Allowable loads on pumps and taks 4

[Eduard47]

After reading all the issue regarding pipe stress analysis, I have this guts to ask help from those expert on the subject. I am going to perform a stress analysis on a 12" line, 400 psi/450 deg F with the layout not known to me yet other than it is a line interconnecting two skids, one with the pump and the other with a tank (reactor). If i run this in the program, what are the allowable loads maximum that can be transferred to this equipment? Assume by now fluid is hot water and the pipe is stainless steel. Are their any preferrable type of expansion joint I can use if in case I need one? I am very thankful if somebody can give me references. Thanks in advance.

Eduard47 (Process)

 

[KernOily]

How do you run the analysis without first knowing the layout of the line?

At any rate - the allowable loads for the pump will come from the pump manufacturer.

The allowable load for the tank is another matter. Is this an ASME pressure vessel? Is it an API tank? You might be able to get the allowable nozzle loads from the tank manufacturer. Otherwise you have to calculate them using the methods given in the Code. This is not a trivial matter. There are a few other posts on this topic here in the forum; I'd suggest you do a search for them and you'll find a lot of good information.

An expansion joint is the last resort. There are other ways to ensure adequate flexiblity of the line before resorting to an expansion joint: expansion loops, Z-bends, braided SS hose sections, thin wall pipe sections, etc. Thanks!
Pete

 

[Eduard47]

Thanks a lot for a quick resonse Pete and I will follow your advice to start with. I maybe able to know all the details of the layout today. Thank you so much.

Eduard47

 

[StressGuy]

I must confess that I am concerned by this thread. After having followed this discussion:

thread765-14405 I feel that I at least need to ask some questions.

Is this going to be your first stress analysis? A pump system is not usually a good way to start out, particularly if it has much temperature to it, and stainless steel presents an added challenge as it has a much higher expansion rate than carbon steel.

I would agree that expansion joints are something you'll likely want to avoid. Almost every client I've worked for seems to require approval from the CEO, board of directors, and a majority of the shareholders to put an expansion joint on a pump nozzle.

I would recommend, at the very least, that you have your analysis reviewed by an experienced engineer before implementing the design. There are a lot of tips and tricks that come from working under the supervision of those who've been doing this for a while. After 6 years myself, I'm still finding plenty of new things to learn.

So, while I'm not necessarily trying to scare you away from stress engineering, I do want to caution you that plugging numbers into a program is niether analysis nor engineering - that part comes when you see the numbers and have to decide, first, are the results real numbers or some kind of computer enhanced hallucination, and second, what can be done to refine the system to perform as it needs to.

If this is something that your company is pushing you into becuase they don't want to go to an outside company for the stress work, you should give it a lot of thought before you decide to take responsibility for this.

If all of this is off base, I apologize, but the nature of your question made me nervous, so I feel the need to at least raise the questions. Edward L. Klein
Pipe Stress Engineer
Houston, Texas

All opinions expressed here are my own and not my company's.

 

[Eduard47]

Thanks for a nice reply Ed. Actually I was doing this job for the last 2 years (say 80-90% of my time). The allowable load transferred to the equipment has been an uncomfortable scenario to me simply because I have no criteria on hand. I have no Caesar seminar and even with it, I am not sure if it is a part of the seminar. My tutor has told me to play with the exp. jt., piping layout and its support to have a load no more than 150 lbs - something that is hard to achieve most of the time. At one time, a cooling tower manufacturer of the chilled water line we are modeling limit us to no more than 100 lbs transfered load to connecting nozzle. So, are allowable loads on nozzles of almost every type and size of process equipment always available from the manufacturer? Or, is there any general rule in predicting this figures? Are there any reference I can use regarding this subject? Thanks Ed, and to all the generous guys out there for being so helpful.

Eduard47

 

[gunnarhole]

Gentlemen,

Tony Paulin (Paulin Research Group in Houston) noted in a recent paper that he had attended a conference of industry experts where it was proposed that formal calculations be prepared whenever:

Thermal Cycles)/(Age min) > (750 F - Toper)

Where:

Age min = Minimum Age of Piping Checker, Analyst, or Plant Engineer Responsible for System Operation

Multiply (Thermal Cycles) by 2.0 if Pressure is greater than 3000 psi.

Multiply (Thermal Cycles) by 2.0 of GRE pipe.

Multiply (Thermal Cycles) by 2.0 if Dangerous Medium

Multiply (Thermal Cycles) by 4.0 if Pipe is Attached to Rotating Equipment

Multiply (Thermal Cycles) by 4.0 if Expansion Joint is used in the system.

I smiled when I read it but I don't disagree.

Regards,

Gunnar

 

[johnp]

As I read it you are working under a Tutor, have access to Caesar software but have not attended a training seminar on it. Which means that you have to dig out how it works from the manuals.

You have been given an assignment to do stress analysis on a 12" line, 400 psi/450 deg F with the layout not known to you yet, other than it is a line interconnecting two skids, one with the pump and the other with a tank (reactor).

1 - Nozzle loads. The most common statement of nozzle loads is in API 610 Table 2. Nozzle loads on equipment are not always available and some assumptions need to be made. In general the nozzle loads are likely to be less than the API 610 values.
If no information is available you can do some nozzle calculations using a pressure vessel code. With a 400 psi reactor vessel it would need to comply with Code requirements, which means that the manufacturer would have to have done calculations.
2 - At 400 psi you are looking at a massive pump to serve the 12 inch line. At one refinery we had 4 centrifugal slurry pumps in series to get up to that pressure at 450 deg F.
3 - To achieve 100 to 150 lb force at a reactor vessel is very difficult. The 400 psi internal pressure means that the components must be strong enough to resist high pressures, which means they will be stiff, and not react in the way you expect. Any dirt or scale inside can cause the units to jam up and they no longer act as expansion joints. It is not unusual for stress engineers to design in expansion joints, only to find production engineers take them out because they are scaled up and don't work anymore. They are replaced with steel pipe spools.
But since it is your assignment to play with them, by all means do.
4 - Now look at the pipe layout. The equipment is on skids. On the one hand this means that you need a flexible connection to the vessel and pump to allow for differential movement on the ground surface. On the other hand it means that the equipment is not rigidly fixed to bases, but will yield somewhat under nozzle loads. In the calculation even a fraction of an inch will help reduce the reactions.
Your best hope of controlling nozzle loads is in the design of the pipe run.
To get an idea of what you are likely to face choose a pump from a catalogue, choose a reactor of some type (say in a hydrotreater process), rough in some dimensions, and start thinking.
Where would you put pipe support? What type?
Where would the expansion joint go? Does it need guides to prevent jack-knifing the piping system? Where? What type?
That will encourage some realistic questions
about the problem. When the tutor gives you the equipment layout you will have some background understanding to work on.
The following are helpful
--- 1 - Busler & Paulin "Pipe Design for Robust Systems" Chemical Engineering June 1997
--- 2 - Bausbacher & Hunt "Process Plant Layout and Piping Design" Prentice Hall 1993

Fortunately you won't have to take out professional liability insurance on this just yet.

 

[ralphsare]

The piping nozzle evaluation is addressed in APi 610. If the pump has already been sized, then the pump manufacturer should be able to provide u with the allowable loads. For the nozzle tank, consult the API650 or wrc 297 and the supplement 107.

 

[Eduard47]

This is just a tremendous knowledge to learn and encouragement for me to read on responses from Johnp (mechanical). I apologize John for not having really clarify earlier that the tank here as somebody has described to me as a reactor is definitely not on petrochem application but on microelectronics facility application. I don't know if such tank is still under API or ASME. They are still finishing the design probably and I do expect to start on it anytime soon, that's why I am preparing very well on it. Let me ask something: if the tank is not classified as an ASME tank, will it be automatically referred to code and standard of API? Although I have have been taught on a direct "on the job training" since the last two years, on some of the design pointers here, mostly are "soul-filling" to me and I just hope this forum will be continually supported by generous guys like you, Elsinore, Edward Klien, Johnp, Ralphsare, gunarhole, and the rest. Many thanks to all of you and May God Bless you all.

Eduard47

 

[KernOily]

Ummm, please correct me if I'm wrong here, but doesn't API 610 only apply to API 610 pumps? I doubt you will see API 610 spec pumps in a microelectronics facility. API 610 pumps are very stout and it would be an error to apply those allowables to, say, an ANSI pump.

The very first thing you need to do is read the flexibility and allowable stress/design criteria sections of the B31.3 code, and/or the texts mentioned above. This can't be overemphasized. It absolutely irks me when I see folks around here running CII or one of its competitors without knowing the first thing about, for example, how CII calculates its allowables. Talk about putting the cart before the horse. CII and its competitors are very easy to run. A trained chimpanzee could run them. The catch is (1) making sure your model is an accurate representation of physical reality and (2) making sure you know how to interpret the results and then working with the designer to ensure your recommendations are accurately implemented. I made a lot of mistakes in the past and learned some things the hard way; fortunately no one got hurt (or sued...). Thanks!
Pete

 

[Eduard47]

Got you too Pete and have hard copies of these messages and B31.3 on my desk. More power to you.

Eduard47

 

[johnp]

Now we start talking business

1 - You are working on a microelectronics facility. Find out what Engineering Codes you will be working to. Your tutor should be able to advise.

If not, then you have to make a statement that for this Project you are using the "following Standards and Codes". Your Caesar package will give you a wide choice as a guide.

2 - For Pump design API 610 is a convenient summary of acceptable nozzle loads. You are not working for the petroleum industry, but the API standards are often applied for the design in other industries - e.g. mining, because there is no other suitable standard (For example API 650 is often used as a design criteria for dry bulk storage tanks. Yet when you read the API 650 it says nothing about bulk storage, and specifically excludes such applications in the Code).

3 - Your vessel is being designed. The designers must know what the design criteria are. It does not have to be API, or ASME, but if they are going to have the vessel registered with the Inspector of Machinery, and Insurance Company, it must comply with the local legal requirements. So find out what they are.

It could be the Japanese Code, Naval Code, British Code, and so on. Nothing is automatic unless stated in the legal requirements, or in the Contract Documents.

4 - Then pay attention to the advise of 74Elsinore. The strange thing about Caesar II is you input numbers, it will spit out numbers. But do they make sense? Many times they don't. That's where you need someone of 74Elsinore's experience to help guide you.

5 - Finally, if you do go ahead, find out who takes the professional liability. At 400 psi & 450 deg.F you are looking at a lethal weapon. If you want to sleep at night recommend that your company employ someone like StressGuy.

I've had guys who have had years of experience running Caesar II, who took one look at some of our stress problems on site - and ran away. The possibilities for something going wrong frightened them too much. I've had others who will run the Caesar calculation, but don't want to get involved with the practical interpretation. They know the theory, but lack understanding of the real engineering decisions and how they are made.

 

[ralphsare]

I've had guys who have had years of experience running Caesar II, who took one look at some of our stress problems on site - and ran away. The possibilities for something going wrong frightened them too much.

wouldnt professional ethic dictates thatknowing something is very wrong with the piping layout should at least bring it to the attention of the owner of the plant instead of running away. if the develop stress is unusually high, there are ways to bring it down to an acceptable level. CII is just a modeling tool, layout will still undergo review by the head honcho of the piping design group who play with the iNtergraph then submitted to the fabricator and handed to the piping contractor for installation. surely along this path, a soln can be formulated that will satisfy the code.

 

[StressGuy]

Now this is what I really love about this forum - a group of people dedicated to our profession who share their experience and ideas so that we, hopefully, don't have to reinvent the wheel from generation to generation.

I'm not sure what piping code is going to apply in a microelectronics facility. B31.3 is certainly going to be a useful general resource, but you'll definitely what to check with your supervisors as to what the governing code is.

In truth, it is rather difficult to analyze a pipe routing that doesn't yet exist. However, you have a golden opportunity that rarely comes along for a stress analyst - involvement in the early stage. Since the layout has not been made, get involved with the guys doing the equipment layout, and see what can be done to make the the tank and pump skid arrangment advantageous to solving the problem. Most of the worst stress problems I've had to work through are the ones where a layout guy has found the worst possible arrangment for pieces of equipment that interact with one another.

One other item to consider - how with the pump skid be piped up? If your group is responsible for piping up to the pump nozzle, then you will be responsible for satisfying the vendor nozzle load requirements. However, if the skid vendor is piping from the pump to then edge of the skid, they will be the ones responsible for the pump, and you may not even get a detail drawing of the pump itself to determine the allowable loads. In that case, the skid vendor will have to supply you with allowables at the interface. Although, in reality, you're going to probably have to beat the routing out of the vendor so you can analyze it yourself - in my experience, most of these groups that design and build skids have very little understanding of stress, nozzle loads, etc. and will probably have put very little thought into the routing of the skid piping.
Edward L. Klein
Pipe Stress Engineer
Houston, Texas

All opinions expressed here are my own and not my company's.

 

[ralphsare]

let us say that the location is an area that follows the ASME code. For piping, you can forget .4 thru .11. that leaves u with either the b31.1 or .3 since its not a power palnt, then the nearest code that will apply would be the one for the process industries (makes sense).

For the tank or vessel or drum inquestion, youre choose is ether the API or B&pv Code. but if youre in location that does not adopt any industry code, then code compliance is the least of your worry .

 

[johnp]

ralphsare,
We were working on upgrading a plant processing mineral sands. My area was Digestion. Sixteen vertical shell and tube heat exchangers in two streams. Live steam at about 400 psi, 450 deg.F came in at the hot end. Exhausted flash steam was then passed on from unit to unit. The last unit was running just above atmospheric pressure.

We hired a fellow experienced in Caesar II, and asked him to model the flash steam piping system because we were making modifications to the interconnecting piping. His background must have been in industrial work. He had never worked with 18 to 24 inch steam pipes before. So he just packed his bags and went. We had to go to a Stress Consultant to help us do that analysis.

We hired another experienced Caesar II fellow. He did the modelling OK, but did not get involved in the detail of what to do next, to reduce nozzle loads within acceptable limits.

Then the management decided that they had spent a big bundle of money on stress calculations, and cut all further analysis. It wasn't a very happy situation to be in, but then mining is a lot less demanding in these areas than petrochemical work.

One of the curly aspect was that we had a live plant, which had been running for years as a full scale model. All the drawings were done by hand. The stress calculations were probably done way before Caesar I, when you had to put your data on punch cards, and take them to a central Mainframe computer service.

The Caesar models showed high stesses, so it was a challenge to reconsile the calculated values with the measured values in the field. You have to use all kinds of tricks - local stiffnesses, friction at supports, and so on to come up with a model which reasonably reflects reality. Fortunately pipe stress is fairly forgiving. It is flange connections which give the trouble by leaking from thermal movement.

The plant upgrade has been installed now for some 3 years and from all reports is working well.