN2. Since GN2 is practically harmless, Luke approved us to stay on-site with the rocket until we reached a pressure of 100psi. But even with our gas cylinders all the way open, we simply couldn’t get above 25psi.

Monica, Will, and Luke descended upon the LOX tank, assuming it (or one of its fittings) was the culprit. After a lengthy, methodical search, they found the problem: A hairline crack had formed on the very bottom of the tank. This was a deal-breaker.

Here’s the deal with cracks: They aren’t good for structural stability. Even at pressures as low as 25 psi, the crack was undoubtedly growing. Undetectably slowly, maybe, but definitely growing as GN2 tried to force its way from high to low pressure. Had we increased the system pressure any more, this effect would have increased dramatically — ending in a catastrophic burst as all of the GN2 left at once.

A tank exploding onboard the rocket was not what we wanted — so we unfortunately had to call off the rest of our testing schedule.

Moving Forward

So, what happens now?

After careful consideration of the data, the team concluded that the hairline fracture had occurred from “temperature cycling” the COPV. The tank went through a series of contractions and expansions as it got subcooled then superheated, which delaminated the layers of the COPV. In much the same way that ruffling a phone book puffs it up, this had introduced space between the “pages” of the tank wall, eventually leading to a full crack.

So, Castle Point Rocketry is still in our Testing Phase. Though we weren’t able to get through all five Full Stack Tests this week, we will soon have another tank. In the meantime, though, it was nice enough just to raise a rocket in the air — depressurized, of course — and look at what we had built.

For ease of access to the internal components, we didn’t add the nose cone, fuselage (skin), or fins. In a launch scenario, the rocket would look different.

News

Projected Testing Schedule

July 12, 2019July 19, 2019
by Dakota

Assuming the final pieces fall into place (a chemical delivery here, rocket fuel there), we will begin our testing procedure soon. We’ve laid out an explanation, but now… Here’s a timeline. Be sure to also check our Facebook page every once in a while for any updates!

Day Three:

Castle Point Rocketry is expecting a delivery of pressurized gas and cryogenic liquids on the morning of Friday, July 12. Once those come through, we will begin pressure testing the subassemblies of pipes and fittings. We have already begun removing small subsections from the full stack in preparation for this.

We have also pulled the liquid oxygen (LOX) composite overwrapped pressure vessel (COPV) from the air frame. In the afternoon, we will have the time to move on to our first official test, TT.01: COPV Cryogen Validation.

Though we are not rushing, we also look forward to having enough time to complete TT.02: COPV LOX Compatibility Validation today.

Day Four:

Following closely on its heels, we are waking up early Saturday morning for more testing.

TT.03: COPV Pressurized LOX Validation will start first, right after we drive to purchase our surrogate rocket fuel from a nearby airport.

That afternoon, we will finish FST.01: Full Stack Pressurization Test.

Day Five:

Sunday will be a long day. We have scheduled both FST.02: Cold Flow Test and FST.03: Ignition Sequence Test, when we first load the fuel tank with actual fuel.

Day Six:

Monday morning, we’ll be up early again to test FST.04: Hot Abort Test. Assuming it goes well, that afternoon will see our first full launch setup with FST.05: Full Stack Hot Fire Test.

Day Seven:

We have reserved some time on Tuesday for a second hot fire test, in case we need it. It also serves as a good spill-over time for any tests which take longer than expected, or a rain date for any other tests.

And then we drive home!

News

Testing: Days 0 – 2

July 12, 2019July 12, 2019
by Dakota

These last three days were dominated by the minutiae associated with launching a rocket. Sure, we did as much setup as possible beforehand. But there’s much more to do once the rocket actually gets here.

Day Zero: Packing Up

Tuesday, July 9, was particularly chaotic. Myriad administrative snafus demanded quick action from the whole team. Dakota grabbed a truck so we could transport our rocket. And, most pressingly, we needed to move everything out of our Griffith work space and go mobile.

Monica, Nathan, Ben, Dan, Will, and Dakota (behind camera) take a breather after lifting our rocket onto the dolly. More straps and plastic wrap were added to further reduce movement.

Due to everything else going on, we didn’t even get around to packing until 6:00pm. By that point, we knew we would be leaving late, but we hoped 1:00am would be the limit. Yet 1:00am came and went.

At 2:30am, we gently loaded the rocket into the truck. (Lift gates are a blessing.) It took all of the remaining six of us, but it happened without a snag. Come 4:00am, we were practically finished.

Our lab and rocket, all crammed into one 26′ box truck.

The lab looked lived in, but at least it was devoid of much our mess. And what better time than 4:00am to start 2.5 hours of highway truck driving?

Day One: Tying It All Together

Dakota scrambled into the driver’s seat. Ben and Rodrigo piled in and kept him company to Salem County. (And kept him awake.) After all, there’s that old saying: “Six eyes are better than two with a $500,000 rocket and boxed-up laboratory in tow.”

They arrived okay — and just in time to volley off the first round of calls to gas/cryogenic companies to confirm our delivery schedule. Starting at 8:00am on the dot, Dakota, Will, and Nathan were on the phone until noon. Then, it was just a matter of waiting for the team to assemble.

The shed at left will store our chemicals. Ground control is barely off-screen to the right. The test site is 2,000 feet behind the camera.

A brief team meeting preceded driving to the test site and setting up ground control. The remaining daylight hours were swallowed up by running 2,000 feet of fiber optic cable and cleaning out the on-site shed for our chemical deliveries.

Day Two: Finishing Touches

It took a little under 14 months, but it’s finally happened — Castle Point Rocketry has received a shipment of rocket propellant. (The oxidizing half, at least.) Shortly after a late start to the day, the first cryogen delivery came through: liquid nitrogen and liquid oxygen. It’s a good thing we got that shed cleaned out!

For the rest of the day, we split up into small project teams. The test stand was bent and needed a quick weld — Rodrigo made short work of it. Over the weekend, the gantry hoist slipped on the muddy ground. We righted and reinforced it. Ben and Nathan focused on getting the avionics and valves attached and laid out.

CP Rocketry Test Site, featuring the Mayor’s Soy Beans

Though all parts of plumbing were screwed together, Will and Monica still had some tightening to do. They spent most of the day in the truck-lab working on the rocket. Dakota manned the ground control station, where there is WiFi and a view of the road, to finish up some remaining purchases and administrative duties.

The inside of our truck-turned-lab. (There are no chemicals stored inside.)

When interrupted by a small storm, we broke for lunch. Then, after three more hours in the sun, a massive thunderstorm rumbled in from Delaware. We took it as our cue to leave before it got dark out — but not without first getting absolutely drenched.

What’s Left

We’re zooming in on the first day of true testing. It’s highly likely that, after a delivery Friday morning, we will be testing Friday afternoon. In the meantime, the team still has a short laundry list of tasks to accomplish. First and foremost, we need to reestablish ground control. It kinda tipped over in the rain, and we had to rescue it.

Stay tuned to our Facebook page for a live feed of testing. (Or the live feed itself.)

News

“Always Open”

June 24, 2019June 25, 2019
by Dakota

Those of you who follow us on Facebook may have noticed a quirk. Under the “About” section, just under the map of Hoboken, it says “Always Open.” This is no mistake.

Here’s a little peek into what we’ve been up to this weekend — at all hours of the day.

Tank Cleaning

As you may recall from a few weeks back, the vast majority of our propulsion system needs to be “Clean for Oxygen Use.” We have finished pipes, fittings, and adapters and are onto the bigger pieces: our tanks.

The helium and oxygen tanks we use were made special for us by Infinite Composites Technologies. We are rigorously cleaning both with isopropyl alcohol baths to dislodge any remaining construction materials from the inside.

Since the alcohol coming out is dirty, we also needed to clean it for reuse — about 34 liters (9 gallons) worth. Monica and Dakota spent much of Saturday vacuum filtering all 34 liters.

The process of vacuum filtration. Dirty isopropyl in the top, clean out the bottom.
We changed filters once every 2 liters — about when they started to look like this.

Load Cell Calibration

A by-product of last weekend’s Dry Run Mechanical Test, we are confirming all of our load cells work. In order to accurately measure the thrust of the rocket, we will attach it to the ground with cables. These cables will pull on our load cells, which tell us how much thrust the rocket has.

Ben was hard at work making sure the code was solid while Will used the engine hoist to test a few known loads.

Will’s load cell testing apparatus.
Ben working on code — as seen through the clean room walls.

Now that it’s Monday, Will, Tom, and Abe are out on Walker Lawn with the load cells and duck bill anchors. The anchors are being used to test the load cells, and vice versa. We want to make sure our duck bill anchors are rated properly. After all, the last thing we would want is for an anchor to come out of the ground during a test.

Fitting Tightening

Last but not least, we have our piping. As mentioned above, all of the propulsion subassemblies have been cleaned for oxygen use. Now, it’s just a matter of preparing them for testing and launch.

Our propulsion system has threaded joints from two rival piping standards: JIC (Joint Industry Council) and NPT (National Pipe Tapered). Each of these two standards comes in multiple sizes — and each size requires a unique tightening force. Larger threads require greater tightening force — as much as 100 foot-pounds.

Nathan and Will using two wrenches and a vise to tighten a JIC-12 fitting to specification.

In order to accurately tighten each joint, we use both a torque wrench and a crescent wrench. (One to twist, one to hold the rest of the subassembly.) For some subassemblies, more advanced methods are needed, though. In the case of particularly wiggly or oddly-shaped pieces, a vise is necessary to get a good grip. Thus the above picture outside the clean room — as long as the interior is not compromised, the outside of the subassembly can always be cleaned again.

A Little Fun

We also manage to have a little fun after a long day’s work. (And usually right before another long night’s work.) Friday night, we all stepped outside to enjoy a barbecue dinner. Because how else would we ring in the first day of summer?

News

Industry Advisor Review

June 14, 2019June 14, 2019
by Dakota

It was June 10th, 2019. A thick, foggy mist had swallowed up New York City. Hoboken traffic was, unsurprisngly, backed up half an hour. And in the back of a machine shop at Stevens Institute of Technology? Nine rocket enthusiasts were ironing out a testing procedure.

Castle Point Rocketry invited our industry advisors, Rich and Luke, in to review our final testing procedure. Somehow, there were still some introductions to be made, too!

Rich Kelly (left) and Luke Colby (right) introduce themselves before we get to work.

Luke Colby is the President and CEO of Triton Space Technologies, providing engineering design services out of Boston, Massachusetts. Luke has been advising our project by phone since Fall 2018, but we have never met in person. His company also manufactured a handful of valves that will travel aboard our rocket.

Rich Kelly is a Senior Project Engineer with Valcor Engineering, based in Springfield, New Jersey. Due to Valcor’s proximity, he has visited our lab many times over the last few months. And they also manufactured several valves we will be launching into space!

After introductions, we quickly showed Rich and Luke the latest work we had done on the rocket. Then, it was down to business.

Sitting down to hammer out the details of chemicals testing.

We crowded around an imaginary table in our makeshift conference room. (Spoiler: It’s our lab with a portable projector screen.) We had less than six hours to go through the entire 64-page Propulsion Testing Document, so… there was little time for games. (There was, however, time for lunch. Self-care is important and, as Rich reminded us, “The food’s not getting any warmer!”)

The team led Rich and Luke through our testing plan page by page, halting when there were questions or suggestions. After reviewing three tank tests, five full-stack tests, and ten procedural methods, we reached the end of our packet. We called it a day, but Luke and Rich left us with a few pointers. Among other things, the team is revising our waste management plan, redesigning the igniter (again!), and renting more robust pressure regulators.

Just some happy nerds doing space stuff.

It was then time to set our sights on the next big exciting task: Dry Run Testing!

Uncategorized

“Clean For Oxygen Use”

June 11, 2019
by Dakota

We can all probably agree with the relative levels of cleanliness. Around the bottom of the scale, there’s “I Am Comfortable Living In This.” A little more clean, you probably find “Company is Coming,” closely followed by “My Parents Are In Town.” Near the top of your list, you probably find “Apartment On The Market.”

“Clean for Oxygen Use” may top the charts. It’s certainly not a household standard.

Bottom shelf: Unclean. Top shelf: Clean.

This cleaning method is the entire reason we constructed our clean room. Much of our rocket will come into contact with high-purity oxygen, whether in liquid or gas form. Gaseous oxygen loves lighting things on fire, and liquid oxygen freezes most substances solid — so we need to be sure everything is as clean as humanly possible. To do so, we have a six-step cleaning process.

Step One: Alcohol Bath

After we identify a subassembly to clean, we remove each piece from storage. We bathe each individual fitting, pipe, adapter, and valve in isopropyl alcohol. (That’s the same alcohol you put on wounds to clean them.) For 12 minutes, they rattle around inside an ultrasonic chamber. By vibrating them very, very quickly, the machine dislodges defects, dust, and other gunk that is clinging to them.

Our ultrasonic bath is located on the left.

Isopropanol is also a dehydrant. This agitation bath ensures every out-of-the-way nook and cranny is water-free. Any water left in the system would freeze in contact with cryogenic liquids, decreasing functionality and making the rocket explosion-prone.

Step Two: Nitrogen Purge

After they’re removed from the bath, each part is individually inspected for remaining debris.

A tee junction during the nitrogen drying cycle.

Then, every part is dried with a pressurized jet of filtered nitrogen. Not only does this ensure no isopropyl alcohol is left on the part, it blows away any remaining foreign materials.

Step Three: Alcohol Rinse

As if Step One weren’t enough, we then subject each component to yet another round of alcohol. This time, the isopropanol is targeted in a stream. The entire part is washed beneath a squeeze bottle before moving on to Step Four.

A tee junction having an isopropyl alcohol shower.

Step Four: Nitrogen Purge

More drying! Like most alcohols, isopropanol is flammable so we need to make sure each part is bone-dry before assembly. This last round of nitrogen is usually enough to get the last bits of stubborn junk off of our fittings.

Step Five: Critical Inspection

Once the second nitrogen blow-down is complete, we are fairly certain nothing remains. But just to be sure, though, we inspect each piece from every angle for leftovers. Inside and outside, nothing is allowed to escape our prying eyes. And on the off-chance we still find refuse holding on? We restart the whole process from scratch. We bought smaller ultrasonic bath just for that purpose.

Step Six: Assembly

Finally, we are sure that our parts are Clean for Oxygen Use. We bubbled, tossed, dried, washed, and dried most everything (and even brushed some with a high-grade pipe cleaner), and it’s time to put the pieces together. One by one, being sure not to stir the air or drop anything, the rocket starts taking shape. We have 24 subassemblies ranging in size from one component to thirteen.

One of the subassemblies we will be using for tank testing.

Each über-clean subassembly is then given a new home on the high shelf in our clean room. Small subassemblies are bagged and given a unique name so they don’t get confused down the road.

And that’s how you make a rocket Clean for Oxygen Use!

News

Successful First Igniter Test

April 26, 2019April 26, 2019
by Dakota

As those of you who follow our Facebook and Instagram (@cprocketry) pages might know, we had an exciting night this past Monday.

After several weeks of deliberation on how to ignite, we settled upon a cocurrent jumpstart from a solid rocket motor. Imagine our large engine housing a smaller engine. The heat of combustion of the small engine ignites the rush of fuel traveling through our bigger engine and — POOF. We have lift-off.

Capturing footage of a solid rocket motor — plain, and without our igniter design.

The Theory

Combustion theory is hard. A good assumption to make, one that doesn’t need a Ph.D., is that our oxidizer and fuel will both need to be vaporized before burning. That is: They need to be a gas. This, at least, is handled by the injector.

The second part is the orientation of the starter flame. To summarize many, many studies and papers in one quick sentence: “The igniter should be oriented the same direction as fluid flow.” (Downwards.) Doing so reduces turbulence inside the engine, reducing the chances of the engine exploding. It also orients the igniter flame away from the shower-head face of the injector, which could cause some orifices to close up.

The big question, then, is how to take a solid, put it inside a liquid shower, turn it all into a gas, and light the whole thing on fire.

The Model

A few ideas came to mind. Do we hang it from a shepherd’s crook? Adhere it to the bottom of a plate? Stick it to the side of a wood pole? The problem with each of these ideas was two-fold. One: With the relatively small size of our engine, how do we ensure that these bulky geometries don’t increase the turbulence? Two: How do we take the straight flame from a solid motor and fan it out in all directions to ignite as much fuel as possible?

Luckily, a little bit of CAD helped solve all of these problems. We settled on a cylindrical igniter plug. This design allows for a variety of motors to be installed and incorporates a diverter to spread flames in all directions. (You can see the outline of our model in the video below.)

The Test

After a watching our first prototype plug print for a few hours, we couldn’t wait to test it out. Even though it was printed from polylactic acid (a plastic which melts at 220°C) and solid motors burn MUCH hotter than that, we figured we could get a few seconds of slow-motion burn time on camera.

And boy, did we get a show.

Proof of concept: The igniter plug design works!

After slowing five seconds of burn time down to several minutes, we had 10 seconds worth of good footage. You can see in the video that each of the radial flame outlets has even distribution and steady flow. Which is just what we want!

You can also tell when the melting plastic starts to disrupt the flame dispersion — about 7 seconds in. The result of that were the smoking, charred remains of our plastic igniter plug. It gave renewed meaning to the word “pungent.”

The Future

Though the test was a success, we began optimizing our design immediately. We inverted the design to allow for ease of access during testing and launch. Loading the motor from the bottom required the flame to shoot upwards, though. This, in turn, required us to reorient the flame outlets.

The team also bought several cylindrical bars of aluminum. After a successful test of our new design, we plan to machine a couple out of aluminum — which hopefully won’t be reduced to a smoky pile of ooze.

Currently, the greatest design hurdle is how to remove the igniter from the engine during testing. Since the rocket is attached to the ground and the current igniter design doesn’t move, we’re back to square one — disrupted Mach flow in the engine. (A result of turbulence.) Luckily, we’ve got a crack team of young rocket scientists working on it.

So that’s where we stand! Proof of concept confirmed, future planned out, and a moveable design in the works. Just in time for our rapidly-approaching testing timeline.

Uncategorized

The Hunt for the Best Igniter

March 14, 2019March 14, 2019
by Dakota

So, rockets burn fuel. That much makes sense. And that combustion of cold liquid creates a lot of hot, pressurized gas that makes the rocket go upwards. Got it. But how does that begin?

You can’t light a kitchen stove without the internal igniter sparking. In the forest, you can’t start a campfire without flint and steel. (Or lighter fluid and a barbecue lighter.) But inside a rocket engine… things are a little more complicated. Castle Point Rocketry has been upending shelves worth of books (all online, don’t worry) searching for the question on the forefront of our minds: “How do we start our engines?” And we’ve narrowed it down to three major contestants.

“The Cotton Ball”

Soak a cotton ball (or other highly-porous material) in something really flammable.
Stick said drenched cotton ball on the end of a metal stick.
Set rocket over stick, with cotton ball inside combustion chamber.
Light cotton ball on fire.
Release the fuel and LOX.

Proof of concept: It’s been done before.

Potential drawbacks include the sudden onslaught of liquid, though flammable, extinguishing the burning cotton ball.

“The Salt Crystal”

Finish researching oxidizing rock salts. Some salts, when heated, spontaneously burst into flames and release copious amounts of oxygen — which helps fuel more decomposition.
Acquire a small-ish amount of the chosen salt.
Carefully pack the salt into a small container on the end of a large stick.
Gently place the rocket over the stick, with salt container inside the engine.
Warm the container and wait for sparks, then release the fuel and LOX.

A snippet of molten oxidizing salt shooting flames.. (1:55 – 2:10.)

Potential drawbacks include the risk of salt decomposing before the igniter set-up is prepared.

“Engine-ception”

Now imagine, if you will, an engine inside an engine.

Source a suitable solid rocket motor, given its thrust/time curve.
Semi-permanently affix the motor to the bottom face of the injector.
Ignite the solid rocket motor.
Release the fuel and LOX.

Solid rocket motors produce a very well-regulated flame over a set period of time. Additionally, this set-up allows both flames (from the starter and the combustion) to travel in the same direction. By doing so, we can reduce the chance of the starter blowing out!

Uncategorized

Full Circle

February 12, 2019February 12, 2019
by Dakota

One of the most competitive facets of our rocket is the engine assembly. Rather than the hundreds (and sometimes thousands!) of parts used to create a traditional rocket engine, our team has streamlined the process into just two: an injector and a nozzle.

Since they’re so integral — and so specialized — we’ve spent hours upon hours designing, modeling, and simulating these parts. And we’ve come full circle.

Phase One: Solidworks

The CAD starter kit for any Stevens engineering student, Solidworks helped us make our first few rounds of injectors. It was pretty cool. Tom spent a lot of time on it.

Injector, Mk. 1. It was bulky, but it got the job done.

Phase Two: COMSOL

Once happy with the geometry, we needed to simulate fluid flow through the manifold. What better software to use than a COMSOL, taught to grad students and marketed as a multi-physics program with a hefty computational fluid dynamics (CFD) engine. Dakota’s self-taught COMSOL regimen came to a standstill when, no matter how finely we meshed the part, physics kept saying, “Nope.”

COMSOL needed the injector flipped inside out. Crazy, right? This part seemed okay…

… But this one obviously had some problems.

They ended up both not working out. One kept growing gnarly spikes, the other had faces that wouldn’t meet up. In the end, no amount of curve smoothing could fix it.

Phase Three: Ansys Fluent

So, what now? Let’s try another CFD! Stevens also offers Ansys to its students, which comes with the handy Fluent plug-in. Abe and Dakota worked on modeling the part in early January, only realizing after returning to campus why it wouldn’t work: Even when simplifying the models by excluding symmetrical pieces, the parts were MUCH, MUCH too big for our Ansys versions to handle. (If years of math has taught me one thing, it’s that 1.1 million “cells” is larger than 512 thousand “cells.”)

Phase Four: Back to Solidworks

The most recent iteration of the injector. We’ve come quite a long way!

We were lost. What do we do? Our school’s CFD programs weren’t working. We had tried simplifying the parts to no avail. Were we just going to hope our applications of what we read in NASA journals would work? Would our rocket engine be held together by dreams and back-of-the-napkin calculations?

Of course not. In the words of Professor Aziz, who teaches courses in Modeling and Simulation at Stevens, “You started
in Solidworks. Why did you leave?”

This simulation tracks individual particles traveling through the injector manifolds. Red means high speed, blue means low speed. Good news: They don’t stop!

So we’ve come full circle. We’re back where we started. We’ve been running fluid flow simulations fairly smoothly, these last couple of weeks. Sure, we keep iterating and optimizing. But at least it’s all built-in, now.

And boy, are the simulations colorful.