PRV & Burst Disc For Pressurized Tank Holding Highly Concentrated H202 4

[MM904]

Hi all,

I have postedI am working on a rocket engine that will use HTP (98% concentrated hydrogen peroxide) as an oxidizer. Right now I am designing the hot fire testing tanks. Peroxides are awfully hard to work with because they continuously decompose, but I've found resources to get me along: Papers regarding storage of H202 (generally long-term, not pressurized) and typical ASME boiler standards for pressure vessel sizing, but most all of the ASME stuff have some form of "Not applicable for storage of H202, see reference X" which just brings me back to the first type of papers.

Even so, the design is 90% done. For reference, the tank specs are roughly as follows:
Capacity ~ 200 gallons
Max operating pressure: About 14 MPa
Material: 316 SS
Dimensions: Cylindrical shape, roughly 4 feet in diameter and 3 feet tall, with a wall thickness of about ~2 inches.
Storage duration: Less than a week.
Pressurized storage duration: Less than a day.

The last issue is sizing the pressure relief valve and the burst disc, which is giving me all sorts of issues as the two types of resources I've found have zero overlap on this part. I will need the tanks to be manufactured to the ASME code, however the manufacturers I've spoken with won't touch the sizing of the PRV and burst disc, they have asked me to provide that info for them, then they'll make me the tank. Any pointers or tips on how I can size these components correctly considering I'm holding HTP at high pressures would be much appreciated!


P.S. I am not opposed to simply oversizing the hell out of everything and calling it a day. However, to oversize something, you must first size that thing.

 

[gerhardl]

My advice is to take all technical discussions and procurement directly with specialized producers for bursting discs and PRV.

You need to decide bursting pressure, necessary tolerance of BP, and likely number of eqal discs to be produced/bought. All will inflence price.

If you want to estimate yourself the size of bursting discs, and to a degree PRVs, one rough way is to select an escaping speed 'normal' for gases in pipelines, say 40 m/s, then determine roughly how much have to escape in the first critical time, then calculate the necessary area, not bothering with temperature and pressure reduction.

PRV producers will normally have seizing tables to help you

Ask the producer for preliminary technical priced, not binding offer. You will get advice there of type, material and necessary flange dimentions for your tank. Your tank producer can then install after your procurement.

 

[georgeverghese]

These authors seem to know a thing or two about HTP decomposition kinetics - see page 8 and 9 of the attachment which states kinetics is first order. See if these folks can help you. I would imagine one of the failure scenarios which determines design case relief load would be some accidental contamination of the HTP - chlorides and bromides also catalyse H2O2 decomposition, so is it possible that an operator may have unknowingly filled a partially full tank with contaminated water? You would have better knowledge of what this HTP vessel is used for, so examine all these operations and formulate a credible failure scenario, which results in catalysed decomposition.

Why is pressurised storage time <24hours?

https://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/20010102648.pdf

 

[MM904]

George,

Thanks for the assist. The pressurized storage time is <24 hours because it is used as a propellant tank for hot fire tests.

 

[georgeverghese]

So in pressurised mode, decomposition is accelerated by higher ambient temperatures during hot fire test? So why isn't this tank insulated? Tell us more about what happens during a hot fire test which results in accelerated decomposition.

 

[MM904]

Georgeverghese,

Here is some background, I have erred on the side of being too detailed so I apologize for the length:

The purpose of this tank is to be used as a propellant tank when hot fire testing rocket engines on a test stand that use hydrogen peroxide as an oxidizer. For a test fire, the tank will be pressurized with nitrogen gas up to the desired tank pressure. The rest of the time, the tank will not be pressurized. During a test, the tank is far enough away from the engine that it will not experience any higher ambient temperatures.

On the day of a test fire, the tank will be loaded with the 98% h2o2. H2o2 undergoes continuous decomposition, albeit at a very slow rate (something like ~0.3% a year, I don't remember exactly and it depends on a lot of variables) when there are no catalysts to the reaction. When there is a catalyst to the reaction, the decomposition, which is exothermic and produces gaseous oxygen and water vapor, happens extremely rapidly. In terms of safety, some sort of contamination occurring while the h2o2 is in the tanks would cause this rapid decomposition scenario.

Basically, if everyone does their job right and the h2o2 and the tanks, valves, propellant delivery lines, etc. don't have any contaminants in them, then there would be no problem. We can't guarantee that obviously, nor can we guarantee that everything is perfect in terms of design and manufacture (all of which is why process safety engineers have jobs), and so we need to install pressure relief systems in the tanks.

Even in the case of no contaminants, we still need a PRV and a burst disc in case the nitrogen regulator fails and over pressurizes the tank. However, this is not the limiting factor on the sizing of the PRV and burst disc. If there was some rapid decomposition of the h2o2 due to contamination, in order for the tank not to detonate due to that increase of pressure and temperature and the presence of gaseous oxygen (which may, because of the high temps and pressures, have an interaction with whatever the contaminant is and cause a combustion), the pressure relief system must be designed for that case.


Again, sorry for the length. I hope I've clarified everything.

 

[georgeverghese]

Okay, sounds like the HTP in the tank flows over a catalyst to enable catalytic decomposition to O2 for use as an oxidant during hot fire test mode. The article from the 3 co authors states the use a silver catalyst if I remember correctly.

Since this is an exothermic reaction, and higher temperatures favor an increase in the rate of decomposition, it seems obvious that that both of the following should be enabled to prevent a runaway uncontrolled decomposition(a)dispose of high pressure O2 and steam evolved and (b) remove the heat of reaction by some means to prevent further increase in reaction rate. The PRV alone wont be adequate if you ask me since this is 98% HTP.

 

[MM904]

Georgeverghese,

Nailed it. We are using a different catalyst, but it is all the same idea.

Part a) can be taken care of by making the burst disc as large as feasibly possible.

Part b) can be taken care of with proper temperature monitoring and a tank cooling system. If the temperature inside the tank is X degrees higher than the ambient temperature, or if the rate of temperature increase is above some Y degrees/minute, then begin dousing the outside of the tank with water to help remove the heat of reaction. If the levels stay in unsafe rangers after that, then maybe it would be possible to initiate some "dump" procedure to remove the remaining liquid hydrogen peroxide from the tank. As to wheat we would "dump" the remaining h2o2 into, I haven't thought that far ahead. Maybe some large container that is open to the air?

 

[georgeverghese]

Presumably, the reaction vessel which enables catalytic decomposition is downstream of this holding tank, and unrelated to the accidental decomposition scenario in this holding tank.

So, for part (a), what would be the rate of decomposition you plan to use as the premise the RV or RD sizing? One safe way would be to assume the base reference reaction rate constant "k[sub]o[/sub]" in the first order reaction here to be same as that occuring in the reaction vessel downstream. Then use the Arrhenius equation to derive the reaction rate constant k[sub]t[/sub] at the max permissible operating temp you would allow. To be sure, it would be good if you got at least one other value for k[sub]o[/sub] from some other source. From this k[sub]t[/sub], you can then further get the reaction heat that needs to be removed in order to keep reaction rate constant / reaction temp constant.

For part (b), dousing the tank with coolant on the outside may be okay to remove small reaction heats, but for higher reaction rates, it may be inadequate.

 

[TiCl4]

MM904: For sizing relief vent requirements for reactions in which the kinetics are not completely or well understood, companies like Fauske perform a VSP2 test using material you send to them. The test will give you the maximum temperature and pressure of the reaction, and they can also use the pressure/temperature rise to size relief vents. They can also perform the test while relieving the material to a receiver tank to determine the flow regime of the two phase flow, if it exists.

I have used these services in the past to size runaway emulsion polymerization reactions.

 

[don1980]

Reinforcing the point made by TiCl4, you need to be very careful designing relief systems that have potential runaway reactive materials. That's exponentially true when you're dealing with an inherently unstable material like concentrated H2O2. This case absolutely calls for expert analysis and an expert relief design generated by a company such as Fauske & Associates.

BTW, are you sure that H2O2 is the right choice for fuel? A lot of people have died because of that decision, and I thought using H2O2 as a propellant fuel was a thing of the past.

 

[moltenmetal]

When diffusing a bomb, you hire a bomb disposal expert. When designing a system for reactive relief which, if not done properly can result in a detonation case, you hire a reactive relief expert. You don't learn by doing on this one.

 

[TiCl4]

Due to rather unique application here, I am wondering how this one turned out. Any updates, OP?

 

[Sabeau]

Have to agree with the comments above, for applications outside of the bread and butter blocked flow and fire cases outlined in API 520, these should be reviewed by a specialist consultant like the one mentioned above.

I believe Smith and Burgess and The Equity Engineering Group may offer these services.
I am not associated with either but I have done some of the E2G training and I found it quite useful.

From a manufacturer's point of view, it's better to see relief load analysis from the client themselves or someone they have engaged to conduct the study for them.

 

[MM904]

An update, as requested by TiCL4, and wrapping up other loose ends:

1) We fully designed our own system (minus the portion requiring input from Fauske's vent sizing service), and once that was finished we contacted Fauske. The team at Fauske was great, and provided us good feedback and two quotes - one for vent sizing, and another for an overall safety review. We decided both of these were necessary for safe operation and took the quotes to the department head, at which point it unfortunately was deemed too expensive to continue with the project.

2) The use of HTP as an oxidizer is not standard in industry for liquid-fueled rockets because it is relatively less powerful than straight liquid O2. However, it is in and of itself a great oxidizer - the devil is in handling the stuff. For more on HTP, please find a copy of Ignition! it is a highly entertaining and educational book on the development of liquid fuels and oxidizers. In our design, maximum performance was not the #1 requirement, and other characteristics of the oxidizer (a non-exhaustive list: relative stability at high concentrations, its ability to be a hypergol, more dense than liquid O2, and no need for cyro prop delivery systems as with liquid O2) drove us to selecting HTP as the oxidizer. Unfortunately, we clearly erred in neglecting to do extensive research on the safe handling and storage of high concentrated hydrogen peroxide - if we had done that homework more thoroughly, our selection might have ended up being something less awful to work with.

3) The team has shelved the engine design that uses HTP and is working on applying the lessons learned to other less powerful variants to further develop and refine competencies needed in liquid rocket engine launch procedures as a way to keep moving forward until more funding from the department is available.


Special thanks to georgeverghese and and TiCl4 for your help! Happy New Year to all.