Monday, May 8, 2017

Just a BABI

If you want to get the attention of someone at CSBF in a negative way just include the word "pressure" post scripted by the word "vessel." CSBF was concerned that our Bernoulli Air Bath Intensifier (BABI), a 28 oz. plastic (PETE-1) container modified as an atmospheric accumulator (a.k.a. a Great Value peanut butter jar) might be a pressure vessel but it's specifically designed to not be a pressure vessel. BABI has a single design goal: to circulate air across the sensors in order to sample fresh air throughout the flight, so that we are not taking stagnant readings of the air inside our hull. We can not reach the pressures preferred by our sensors, but testing shows that we can improve the amount of air available at float, and we are hopeful that we’ll be able to record some meaningful data. This challenge of this project has moved away from our original science and has morphed into an engineering design quandary about moving air in the stratosphere.
CONSTRUCTION
For reproducibility throughout our prototyping, we make our capillary holes with a Cobalt Steel Micro-Size Drill Bit (0.0020" diameter) in a thin piece of aluminum from the wall of a beverage can. We then drill a 1/16” hole in the wall of the plastic jar and seal the perforated aluminum patch over it, aligning the holes. We have yet to settle on epoxy or silicone caulk as the best sealant; both have benefits and drawbacks and we have to consider what will react with our sensors.





For each BABI, ¼” brass hose nipples are mounted in the sturdy plastic at the “bottom” of the jar, sealed, secured in place by washers sandwiched onto the plastic, then left to cure in low vacuum. Our 7.5-to-15 PSI (0.52-to-1.03 bar) safety valve is similarly mounted.
A slot is drilled into the lid through which the wiring is fitted and then sealed. This is currently an unsatisfactory and fragile seal, and we are investigating better ways to install a pass-through.
The plastic lid threads are wrapped in 5 turns of teflon plumber’s tape before the lid is sealed. Structural elements such as the mounting brackets and sensor standoffs have been omitted from this drawing/description for clarity. The pump in question is a SparkFun ROB-10398 that runs on 12VDC and outputs a stated flow of 2.50x10-4 kg/s (12 LPM stated by manufacturer, and assuming standard atmospheric density at sea level.)
IMG_20170409_202438.jpg
The BABI (bottom left) inside our BestValueVacs 3 gallon vacuum chamber with our SparkFun pump and Arduino Mega.


NORMAL OPERATING PRESSURE for GROUND AND FLIGHT OPERATIONS
We have made several tests both at local ambient pressure as well as inside our vacuum chamber. All tests have been made at room temperature; thermally-controlled tests at -40°C and at +40°C are planned.
Test Run
External Pressure Max (bar)
External Pressure Min (bar)
Internal Pressure Max (bar)
Internal Pressure Min (bar)
Comments
Hole
(in.)
Pump
PB BABI - 1.1
(03/29/17)
ambient
ambient
1.660
ambient
bottom dimpled out
none
SparkFun at 18V
PB BABI - 1.2
(03/29/17)
ambient
ambient
1.061
ambient
leak rendered this test useless
none
SparkFun
PB BABI - 1.3
(03/29/17)
ambient
ambient
1.056
0.07
Considerable and consistent leaks at lid threads
none
SparkFun
PB BABI - 1.4
(03/29/17)
ambient
ambient
1.035
0.07
No deformation. Small leak at lid threads
none
SparkFun
PB BABI - 2.1
(04/9/17)
1.005
0.069
1.555
0.09
Minor Leaks through lid
.008”
SparkFun
PB BABI - 2.2
(04/9/17)
1.003
0.071

1.555
0.09
Minor Leaks through lid
.008”
SparkFun
PB BABI - 2.3
(04/9/17)
ambient
ambient
1.015
ambient
Minor Leaks through lid
.008”
SparkFun
PB BABI - 2.4
(04/9/17)
ambient
ambient
1.04
ambient
Found leaks part way through our test. Covered it up with our hand to bring us up to our max pressure.
.008”
SparkFun
PB BABI - 3.1
(04/12/17)
1.01
0.05
1.563
0.06
Minor Leaks through lid
.008”
SparkFun
PB BABI - 3.2
(04/12/17)
1.01
0.05
1.170
0.06
Massive leakage
.008”
SparkFun
PB BABI - 3.3
(04/12/17)
1.009
0.05
1.163
0.06
Massive leakage
.008”
SparkFun
PB BABI - 3.4
(04/12/17)
1.01
0.05
1.166
0.06
Massive leakage
.008”
SparkFun


Maximum Pressure Testing
METHODOLOGY
We performed these maximum pressure tests to observe the behavior of BABI at extreme internal pressures. We manufactured several BABIs with only an inlet port (¼” barbed hose nipple). We used much stronger pumps for two early trials, then the SparkFun pump for all later trials. One trial used an air compressor capable of 250 PSI and the other a standard bike pump.


CONCLUSIONS
The result was a consistent maximum pressure reading of 1.56 bar after 40 seconds, with minor leaking around the threads. The two ‘extreme’ trials resulted in a breach where the lid was forced off the threads. There were no projectiles created from the breach and no permanent deformations of the assembly.
Test Run
Ext.  Max P (bar)
Ext. Min P (bar)
Int. Max P (bar)
Int. Min P
(bar)
Comments
Capillary Hole Size (in)
Pump
PB BABI
(3/29/17
ambient
ambient
1.38 ± 0.69
ambient
jar lid came off threads
none
Bicycle Pump
PB BABI (03/30/17)
ambient
ambient
2.89 ± 0.13
ambient
Considerable and consistent leaks at lid threads
none
250 PSI Air Compr.
PB BABI
(4/16/17)
ambient
(1.00 ± 0.01)
1.00 ± 0.01
1.56 ± 0.01
1.00 ± 0.01
Minor leaks through lid
none
SparkFun
PB BABI
(4/16/17)
1.00 ± 0.01
1.00 ± 0.01
1.56 ± 0.01
1.00 ± 0.01
none
plugged for overpressure test
SparkFun
PB BABI
(4/16/17)
1.00 ± 0.01
1.00 ± 0.01
1.56 ± 0.01
1.00 ± 0.01
none
plugged for overpressure test
SparkFun
PB BABI
(4/16/17)
1.00 ± 0.01
1.00 ± 0.01
1.56 ± 0.01
1.00 ± 0.01
Consistent pressure despite leaving pump on for longer
plugged for overpressure test
SparkFun
PB BABI
(4/16/17)
1.00 ± 0.01
1.00 ± 0.01
1.56 ± 0.01
1.00 ± 0.01
Perhaps minor leak at hose fitting
plugged for overpressure test
SparkFun
PB BABI
(4/16/17)
1.00 ± 0.01
1.00 ± 0.01
1.56 ± 0.01
1.00 ± 0.01
none
plugged for overpressure test
SparkFun
PB BABI
(4/16/17)
1.00 ± 0.01
1.00 ± 0.01
1.57 ± 0.01
1.00 ± 0.01
Perhaps pump cannot handle back pressure
plugged for overpressure test
SparkFun




RISK MITIGATION
We understand that CSBF has a vested interest in equipment that does not explode. To that end, we have implemented a complementary array of safety features to prevent an overpressure disaster, listed here in descending order of probability/reliability:
  1. Intelligent automation of pump:
We are not running the pump until our barometers indicate that we are at ~100,000ft altitude (~0.01 bar), and the pump will automatically shut off when ambient pressure rises over 0.01 bar. Similarly, if the pressure inside BABI rises over 1.25 bar, the software shuts the pump off. This is our first and best defense against overpressure.
Our vacuum chamber can only draw down to around 50 mbar, which is significantly higher than the 5 mbar expected at float. However, at 50 mbar we saw that the pump could only create a pressure differential of approximately +20 mbar (50 mbar ‘ambient’, 70 mbar inside BABI); mathematical modeling shows that at 5 mbar this will be even less. In our estimation the likelihood of a breach is minimal.
Ambient tests have shown that, with the pump off, BABI de- and re-pressurizes only slightly slower than our vacuum chamber. We are not concerned about a rupture (neither implosion nor explosion) due to ambient pressure differentials on ascent/descent. Note how closely the curves of the BABI and its “ambient” vacuum chamber pressure align over time as shown below.

  1. Pump physically incapable of overpressuring jar
Our pump has shown itself incapable of pushing more than an extra 0.5 bar of pressure. In sea-level-pressure tests without any capillary holes and a safety valve set to 25 PSI (1.72 bar), we were unable to exceed 22.6 PSI (1.56 bar). We heard a distinct fluttering noise and isolated it to the pump housing; this was repeated on subsequent tests. While the stated max output of our SparkFun pump is 32 PSI (2.26 bar), we have found that it just can’t push that hard.  The only time the lid came off the threads, we were using more powerful pumps (air compressors and bicycle pumps.) Even in these extreme cases the threads of the jar gently gave out before a fatal burst could be considered.

  1. Safety valve
The pop safety valve is set to activate when the internal pressure of BABI rises 0.5 bar (7 PSI) above ambient pressure and testing has shown that the safety valve is effective. The mass flow rates for the safety valve, capillary hole, and the pump were derived experimentally and mathematically. The pump’s maximum mass flow rate at ground level in Raleigh was consistently observed to be 1.405 * 10-4 kg/s, despite its higher specifications. Taking the combined mass flow rates of the capillary hole and the safety release valve, we observed a combined mass flow rate of 9.05 * 10-5 kg/s.
In the appendix are two samples of our data and calculations. We were able to measure ΔP/s and, assuming atmosphere is ideal, apply the law PV=mRT.  We considered the mass differences per second to give us a mass flow rate for each individual flow component of BABI: the capillary hole, the safety valve, and the pump.
The results of this experiment provide evidence that, when the safety valve is opened, the mass flow rate out of the BABI is less than the mass flow rate of the pump. This means that even when the valve is open we are still able to achieve a slight gain in air being delivered to our sensors.
The team would like to do more testing in order to optimize the BABI. Ideally, we would be able to maximize the airflow through the container while still providing a sufficient leak. We are working on a combination of testing, mathematical modeling, and engineering the assembly.  The BABI assembly leaks when the internal pressure is raised; we have been unable to accurately account for the mass flow through the various leaks, which result in a lower calculated mass flow rates for all three components. Further experiments with our wiring, materials used for patching holes, adhesives, and valve placement should help us fine-tune the BABI. We want it to not be a pressure vessel but we don’t want it to be this much of a non-pressure-vessel.

  1. Sealant faults
Our recorded “best” pressure (running the pump at 12V) was 1.56 bar, after which point (while testing at sea level pressure) we are distinctly able to identify leaks through our lid threads as well as through some of our homemade vents. The more the pressure, the greater the surface area of the leak.
(NOTE: even with these leaks, we are still building up a small ΔP; just nowhere near enough of a differential to be dangerous.)
  1. Threads fail before jar
In overpressure tests with a proper air compressor able to output 250 PSI, we found that the jar/lid interface deforms enough to allow large volumes of air to escape at the lid. The deformation occurs consistently around ~2 bar. During several trials the pressure then declined with the large volume flow rate of the compressor still active and lid still attached.  Two of the trials resulted in the lid coming off the threads with minimal momentum. Our bracket will constrain the lid from carrying completely away, but likely will not constrain the lid enough to prevent it from lifting slightly and venting. Further tests for this scenario are planned.
  1. Tube fastener fails before jar
In one other overpressure test, the vinyl tube with small zip ties popped off the barbed fitting before the jar lid popped loose. In all other tests the jar lid let go first.
  1. Contained in 0.0625” aluminum hull
Worst-case scenario: If our pump has somehow stayed stuck in the on position, and our code can’t cut its power, and we can’t override it remotely, and the pump surpasses all previously recorded efficacy, and the safety valve has stuck, and the jar self-seals all of its various leaks due to the heat, and the low-grade threads hold, and the tube stays on, then our BABI will become a true pressure vessel and may rupture. In this worst-case scenario, our hull is bolted together on all 6 sides with solid 6061 aluminum; any ‘shrapnel’ will be contained within and should provide no danger to HASP/CSBF hardware or personnel.


PAST FLIGHT HISTORY OF PRESSURE VESSEL
To the best of our knowledge and research, no one has ever flown a partially-pressurized perforated peanut-butter jar to near-space; this will be a first.

Fig. 13: An engineering concept image showing BABI operating at float altitude. All dimensions to scale.

CONCLUSION
The common definition for a pressure vessel describes it as a container that has a pressure radically different than its surroundings. Our data shows that our container does not meet that definition and will not build enough pressure to ever become a flight risk. Our team is confident that the possibility of dangerous pressure buildup is miniscule.











2 comments:

  1. That's really interesting. What do you opine? What's a best air compressor? I think, along with the best features when a compressor can fulfill one's need of air compressor with some added values,that is it.

    ReplyDelete
  2. Thank you because you have been willing to share information with us. we will always appreciate all you have done here because I know you are very concerned with our. the garage compressor

    ReplyDelete