| Applies To | |||
| Product(s): | PULS | ||
| Version(s): | XM & V8i | ||
| Environment: | N/A | ||
| Area: | modeling | ||
| Original Author: | Bentley Technical Support Group | ||
Comments, Questions, and Answers about modeling with Bentley PULS
Items #1:
**Attention**
Please see the following PULS help section, Help > Tutorial>
Or
Locate the following file on your computer: C:\Bentley\PULS XM\PulsTutorial.pdf
This help has been provided in order to give users ideas for modeling typical piping arrangements. The steps shown in each example should not be taken as the only method available to create models. In addition, the intent of the examples is to present ways to create adequate models for analytically purposes.
Tutorial #1 : Acoustic Mode Shapes
The purpose of this tutorial is to introduce the basic operations of the PULS application. The tutorial will show you how to produce acoustic mode shapes of a simple system. It will walk you through the various steps in model building. You will define the piping system by specifying system geometry, fluid properties and boundary conditions. You will learn how to view and print your results.
Tutorial #2 : Modeling a Reciprocating Compressor
The purpose of this tutorial is to introduce the basic operations of PULS to model a reciprocating compressor system. This tutorial assumes that you have not completed Tutorial 1, therefore, some of the information may be repeated.
Item #2:
The following model files are included with PULS :
2CYL.LP1 - TWO CYLINDER COMPRESSOR EXAMPLE
COOLER.LP1 - GAS COOLER
EX-BOT.LP1 - SUCTION BOTTLES AND COMPRESSOR
EX-SUCT.LP1 - SUCTION SYSTEM
METER.LP1 - METER RUN : Globe Valve Source
PISTON.LP1 - Piston (Volume Velocity) source.VS = 0.01 m3/sec, Natural gas/ sound speed = 494 m/sec
PRESURE.LP1 - Press source for a branch system., PS = 10 kPa, Natural gas, sound speed = 494 m/sec
PUMPFIX.LP1 - PULS models involving a reciprocating triplex pump
PUMPSUC.LP1 - PULS models involving a reciprocating triplex pump
RECIPP. LP1 - RECIPROCATING COMPRESSOR EXAMPLE
T1. LP1 - Pulsation Source: T1 - Mode 1
T1-1. LP1 - Pulsation Source: T1 - Mode 1
Items #3:
We would like to know if PULS (Option 3) is suitable for modelling the shaking forces in dead-leg side branches from vortices generated by a corrugated pipe with flow. We already know the amount of energy and flow conditions which the vortices generate in the corrugated pipe, so PULS does not need to calculate them. We just need a way of putting in an energy source in the (e.g. middle of the corrugated pipe). We have considered using a suitably scaled velocity source or pressure source, but both of these would impose reflective boundary conditions (like an open or closed end) which do not exist in reality. Is there a way of modelling this in PULS?
Answer:
Suggest to try and insert a tee point on the tee branch very close to tee point. The length would be short and you would add a velocity source there. They can scale the source to give a known pulsation at the dead end if known.
Item #4:
When I try to adjacent flow directions, I mess the pulsation, and when I go to adjust the pulsation, it messes the flow direction. Have been doing that for the last hour. Is there anyway to avoid this interactions ?
Answer:
These should not interact usually. Send model for review.
Item #5:
If I have two nodes which I would to turn them into one, what do I need to do to equate them?
Answer:
Delete the element joining them and then you update one element and change the node number to be same as other node.
Item #6:
Does adding a Volume element have the same acoustic effects as adding an enlarged pipe say that is having the same volume?
Answer:
Not exactly. The pipe element has a length and hence it has some possible frequencies. But these may be ignored since they are usually high, but not always. It is more accurate to model using lengths since most vessels or passages have lengths.
Item #7:
Could you provide some instruction on how to make liquid pump model and check it? could PULS support liquid's model building and checking?
Answer:
PULS does not have a centrifugal pump. If you have to model it I would use a pipe or reducer or something to simulate volume or length. Noise cannot be simulated from centrifugal pump. If you have to, you may try a velocity source at a dead end near the pump. You may create a fictitious short branch and add a velocity source at the dead end. For reciprocating pump you can choose the recip boundary condition and works the same way as recip compressor. We have two pump examples pumpsuc and pumpfix and they are documented in the comparison to experiment and on chapter on liquid modeling in the PDF manual mentioned in #1 above..
Item #8:
Adding side streams - single system with multiple gas compositions
Answer:
Please see the following procedure:
1. Select Tools> options> change fluid to "user defined".
2. Make a selection of piping
3. Press Modify++> properties over range> fluid> enter a new fluid property or select an existing one from the list provided.
Item #9:
Designing multi-chambered bottles:
Answer:
See attached image for modeling a muffler with a choke tube.
As you already know, bottles can be modeled as a combination of pipe, two-port, and three-port elements. In addition, you can also use a half-pipe or half-vessel(e.g., for the internals of a pulsation bottle). Suggest to model the multi-chambered bottle as designed with components mentioned above, you may need to add a small amount of length to some components to be modeled correctly. See example model C:\Bentley\PULS XM\Project\ EX-BOT for additional suggestions.
Item #10:
When there are multiple compressors in a single acting condition. If you choose closed for the single acting side (head end or crank end) you must also have a value for the flow. When the machine is single acting, there is no flow through one side of cylinder. So, do you leave the single acting head end selected as unloaded when the corresponding crank end side is selected as closed?
Answer:
When a double acting compressor is in a single acting mode, one side is unloaded while the other is loaded. Unloaded for discharge is equivalent to closed. Unloading in suction is a very severe scenario and it is full sinusoidal piston motion with no flow. The flow comes from the loaded end. For single acting compressor, it has only one end and would be set to either loaded or closed.
Again, Loading and Unloading can only be used on multiple cylinder compressor. This technology cannot be applied to a single cylinder compressor. Unloading a compressor keeps it from shutting off when the demand is less than full capacity. Otherwise, the compressor would turn on and off frequently causing more wear on the equipment than continuously running.
Bentley technical support does not have any example models. However, the closest are the muffler models in the comparison to experiments in the online help.
Item 11:
Does the program model reciprocating pump / compressors?
What about multi-cylinder and multi-pump/compressor scenarios?
Answer:
Puls can Handle Centrifugal and Reciprocating Compressors, Reciprocating Pumps, or the ability to accommodate a Pressure or Velocity source.
Multi cylinders are handled nicely using phase angles between cylinders. For multi-compressor units, each unit is analyzed separately assuming the others closed or inactive. User then has to combine the responses manually for the most conservative response since phase angles are unknown. Since this approach is time consuming, many assume all compressors are in phase initially and they may evaluate phase information for just a few critical points.
See the example in the online help and additional information here:
Item #12:
In the online help, Valve Delay, but there is no way to enter this value in Puls dialog.
Answer:
The Valve delay option is on the Recip Pump dialog (see below). However it is only available for systems with Liquid contents and NOT gas contents.
Item #13:
How to distinguish single acting and double acting cylinder in Puls?
Answer:
See online help, search for "acting": and select the "Theory" topic:
For detailed pump modeling, PULS assumes each cylinder has a head end valve for single acting and a head end valve and a crank end valve for double acting cylinders. For the simpler pump nozzle model, you can enter the pump type, such as triplex as described later in this section. In this case, the pump detail passages are not necessary and the cylinder number refers to the unit number, since every boundary condition can have up to nine single acting cylinders.
Item #14:
Will PULS analyze systems based on API 618?
Answer:
Puls has the 4th Edition API 618 Guideline with 1992 NOA shaking Force Guideline and the 5th Edition API 618 Guideline
Item #15:
Can PULS do different gas streams with different mole weights in the same system?
Answer:
No, in PULS you define 1 gas stream with a fixed composition.
Item #16:
Can PULS model multi chambered bottles?
Answer:
Yes, see online help, search for "Comparison To Experiments"
Item #17:
Can PULS calculate pressure drop across an orifice?
Answer:
Yes, PULS calculates the pressure drop across an orifice.
Item #18:
Can PULS calculate flow velocity in choke tubes?
Answer:
Yes, see online help, search for "Comparison To Experiments". The velocities will be calculated for each component in the model.
Item #19:
How many unique flows and gas streams can PULS handle?
Answer:
In a single model, PULS can handle only 1 composition, but at various temperatures, pressures, the flow can be split with tees.
Item #20:
Can I import an AutoCAD file into PULS?
Answer:
AutoPIPE can import or insert the center-line from AutoCAD and then create a neutral file that PULS can read. Therefore using AutoPIPE as a translator, Yes Puls can import an AutoCAD file.
Item #21:
I could not connect both suction and discharge pipes to a cylinder
Answer:
Suction, inter-stage and discharge are treated in separate models.
Item #22:
I could not set up a simplified model with head end and crank at pulse at one node
Answer:
Head and crank should have separate nodes. Compressor passages between these valve points are very important to the solution for compressors.
Item #23:
although PULS has a graphical user interface I get very little control over the model and input parameters. It is beyond the scope of my evaluation to summarize it all. only resonance graphs for existing nodes, so in a simple model with just 2 nodes and 20 meter pipe there is no usuable graph and I have to add additional nodes. However a user does not think in nodes but in elements (like an elbow or tee).
Answer:
Intermediate nodes can be added use the split element feature.
Item #24:
it was unclear to me how I set up a calculation at nominal design conditions and then vary for instance the speed of sound within a certain range (let's say +/- 20%). Has this to do with the fact that PULS is limited to the frequency domain?
Answer:
Speed of sound can be varied by increasing range on compressor rpm.
Item #25:
For liquids interaction effects are more pronounced, but have been ignored due to complexity of such analysis. In addition, we are concerned about one point in that the mode frequencies of the pipe deformations may depend on the speed of the flow within the pipe. This is a very important point for us since we operate our pumps over a range of different speeds during our operations. With this said, I was wondering if AutoPIPE and PULS have the capability of modeling the phenomenon I just stated.
Answer:
In PULS the mean flow is a parameter in the transfer matrix formulations and it would affect frequencies as well as damping. AutoPIPE ignores flow velocity effects in determining the mechanical resonance frequencies.
Item #26:
Does PULS have the capability to carry out frequency response analysis on a pipe network that is being forced by a pulsating flow such as a positive displacement pump?
Answer:
YES
Item #27:
According to the PULS brochure, PULS contains ANSI/ASME or DIN standard libraries; however, I was wondering if I could create custom objects to these libraries from data that I have?
Answer:
These libraries are for pipe sizes, no custom libraries possible, but you can easily enter any non-standard pipe size and thickness.
Item 28:
To what extend are the reciprocating pumps in PULS able to model the behavior of a positive displacement pumps? Can I model a triplex (i.e. a pump with 3 plungers) and a quintuplex (i.e. a pump with 5 plungers)? Can I adjust the stroke and bore diameter of the pumps so that I can mimic the volumetric flow rate more accurately? Do the models of the pumps take into account valve ringing (i.e. the effect that the valves in the positive displacement pump don’t open and close instantaneously)?
Answer:
Yes you can do all that. We have duplex to nanoplex pump conditions. You can also model pump internals using head and crank ends for each cylinder for more accuracy.
Item #29:
Is the capability of importing CAD models from AutoPIPE included by default in PULS?
Answer:
Yes
Item #30:
Can I import CAD models using STEP or IGES files into PULS?
Answer:
No.
Item #31:
Is pulse capable of modeling the effects of cavitation in piping?
Answer:
No, but you can check if cavitation will happen to avoid it.
Item #33:
Also as I have asked for PULS, can one carry out frequency analysis on AutoPIPE to determine the structural effects of acoustic forcing functions on the piping network?
Answer:
Yes, shaking forces from PULS can be imported as harmonic forcing functions to perform harmonic frequency analysis.
Item #34:
Do either AutoPIPE or PULS have the capability to perform FFT at a particular point on the pipeline so that the contributions of different harmonics in a pressure or acceleration/strain signal can be interpreted?
If this is not the case, can simulated pressure, acceleration, strain, and stress vs. time data can be ported from AutoPIPE and PULS to other programs such as Excel or MATLAB to carry out FFT Analysis on the signal.
Answer:
In PULS FFT is done to create pump harmonics and then IFFT is performed to show time domain results for pressure/velocity and forces. There is no FFT in AutoPIPE, but AutoPIPE takes FFT components in the harmonic shaking forces and solve for the harmonics and adds the response.
Item #35:
The group I'm working for is searching potential software suites that would help us in analysis failures of high pressure piping (20,000psi) in our stimulation operations of hydrocarbon-wells.
The analysis that we would like to do involves both the fluid acoustics in the piping (caused by a set of positive displacement pumps) and the structural behavior of the piping as a consequence of the fluid acoustics. In addition, we would like to study the effects of the fluid-structure-interaction to determine whether the interaction is "one-way" or "two-way" - "one-way" the fluid acoustics influences the piping structure but subsequent changes into the piping structure do not influence the fluid acoustics (i.e. feedback does not occur) whereas "two-way" fluid acoustics entails feed-back. With that said, would you happen to have further information whether the dynamic analysis in AutoPIPE and Pulse provides an avenue for two-way FSI or is the software just capable of analysis piping networks with just one-way FSI?
Answer:
The analysis in PULS/AutoPIPE ignores interaction, so it is one-way FSI.
Item #36:
How to Cut, Copy, and Paste in Puls?
Answer:
At this time these features are not available in the program because they were not thoroughly tested.
These options can be enabled and used at your on risk. If you accept this, perform the following procedure:
(1)Exiting PULS
(2)Editing the PULS.INI (Configuration settings) file in the C:\Bentley\PULS XM folder using NOTEPAD
(3)Setting the flag: Enable Copy/Paste to True. Save and close the file.
(4)Start PULS and notice Cut/Copy/Paste icons are enabled
(5)Read about it in online help "Command reference/Edit Menu" Cut/Copy and Paste
The feature is stable and was informally tested and many issues fixed. But no test plans was developed or executed.
Again, use this feature at your own risk.
Item #37:
How to input variable pressure or velocity source file in the program.?
Answer:
The variable pressure file needed to be in text form each line having frequency Hz, pressure kPa and Phase angle degree, separated by one space.
Item #38:
How to model a Cross pipe component in the program.?
Answer:
Unfortunately at this time you cannot model a cross or a 4 leg tee in PULS. The only workaround is to model a tee and then another tee very close to the first.
Item #39:
When using the "Variable Velocity Function" boundary, we can only get the volume velocity as a function of Frequency. But the curve is continuous. This is quite different from the "Reciprocating Compressor" boundary. How to explain this situation?
Answer:
Two approaches to boundary conditions like physical condition i.e. reciprocating compressor vs. theoretical condition i.e. volume velocity came from different mathematical modelling. We need to remember that we work with models only and this is why only simplified approach can be applied. PULS solves series of equations for specific frequency, so even if you think it is an analytical solution in continuous frequency domain - it is not.
Results are a series of analysis in given range of frequency, with assumed step, that are combined. In case of Reciprocating Compressor it works the same. The only difference is that theoretical boundary condition is calculated by the program in a 'preprocessing' phase.