valvole di controllo e intercettazione, sistemi di ... gds valvole ati aprile... · valvole di...

21
VALVOLE DI CONTROLLO E INTERCETTAZIONE, SISTEMI DI AZIONAMENTO, DISCHI DI ROTTURA E DISPOSITIVI DI SICUREZZA UTILIZZATI NELL’INDUSTRIA DI PROCESSO Milano, 18 Aprile 2018 Auditorio TECNIMONT

Upload: truongcong

Post on 14-Feb-2019

218 views

Category:

Documents


0 download

TRANSCRIPT

VALVOLE DI CONTROLLO E INTERCETTAZIONE, SISTEMI DI AZIONAMENTO,DISCHI DI ROTTURA E DISPOSITIVI DI SICUREZZA UTILIZZATI

NELL’INDUSTRIA DI PROCESSO

Milano, 18 Aprile 2018Auditorio TECNIMONT

FLOWSERVE VALBART – FLOW CONTROL DIVISION

Flow Coefficient validation through Computational Fluid Dynamics (CFD)of Control Ball Valves for Oil&Gas applications

Abbiati Luigi

FLOWSERVE VALBART – FLOW CONTROL DIVISION

How & where to get all those numbersfor an entire product range?

Numbers matter

• Correct sizing & selection of a control valve is a keyactivity, that relies on an extensive database of values,coefficients and parameters.

Instrument datasheet Control valve

FLOWSERVE VALBART – FLOW CONTROL DIVISION

Experimental testing

• Standards IEC 60534-2-3 or ISA 75.02, define testingprocedures, test rig arrangement, measurements todetermine the main variables of interest (CV, FL, XT, etc..)

FLOWSERVE VALBART – FLOW CONTROL DIVISION

Experimental testing

Experimental testing is fundamental,but has practical limitations:• Test rig capacity & availability• Investment in prototypes

Usually, few prototypes ofsmall size can be tested

What about theuntested valves?

FLOWSERVE VALBART – FLOW CONTROL DIVISION

Extending the experimental results

• Not all the dimensions/components scale accordingly withthe valve size

• Fluid dynamics do not simply scale with the size• Extrapolation of results from few concentrated points is not

recommendable

Simple scaling of theresults is not accurate

enough, especially whenconsidering more

“complicated” geometries.

Predicted Cv

Actual Cv

0

2000

4000

6000

8000

10000

12000

0 10 20 30 40 50 60 70 80 90 100

Cv

Travel %

FLOWSERVE VALBART – FLOW CONTROL DIVISION

Computational Fluid Dynamics

+ Pros• Virtual, no real physical

prototypes• Can simulate different conditions• Cost & time saving

- Cons• …Still need some money

investment• Need to be used with adequate

knowledge

FLOWSERVE VALBART – FLOW CONTROL DIVISION

CFD as reliable tool

3D Model• Must be accurate…• …but not over-detailed• Must take advantage of symmetries

Fluid domain

Full valve3D model

Save computational resources &time without sacrificing accuracy!

FLOWSERVE VALBART – FLOW CONTROL DIVISION

CFD as reliable tool

Mesh• Size• Quality of the mesh (Skewness, Smoothness, Aspect Ratio)• Sensitiveness of the solution to the mesh

FLOWSERVE VALBART – FLOW CONTROL DIVISION

CFD as reliable tool

Models & Boundary conditions• …must be set correctly (type, location, etc..)• Identify a proper set of boundary conditions to verify and fine-

tune the CFD methodology

Boundary setup

Experimental test conditions

FLOWSERVE VALBART – FLOW CONTROL DIVISION

CFD as reliable tool

Check solution quality• Convergence & Stability• Residuals• Monitor of other variables of interest

FLOWSERVE VALBART – FLOW CONTROL DIVISION

CFD fine-tuning & validation

CFD (first runs) vs. Experimental testing – Comparison• Notable differences observed:

1. CV values at full opening (for some trim solution)2. CV values at minimum openings

0

5000

10000

15000

20000

25000

0 10 20 30 40 50 60 70 80 90 100

Cv

Travel %

Test vs. CFD

Test

CFD

FLOWSERVE VALBART – FLOW CONTROL DIVISION

CFD fine-tuning & validation

1. CV values at valve full opening

Valve geometry changeswith the opening

Different dissipationeffect/mechanism may

become relevant atdifferent opening

FLOWSERVE VALBART – FLOW CONTROL DIVISION

CFD fine-tuning & validation

1. CV values at valve full opening• Introduction of pipe wall roughness into the simulation

0

5000

10000

15000

20000

25000

0 10 20 30 40 50 60 70 80 90 100

Cv

Travel %

Test vs. CFD

Test

CFD

0

5000

10000

15000

20000

25000

0 10 20 30 40 50 60 70 80 90 100

Cv

Travel %

Test vs. CFD

Test

CFD

Significant increase of the CFD accuracy at full open condition

FLOWSERVE VALBART – FLOW CONTROL DIVISION

CFD fine-tuning & validation

2. CV values at valve minimum openings

While at small openings,small differences in the ballrotation can cause relevant

differences in the areas

Precise positioning of theball during the test is

important at low openingsto get the correct CV

FLOWSERVE VALBART – FLOW CONTROL DIVISION

CFD fine-tuning & validation

2. CV values at valve minimum openings

Stroke%

CFDCv Test Cv Diff %

20% (*) 42.62 46.11 -7.57%30% (*) 121.52 128.20 -5.21%40% (*) 201.71 205.40 -1.80%50% (*) 291.99 305.45 -4.41%

Measures of tested valvebacklash where used to correct

the CFD model openingsthrough a simple model

Increased CV accuracy vs. test atlower openings, with minimal tono effect for mid-large openings

FLOWSERVE VALBART – FLOW CONTROL DIVISION

CFD fine-tuning & validation

CFD Fine-tuning observations:• Both the adjustments made to fine-tune the CFD methodology

proved to be:1. Consistent with the experimental data2. Beneficial for the accuracy of the results where they meant to

be so, irrelevant elsewhere

These adjustments have been applied to different valves andcompared to experimental data for final validation

FLOWSERVE VALBART – FLOW CONTROL DIVISION

CFD fine-tuning & validation

CFD vs. Experimental testing – Before and after

0

5000

10000

15000

20000

25000

0 10 20 30 40 50 60 70 80 90 100

Cv

Travel %

Test

CFD

0

5000

10000

15000

20000

25000

0 10 20 30 40 50 60 70 80 90 100

Cv

Travel %

Test

CFD

BEFORE AFTER

FLOWSERVE VALBART – FLOW CONTROL DIVISION

CFD fine-tuning & validation

CFD (fine-tuned) vs. Experimental testing – Comparison

0

500

1000

1500

2000

2500

3000

3500

0 10 20 30 40 50 60 70 80 90 100

Cv

Travel %

Trim N2-3 - Test vs. CFD

Test - N2-3

CFD - N2-3

0

1000

2000

3000

4000

5000

6000

7000

0 10 20 30 40 50 60 70 80 90 100

Cv

Travel %

Trim Z2-4 - Test vs. CFD

Test - Z2-4

CFD - Z2-4

0

2000

4000

6000

8000

10000

12000

0 10 20 30 40 50 60 70 80 90 100

Cv

Travel %

Trim STD - Test vs. CFD

Test

CFD

Increased CV predictionaccuracy vs. test datawith fine-tuned CFD

FLOWSERVE VALBART – FLOW CONTROL DIVISION

Extending the experimental results throughvalidated CFD methodology

• Should guaranteeadequate coverage ofthe product range

• Should considerspecific product design

Trimtype

Pressure class: ANSI 600

10” 12” 14” 16” 18” 20” 24” 28” 30” 32” 36”

STD ü ü ü

Z1-2 ü ü ü

Z2-4 ü ü ü ü ü

N1-3 ü ü ü

N2-3 ü ü ü

N2-4 ü ü ü

CFD campaign based on a matrix of simulations

FLOWSERVE VALBART – FLOW CONTROL DIVISION

Thank you!