9 replies to this topic

[precog1906]

Hello All-

First I would like to say that I have been very impressed with this site, and I wish that I would have found it sooner. My current problem:

I have been tasked with verifying the calculations that another engineer performed on relief valves in our plant to make sure that they were correct. As a new engineer I assume everything is wrong until I can verify it directly, but this is giving me some problems.

The table gives the following data but is incomplete: Process Pressure, Process Capacity, Relief Pressure, and Relief Capacity. For the example below the

Process Pressure =45
Process Capacity=
Relief Pressure=50psig
Relief Capacity=4398

For one of the valves for example he has :
Safety Relief of Stripper Preheater rated for a capacity of 4398lb/hr

He then calculates the total pipe equivalent (including all elbows and tees)

Using the formula for Pressure Drop which is Delta P =(Lvgm^2)/(0.08D^5) where Delta P is pressure drop=13.8 bar, L is length of pipe=35.7m, vg is specific volume of steam=.186m^3/kg, m is mass flowrate, and D is Pipe diameter=50.8mm. Solving for mass flow rate we have 2770lb/hr Therefore a flow of 2800lb/hr would encounter enough friction to completely dissipate 200psia of pressure

My question is should the calculation be more involved? When does pipe equivalents need to be calculated?

I know this was pretty lengthy, but any advice will be grateful. Thanks

Looking for "Relief"

[astro]

My advice would be start with API STD 521 and give section 5 Determination of individual relieving rates a good read. Have a think and come back to the forum.

Where I lose you is what relief contingency does the data that you've posted relate to? Once you've nailed that, then you can go about determining the relief rate required for that contingency.

Here's a non-exhaustive list of contingencies that you could consider:
1 Power Failure - Site (Total)
2 Power Failure - Local (Partial)
3 Steam Failure - Site (Total)
4 Steam Failure - Local (Partial)
5 Cooling Water Failure - Site (Total)
6 Cooling Water Failure - Local (Partial)
7 Instrument Air Failure - Site (Plantwide)
8 Cooling Failure (e.g. loss of refrig., loss of fin fans)
9 Block Valve Misoperation (inadvertent closure or opening)
10 Operator Error
11 Failure of Automatic Control
12 Liquid Overfill
13 Blocked Outlet
14 Reflux Failure
15 Exchanger Tube Rupture
16 Chemical Reaction
17 Hydraulic Expansion
18 External Fire
19 Check Valve Leakage
20 Abnormal Heat Input
21 Accumulation of Non-condensables
22 Entrance of Volatile Liquid
23 Internal explosion ("Dust" systems)
24 Check Valve Failure
25 Control Valve and Bypass 100% Open

You need to work up your own list to suit the specifics of your process as they apply. You'll also note that quite often you might find overlap, e.g. a cooling failure could be the same as local power failure such as for electric drive fin fan coolers.

Once you've considered matters come back and tell us what's relevant. If the contingency is a pool fire, then the relief rate is a function of vessel area. If the contingency is a control valve failure then the critical parameter is valve Cv ... and on it goes. With most of the relief sizing work that I've done, the approach has been to reduce the problem in question to identifying the flow bottleneck or source and disregard a detailed assessment of line resistance on the basis that the bottleneck / source is the rate determinant, which simplifies matters and in turn provides a conservative answer. There are always exceptions to the rule but if you've got significant dPs it stacks up because you'll hit sonic flow at the restriction. Or if you've lost coolant or have a fire, you'll need to dispose of the vapour evolved otherwise this inventory will accumulate and result in a pressure rise.

While your mathematics may be accurate, your information doesn't give me a sense of the relief event. Please expand and provide a sketch with info of the pressures in your HP system at the relief condition and the system pressure downstream of your relief device.

I think though that once you've addressed the points raised above, you will have resolved your problem yourself.

[Qalander (Chem)]

Dear Astro,

undoubtedly very good and accurate reply.

Regards
Qalander

[astro]

Qalander, thanks for the vote of confidence.

Precog, some other points for you, how has this RV calc been set out?

If the methodology and referencing is not formatted clearly, then being a 1st timer at checking RV calcs you've got a challenge on your hands. I suspect that this is the case given the nature of your query.

There are many ways to address calculation presentation. Ideally it should be a self contained document that provides the background to the problem, the objective / purpose of the calc, referencing to document the sources of data used and draw some conclusions along with the worked up content. I like to see a summary of these critical points on the 1st page. This helps when searching for information after the calc is complete.

I've attached a sanitised, cut down example to give you an idea of what to look for. If I was in your position, I personally would reject a calc that didn't come up to a similar standard. If you're being relied upon to make assumptions in your head as you check the calc because of poor referencing/citation, then you need to put the brakes on.

This is even more important for a relief valve calculation. I hope that you've got a senior engineer, well versed with RV calcs, reviewing / approving your check work. Safety device design deserves additional scrutiny given the potential consequences of an error passing through to plant operation.

Attached Files

•  Example_Calc.xls   34.5KB   214 downloads

[precog1906]

Thank you so much for your reply and it makes perfect sense. I will have a senior engineer reviewing my findings, but I think this is a "test" for me. My initial thought process was that the steam to the system is supplied by our boiler, which we operate at 150 psi. All of the equipment with relief valves are supplied off of a steam header from the boiler. So with the equation described in my first post, I assumed that at worst case total pressure loss do to friction in the pipe could not exceed 200 psi. I then used this as my pressure drop in my formula, found the equivalent pipe distances (L), and then solved for mass flow rate.

After I had my mass flow rate as long as my relief valve was rated for a larger capacity, then there was no way that my process could exceed that.

Was that not a correct assumption/approach to the problem?

[astro]

Inlet line pressure drop is very important to ensure functional operation of the relief valve (check the standards API RP 520 & STD 521 for the discussion on valve chatter and inlet line losses). If memory serves, a maximum of 3% inlet losses [from the pressure source to the relief valve] is the criteria (one for you to check). Your assumption about a 200psi delta-P looks a bit big to me for a boiler normally operating at 150psig.

However you still haven't identified the relief contingency. From the additional info you've provided I suspect that "blocked outlet" will be the sizing case. The relief rate is then the maximum continuous rating of the boiler, which is a function of the heat input at the maximum firing rate or max. process duty (for a waste heat unit).

You still haven't convinced me that you understand the concept of relief contingencies and individual sizing cases.

What is your boiler's steam drum design pressure (DP) or maximum allowable working pressure? The steam system's normal operating pressure is important to be aware of but for this kind of work, the DP is where you need to be working from because this is the pressure at which the relief valve will begin to operate.

You mention over capacity of your RV. Take care not to oversize the valve. This is the importance of identifying all relief contingencies and has been one reason why I've stressed this point. If one credible contingency will result in a relief rate significantly smaller than the design case, the RV may exhibit "chatter" for the lower flow rate. The variation in flowrate justifies the use of multiple valves (again a matter for your personal reading to review and digest).

[precog1906]

Okay maybe I am way off base. I will take a specific case for example the mineral oil heater. The relief contingencies are liquid overfill and blocked outlet. The relief valve has already been sized for this equipment again I am only verifying that it is adequate. The relief valve is rated for a steam capacity of 5076 and set at 60psi and located directly on the top of the heater venting to atmosphere. We operate our heater at 25psi. I do not have the DP information handy for the Heater or the boiler, but I can retrieve the info. Again I assumed max pressure loss, but as in your last response that may have not been the right thing to do. Where am I going wrong? Lets assume that I had the DP information for the boiler and heater, would equivalent pipe distances not come into play? What should be my next steps?

I know it may seem like you are "going in circles", but thank you so much for your help and understanding thus far

[precog1906]

I guess if I just had relief valves on all of my equipment to handle the maximum operating conditions of the boiler, everything would be great. i would then be stuck with all oversized RV's and plenty of chatter.

[skearse]

I think what astro is trying to convey, correctly, is that sizing a relief valve is not simply an exercise in calculating pressure drops, maximum flows, etc. based simply on the physical equipment. PLEASE review API 520 and 521 before going any further. In order to size (and verify) you must take a look at any and all potential scenarios that could cause the pressure in the vessel to rise above it's MAWP (or DP as astro referred to it). The 25 contingencies that astro listed are a great start, but you must think about the system as a whole and realize which potential upsets would apply to you situation (and you also need to identify any that are NOT in that list). You must then perform the calculations for EACH of those scenarios, and design an overpressure relief strategy that makes sense for your situation. For example, you state that liquid overfill and blocked outlet are applicable cases. For those, you must look at the pump curve (or whatever would cause the overpressure) and determine what the flow rate through the relief vavle would need to be to prevent the heater from over-pressurizing. You ALSO need to confirm, that other scenarios would not cause an overpressure situation. For instance, can the boiler produce a steam pressure that is higher than the heater steam side MAWP? Can the heat input from the boiler provide heat input that would vaporize the oil and cause the vessel to overpressurize? Could an external fire do the same? Each of these must be considered individually, accounted for, and then, and only then, would you be able to say that yes, the valve(s) are adeqautely sized or that no, they are not, and recommend the appropriate changes.

[Andrei]

precog,

I don't have to add anything more technical to what my predecessors did, I totally agree with what they recommended.
I only have one thing to add, and I would like to start with a quote from your first post:
"As a new engineer I assume everything is wrong until I can verify it directly, but this is giving me some problems".
I think this is not the right approach when working in a professional environment. I know that a lot of people have this kind of approach, I've did the same thing at the beginning.
I think you should always assume that you are wrong, and the other one is right, and if something seams wrong, you are the one that don't understand what was the intend of the other person.
It is always much simpler to criticize other's work than do it yourself, and you should always show respect to others' work.
The first sign of knowledge is when you start to realize how little you know.

Good Luck