oto00130

7/27/2019 oto00130

http://slidepdf.com/reader/full/oto00130 1/80

HSEHealth & Safety

Executive

Examination of the effect of relief deviceopening times on the

transient pressures developed withinliquid filled shells

Prepared by theUniversity of Sheffield

for the Health and Safety Executive

OFFSHORE TECHNOLOGY REPORT

2 0 0 0 / 1 3 0

7/27/2019 oto00130


Examination of the effect of relief deviceopening times on the

transient pressures developed withinliquid filled shells

B C R Ewan, D Nelsonand P Dawson

Department of Chemical and Process EngineeringUniversity of Sheffield

Mappin StreetSheffield

S1 3JD

HSE BOOKS

HSEHealth & Safety

Executive

7/27/2019 oto00130


© Crown copyright 2001Applications for reproduction should be made in writing to: Copyright Unit, Her Majesty’s Stationery Office,St Clements House, 2-16 Colegate, Norwich NR3 1BQ

First published 2001

ISBN 0 7176 1985 0

All rights reserved. No part of this publication may bereproduced, stored in a retrieval system, or transmittedin any form or by any means (electronic, mechanical,photocopying, recording or otherwise) without the priorwritten permission of the copyright owner.

This report is made available by the Health and SafetyExecutive as part of a series of reports of work which hasbeen supported by funds provided by the Executive.Neither the Executive, nor the contractors concernedassume any liability for the reports nor do theynecessarily reflect the views or policy of the Executive.

7/27/2019 oto00130


SUMMARY

A recent Joint Industry Project under the management of the Institute of Petroleum has beenconcerned with the failure scenario in which a shell and tube heat exchanger with a low pressure rated shell and with a high pressure tube-side suffers tube failure.

For water filled shells, the ensuing scenario is one in which a gas bubble rapidly growsaround the burst site generating a rising pressure wave within the shell. This pressure wave iseventually relieved when the protective hardware operates leaving a residual lower pressuretail as the water is driven out of the shell.Tube failure for a shell protected by a relief device therefore gives rise to a transient pressure pulse whose characteristics are influenced by the failure site geometry and location and thelocations and dimensions of the relief points.The study revealed the importance of relief device opening times on the peak pressures whichcould result and the present work being reported has been concerned with providinginformation on such opening times for a range of devices including bursting discs and relief valves.

The experimental work has been performed at University of Sheffield and the data from thishas been subjected to a hydraulic modelling analysis by PSI Ltd providing a good overview tothe sequence of events associated with wave propagation and relief operation.Opening can be defined in a number of ways ranging from the instant at which pressure begins to be relieved to the time at which the relief device is fully open. Ultimately, devicesmust be chosen which meet the relief requirement on a timescale which is compatible with theupstream system design and its failure characteristics. This represents an issue of sizing aswell as relief design choice.The devices studied have not been chosen to meet any particular relief objective, and it became clear that the bursting discs were oversized for the relief duty whilst the relief valveswere undersized. However, the data acquired enabled the times to full opening to be measuredwith a high degree of confidence.

The burst discs ruptured in 1.9-10 msec and this is in line with (and therefore supportive of)the values used in the IP study, which formed the basis of the IP Guidelines.The study also shows that the pop-action of the relief valves (RVs) studied occurred in 2.5-4msec but this finding, in particular, should be viewed with extreme caution and should not betaken out of context. We believe it would be premature if these values were taken as typicaland applied across the industry in general, for all sizes of device and all operating conditions.Overall, our reservations are as follows:Firstly, the study implies that RVs are faster than a metal burst disc but this, potentiallymisleading finding, has arisen by comparing a vastly oversized (8") disc against a severelyundersized, (2in) RV.Secondly, the study has given the unexpected finding of very fast opening times for the RVs,

(4msec) almost 100 times faster than some of the values quoted by others. Although a quick response can be expected from the test case (the valve is very small and it is the pop-actiontype) this may not be typical in the field

ii

7/27/2019 oto00130


iii

CONTENTS

page

SUMMARY ii

1. BACKGROUND AND OBJECTIVES 1

2. PLAN OF WORK 2

3. METHODOLOGY 2

3.1 Test Details 2

3.2 Test Characteristics 4

3.3 Experimental facilities 5

3.4 Experimental Method 7

4. RESULTS 8

4.1 Pressure Traces 8

4.2 Video Tape 8

5. ANALYSIS 9

5.1 Summary of modelling analysis 9

5.2 Preliminary Appraisal 145.3 Open Tube Tests 175.4 Graphite Burst Disc Tests 20

5.5 Stainless Steel Burst Disc Tests 265.6 Spring Loaded RV, High Pressure Test 315.7 Spring Loaded RV, Low Pressure Test 35

5.8 Pilot Operated relief Valve Test 39

6. CONCLUSIONS AND RECOMMENDATIONS 43

7. APPENDIX 45

7.1 Plates 45

7.2 Data Charts 557.3 Video tape record 73

7/27/2019 oto00130


1. BACKGROUND AND OBJECTIVES

A recent Joint Industry Project under the management of the Institute of Petroleum has been

concerned with the failure scenario in which a shell and tube heat exchanger with a low pressure

rated shell and with a high pressure tube-side suffers tube failure.For water filled shells, the ensuing scenario is one in which a gas bubble rapidly grows around

the burst site generating a rising pressure wave within the shell. This pressure wave is eventually

relieved when the protective hardware operates leaving a residual lower pressure tail as the

water is driven out of the shell.

Tube failure for a shell protected by a relief device therefore gives rise to a transient pressure

pulse whose characteristics are influenced by the failure site geometry and location and the

locations and dimensions of the relief points.

The experimental study indicated that pulse widths could be in the range 1 - 10 msec for the

graphite bursting discs used, and this was comparable to results from recent numerical work by

Cassata et al1

for a longer shell, where pulse widths in the range

4 - 20 msec could be obtained with amplitudes of up to 25% of the tube-side pressure.

A recent parallel study2

was concerned with the response of the shell structure within the time

duration of the pulse loading, and concluded that engineering benefits could be secured by

ensuring the pulse widths were kept as short as possible compared with the fundamental period

of the shell structure. In this case, peak pressure amplitudes several times the yield pressure of

the shell could be tolerated. Since shells will typically have fundamental periods in the range 5 -

20 msec, design efforts should be directed at limiting pulse widths to 1/2 - 1/3 of this range. The

choice of relief device is important in this context as well as the number of such devices and

their distribution on the shell.

The objective of the work now being reported has been to measure a number of relief deviceopening times and to consider their effect on the amplitude and duration of transient vessel

pressures arising from high pressure tube failures. This has been carried out through a series of

controlled experiments to monitor transient pressures in a water filled column combined with a

hydraulic analysis of the dynamic event.

1

7/27/2019 oto00130


2

2. PLAN OF WORK

The programme of work undertaken had the following specific objectives :

1. To measure the opening times of a range of pressure relief devices when subjected to rapidly

rising pressure within a water medium,

2. To monitor the rates of pressure rise and peak pressures in the water column, which result

from chosen combinations of gas discharge pressure and orifice size,

3. To provide an interpretation of the key features of the pressure behaviour using current

understanding of transient hydraulics.

3. METHODOLOGY

The shock tube schematic shown in Figure 3.1 and described in greater detail below, was usedfor the experiments.

bursting diaphragm

and orifice location

high pressure

driver section

relief

device

9 m3 m

water filled

driven section

+ + + + ++

+ = pressure transducer

21

Figure 3.1 Outline of shock tube used for experiments

This is instrumented with a number of Kistler pressure transducers along its length.

The shock tube has a 100mm internal diameter and transition sections were available to extend

this to 200mm. It was intended that relief devices be examined whose diameters are

representative of those used offshore and that these should cover a range of types from the

fastest to the slowest operating and include bursting discs, spring loaded and pilot operated relief

valves.

3.1 TEST DETAILS

Since the key features of the effect of relief device opening time could be demonstrated over a

wide range of pressures, it was proposed that tests be carried out with relief devices which

operate at a nominal 15 bar gauge pressure. The shock tube gas driver pressure was set at 100

7/27/2019 oto00130


3

bar and pressurisation of the water column locating the relief device was by means of rapid

discharge through suitably sized orifices.

7/27/2019 oto00130


4

The key issues of interest are (i) the effect of opening times which cover the typical range for

devices in use, (ii) the relationship between the imposed rate of pressure rise from the source

and the resulting peak pressure when combined with these different opening times and (iii) the

demonstration of pressure interpretation and predictive capability through the use of hydraulic

modelling.

It was therefore proposed that the following relief devices be tested within the above water filled

shock tube :

Bursting discs :

Diameters 4 in, 6 in and 8 in

Materials Graphite and stainless steel

Relief valves : 2” Spring loaded (pop-action), 2” Pilot operated, 2” bellows.

Note that relief device dimensions are given in inches to be consistent with manufacturers

specifications for the devices.

The range of discharge orifices to be used to achieve different rates of pressure rise in the water

column was 4mm, 8mm and 15mm in diameter.

The combinations available provided 24 sets of results, including those with an open ended tube

(thin plastic film). These are summarised in the test matrix shown in Table 1, where test

numbers refer to specific data sets generated.

Table 3.1.1

Summary of test conditions and test numbers

Relief device↓

Relief diameter

(in)

4mm orifice 8mm orifice 15mm orifice Relief pressure(bar)

Open tube 39 38 37 0

Graphite disc 4 51 50 49 10

6 55 54 53 10

Stainless steel disc 4 41 42 40 15

(reversed dome) 8 48 47 46 15

2” Spring loaded RV - 59 58 57 15

2” Bellows RV - 62 61 60 15

2” Pilot operated RV - 66 65 64 15

= tests for detailed analysis

3.2 TEST CHARACTERISTICS

In simple terms, the tests are completely different and, as discussed later, these differences have

a significant impact on the action of the relief devices. Thus, the two tests that form the basis of

the mathematical modelling have the following characteristics:

The 15mm orifice gives a high pressure test which subjects the devices to a significant, rapidly

applied pressure which is far higher than the rupture pressure

The 4mm orifice givers a relatively low pressure test.

7/27/2019 oto00130


5

3.3 EXPERIMENTAL FACILITIES

The main facility used to generate a rapidly rising pressure transient was a shock tube of 100mm

internal diameter. This consisted of a driver section which could hold high pressure gas and

which was separated from a short buffer section containing atmospheric air by an aluminium

bursting disc. Discs could be designed to fail at any pressure and all tests were conducted at disc

failure pressures of 100 bar ± 10%. The buffer section was connected to the water filled driven

section of the tube via a plate, in the centre of which was the discharge orifice. To retain the

water during filling, the downstream side of the orifice plate was sealed with thin aluminium foil

or plastic film.

Four Kistler pressure transducers (K1 - K4) were located at positions along the driven section to

record the transient pressure profile. K1 being closest to the orifice was also used to trigger the

data acquisition. K4 was chosen to be as close as possible to the relief device at the opposite end

of the tube to minimise the transmission delay during device opening. A separate pressure

transducer, D1 was used to monitor the driver pressure and provided the starting pressure at the

point of rupture.

Slight variations in geometry were used for different groups of tests and these are represented inFigures 3.3.1 - 3.3.4.

Air was supplied to the driver section from a compressor via an air reservoir and a number of

electrically operated isolation valves.

Raw voltage signals were acquired to a computer acquisition card at a rate of 20kHz on each

channel.

Figure 3.3.1 Shock tube geometry and dimensions used in tests 37 - 41, 49 -

51.

K1 - K4 represent positions of Kistler pressure transducers.

DIMENSIONS IN mm

ORIFICE

PRESSURISING

WATER COLUMN

DRIVER

SECTION

RELIEF

DEVICE

WATER FILLED TUBE

K4K2 K3K1

150 2440 2140 1980 90

7/27/2019 oto00130


6

Figure 3.3.2 Shock tube geometry and dimensions used in tests 46 - 48.

Figure 3.3.3 Shock tube geometry and dimensions used in tests 53 - 55.

ORIFICE

PRESSURISING

WATER COLUMN

DRIVER

SECTION

RELIEF

DEVICE

WATER FILLED TUBE

K4K2 K3K1

150 2440 2140 1205 65

100

150870

200

ORIFICE

PRESSURISING

WATER COLUMN

DRIVER

SECTION

RELIEF

DEVICE

WATER FILLED TUBE

K4K2 K3K1

150 2440 2140 935 65

150

150600

100

7/27/2019 oto00130


7

Figure 3.3.4 Shock tube geometry and dimensions used in tests 57 - 62, 64 -

66.

3.4 EXPERIMENTAL METHOD

Aluminium bursting diaphragms could be produced in house and designed to rupture at any

prescribed pressure. These were designed for 100 bar operation and located so as to seal the

driver section of the shock tube. A 60mm thick spacer was located between this diaphragm and

a discharge orifice which fed into the water filled downstream section of the tube. The orifice

plate also held in place a plastic film to isolate the water column.

The removal of air bubbles from the water filled tube and connections before firing was an

important requirement to avoid contamination of the pressure traces by spurious reflections from

gas interfaces. Water inlets and outlets and the fabrication detail around relief devices were all

optimised such that no air pockets could remain during filling.

It was found initially that small air bubbles would remain immobile along the surface of the tube

during filling. The flow velocity of water during tube filling was important for their removal and

therefore water supply and outlet diameters were maximised whilst the shock tube was also

inclined at a gradient of 1:75. When the tube was full of water, surfactant was mixed with the

feed water to aid bubble flow and this feed was maintained until no further bubbles were

obtained in the outflow. Fresh water was finally flushed through the tube.

This procedure was ultimately judged to be satisfactory although the removal of the smallest

bubbles remained a central part of the experimental procedure before each test.

Each test then consisted of the slow pressurisation of the driver section until the aluminium

diaphragm ruptured. Data acquisition on all Kistler channels and the driver stagnation pressure

was then triggered by the voltage rise on K1. A small proportion of pre-triggering time was also

collected and allowed an exact determination of the pressure at rupture.

The pressure wave then took around 5.5 msec to arrive at K4 and it was therefore considered

sufficient to collect 250 msec of data. Only the early part of this is relevant to the relief device

opening and is represented on the enclosed charts.

Burst disc holders for both types used were fabricated according to manufacturer's specification.

ORIFICE

PRESSURISING

WATER COLUMN

DRIVER

SECTION

WATER FILLED

TUBE

110

4"

K3K1

150 2440 2140

1100

K4920

120

K2 RELIEF

POINT

DIMENSIONS IN mm

7/27/2019 oto00130


8

4. RESULTS

4.1 PRESSURE TRACES

The combined pressure - time traces for K1 - K4 for each test are shown in Charts A.1 - 33

within the Appendix section. In all cases K1 is the first to rise and K4, being furthest from thesource, is the last. Identification of the transducers can therefore be made on this basis. D1 is

not shown but instead the driver pressure at the moment of rupture is given on each trace. The

relief device set pressure is also indicated on each trace in the region of K4. Although pressure

data was collected over a 250 msec period, only the sections of this relevant to the discussion on

relief device opening time is included. For the bursting discs this is up to 15 msec after

diaphragm rupture. For the relief valves there is periodic behaviour and therefore for these, time

resolved traces up to 15 msec are presented initially and are followed by a longer duration trace

up to 50 msec.

4.2 VIDEO TAPE

A video tape record of the shock tube geometry and a number of components used has been

produced in conjunction with this work and the details of this are included in the Appendix. The

video record also includes a high speed sequence of the stainless steel and graphite bursting

discs opening.

7/27/2019 oto00130


9

5. ANALYSIS

5.1 SUMMARY OF MODELLING ANALYSIS

Introduction

The results which have been produced provide an overview of behaviour with respect to

operating conditions and device types. To provide a greater insight into the relief device

operation, a detailed analysis has been undertaken for 10 of these tests. The following sections

therefore summarise the findings of this detailed analysis, with the particular objectives of:

• Determining the opening time of the relief devices

• Providing an interpretation of the key features of the pressure behaviour

Analysis methods

Fundamentally, the detailed analysis comprised a dynamic simulation study using a mathematicalmodel. A hydraulic model of the shock tube was configured with all the project data (i.e. tube

length, diameter, driver pressure, orifice diameter etc.) to produce an accurate mathematical

representation of the system. The model was then calibrated to provide good correlation

between the measured and predicted results.1

Finally, existing models were incorporated of the

burst disc and relief valves to reproduce the tests, as applicable, primarily using trend studies to

determine the opening time of the devices.

In addition to the dynamic simulation, wide ranging appraisals were undertaken. PSI have

extensive experience in the effects of surge pressure changes in pipes and piping systems and

this understanding was applied to the evaluation of the physical test results themselves, as well

as the findings from the dynamic simulation.2

For example, one of the benefits of dynamic simulation is the ability to provide additionalinformation on the behaviour of a piping system, to supplement the data from SCADA systems,

transducers and pressure gauges.3

This information, equivalent to the output from virtual

instruments, provides further diagnostic information and this, together with PSI's experience,

means that evaluation of the test results was particularly useful.

The benefits of this approach were even evident during the initial testing phase. PSI were able to

interpret the test results and suggest that the presence air in the shock tube was generating 'non-

standard' behaviour. Sheffield University then identified the source of the problem and eliminated

it with revised test procedures. Subsequently, very good correlation was obtained between the

measured and simulated results for the Open Tube tests giving confidence that the dynamic

simulation phase with the relief devices would provide a meaningful outcome.

High and low pressure tests

1

Some parameters, such as the internal hydraulic roughness of the tube and the amount of free air in the

test-water are not unique data items and can vary between systems. These are therefore adjusted in the

calibration

2Additional information is provided in Appendix A and B

3This is widely used in industrial applications and is extremely helpful in troubleshooting; it is a non-

invasive and low-cost way of investigating operating problems

7/27/2019 oto00130


10

In general two groups of tests results were looked at, differing because of the size of the orifice

on the driver end of the shock tube. The larger orifice (15mm) gave a high pressure test with

peak pressures of over 60 barg at the test-end of the tube in the Open Tube tests; this is over 4

times the nominal set pressure of the relief devices. In contrast the small orifice (4mm) gave a

low pressure test with pressures which were only a few bar above the set pressures.

These test-groups were therefore completely different and the study shows that these

differences had a significant impact on the action of the relief devices. In simple terms the relief

devices responded fully in the high pressure tests with the burst discs rupturing completely and

the relief valves lifting fully. In contrast, the low pressure tests gave 'marginal' conditions,4

tending to give slower (or incomplete) operation of the devices. However, this project arose from

the investigation into tube rupture within an industrial heat exchanger and, in that context, the

rapidly applied high pressure tests are far more relevant than the marginal, low pressure ones.

Behaviour of Relief Devices

The study examined two groups of devices (i.e. burst discs and relief valves) categorised by the

fact that the burst discs are 'non-closing' devices whilst relief valves (RVs) are designed to

automatically re-close and prevent the further flow of fluid. But, in practice, the findings

confirmed the need to further sub-divide these categories and hence the four different simulation

models that are available. Although there are many similarities between the graphite and metal

burst discs, the graphite discs continue to shatter even if the initial pressure surge decays.5

In

contrast, a metal disc only opens when there is a positive driving force; as seen by the

researchers, the metal discs were only partially opened in some of the tests. In turn, this tended

to impose an effective back pressure on the relief system.

Similarly, the study confirms that the two types of relief valve necessitate PSI's different

simulation models. Both of the valves are characterised by 'pop' action (whereby the design of

the internal forces within the valve means that they open rapidly when the inlet pressure is

slightly above the set pressure) but the pilot valve can have a modulated closure characteristic

while the spring loaded valve also re-seats rapidly.6 As discussed below, these closure

characteristics are particularly significant in 'marginal' pressure conditions.

Accuracy of Modelling

The overall accuracy of the modelling and the reason for the high level of confidence in the

study as a whole is best shown by example. Figure 5.1.1 demonstrates the excellent level of

correlation that was obtained, in this case with the high pressure test on the 4" graphite disc. The

simulated result is the dark line, superimposed over the measured result.

4

The term 'marginal' is used in this report to indicate cases where the surge pressure in the shock tube was

only slightly higher than the set pressure of the relief device

5Under marginal conditions, discs did not always rupture fully, leaving a narrow annulus at the edge. But

this had little effect on the disc capacity and can be classified as materially complete rupture

6The full RV operating cycle is briefly outlined in Section 5.6. More comprehensive information is given in

manufacturers' catalogues.

7/27/2019 oto00130


11

0

5

10

15

20

25

3035

40

45

0 0.002 0.004 0.006 0.008 0.01 0.012

Test 49p 1.9ms Rupture

K4 Pressure Time (seconds)

Pressure (bar)

1. 10

Sheffield Model - Re-Calibrated after Test 39p

Max = 40.87

Min = 0

Figure 5.1.1 Correlation for Graphite Disc, High Pressure Test

Overall, the study therefore confirmed the suitability of the simulation models of the different

relief devices and so the trend studies were used to investigate their response time. In this, all the

test parameters remained the same with the exception of one (i.e. the overall rupture time of the

burst disc), which was changed systematically.

Graphite and Metal Burst Disc

The trend studies showed that the burst time for the 4" graphite disc was 1.9 msec compared

with 10 msec for the 8" stainless steel disc. These compare favourably with the values of 0.1-10msec given in the recent IP Guidelines for the Design of Heat Exchangers.

7However, the

analysis identified other factors.

In the context of the 4" shock tube, the 4" burst disc was oversized, presenting a massive amount

of relief and this was even more marked with the 8" disc. The effect of this was that, even in the

high pressure tests, the pressure wave was relieved within about 10% of the opening time and

this means, subsequently, the test became insensitive to the increasing disc capacity. In practice

therefore the overall burst time of the metal disc represents an effective rupture period of about

1 msec, extrapolated to give an overall best estimate of 10 msec.

Spring Loaded and Pilot Relief Valves (RVs)

The correlation for the spring loaded relief valves was also excellent (e.g. Figure 5.1.2, where

again the simulated result is the dark line, superimposed over the measured result). It is also

interesting to note that, in the high pressure tests, the 2in RVs do not exhibit any of the over-

sizing problems seen with the burst discs. In the test below, the RV remains open after the initial

pressure wave is relieved; the system pressure remains above the valve's set pressure of 15

barg.

7

Guidelines for the Design and Safe Operation of Shell and Tube Heat Exchangers to Withstand the

Impact of Tube failure, The Institute of Petroleum, London, 2000

7/27/2019 oto00130


12

-20

0

20

40

60

80

100

120

0 0.002 0.004 0.006 0.008 0.01 0.012

Test 57p 10% Capacity in 4ms


Pressure (bar)

5. 8

Sheffield Model - Taken from Test 51p

Max = 100.75

Min = 0

Figure 5.1.2 Correlation for Spring RV, High Pressure Test

Strictly, the tested RV is defined by API RP 520 as a 'safety' valve which means that it opens

by pop-action. In this way it differs from valves defined by API RP 520 as 'relief valves' in

which the lift is proportional to the inlet pressure. Similarly, the pilot configuration offers the

same options of either pop-action or modulating-action; there is also a combined option of pop-

open and modulate-closed which was the configuration studied in these tests. And overall these

different options are significant because they mean that the test valves were the fastest

available.

The 'rated' capacity8

of a pop-action valve is reached when it has popped open (although a slight

increase in capacity can subsequently occur on a further increase in pressure). We havetherefore defined the opening time for the RV as the time taken for it to reach the rated capacity

i.e. the capacity for 110% of the set pressure, or 10% over-pressure (10% OP). And on this

basis the spring loaded valve opened in 4 msec and the pilot valve in 2.5 msec.

These values are supported by the high level of correlation shown above and, by inspection,

particularly of the low pressure tests discussed later. However, these results were completely

unexpected; they do not agree with the data given in the IP Guidelines of 80-350 msec and are

faster even than the tested value of 25 msec by Kruisbrink, 1990.9

The finding must therefore be

treated with extreme caution until further work has been undertaken to investigate this further

and to establish points of comparison. In the interim however, we do make the following

observations:

The RV is small (2H3), it is subjected to a significant over-pressure and it is the pop-action typeand, subjectively, we would therefore expect the opening times to be at the fast end of any

performance range. Further testing (or test analysis) is needed to establish whether the findings

are size or pressure related i.e. whether these findings are representative of larger valves at

different over-pressure conditions.

We were also surprised by the finding that the pilot valve opened more quickly than the spring

loaded valve; again, subjectively, a slower response was expected. However, we noted that the

8

The 'rated' capacity is in accordance with international codes

9Modelling of Safety and Relief Valves in Waterhammer Computer Codes, Kruisbrink, A.C.H., Proc. 3rd

Int. Conf. Valves and Actuators, BHR Group, STI 1990

7/27/2019 oto00130


13

pilot operated RV and the spring loaded RV are not directly similar valves. The physical mass of

the moving parts is smaller for the pilot operated valve and so the inertia effects would be lower.

Additionally, the study suggests that the pilot is the type that is characterised by pop action,

followed by a modulating action. This means that the opening time would not be adversely

slowed by the pilot

Behaviour of Devices at Low Pressure

The over-sizing issue mentioned earlier was even more evident in the low pressure tests because

these only created 'marginal' conditions for all of the devices. The most significant case arose

with the spring loaded RV, generating the classic case of valve chatter (Figure 5.1.3).

2in Spring Relief - 4mm - Smoothed

0

10

20

30

40

50

0 0.05 0.1 0.15

Time (s)

Pressure

(barg)

K4

Figure 5.1.3 Measured Test Showing RV Chatter

API RP 520 notes that RVs "operating at low pressures tend to chatter; therefore overpressures

of less than 10% should be avoided". And these are the very conditions that exist in this low

pressure test. It is a serious problem in the field; as shown above, it can lead to pressure

oscillations (with the attendant problems of vibration, pipe movement and damage) as well as

damage to the valve itself and galling of the guiding surfaces. And, as a result valve chatter has

been the subject of other industrially oriented research (e.g. Kruisbrink, 1990 and Auble, 198310

).

The detailed study of this phenomenon was therefore well outside the scope and the aims of this

study but we still obtained good correlation with the opening phase. Moreover, the results support

our estimation of the opening time for the RV of 4 msec. Figure 5.1.3 shows an oscillatingfrequency of about 80 cycles per sec, giving an opening/closing sequence within 12.5 msec.

The detailed analysis of the performance of the pilot under the 'marginal' conditions was also

beyond the study scope although again we showed good correlation with the opening phase. The

effects of over-sizing were still a potential problem although, in this case, the modulating action

of the pilot totally eliminated the chatter seen previously with the spring loaded RV. However,

this is not the only way of eliminating chatter; for example, it can be avoided by using dynamic

simulation methods to ensure that the spring loaded RV is correctly sized in the design stage.

10

Full Scale Pressurised Water Reactor Safety Valve Test Results, Auble T.E., Testing and Analysis of

Safety/Relief Valve Performance, 4th National Congress on Pressure Vessel and Piping Technology,

Portland Oregon, 1983

7/27/2019 oto00130


14

Conclusions

The overall aim of the study was to determine the opening time of relief devices that may be

used on industrial heat exchangers to provide protection in the event of a tube rupture. And this

has been achieved with the study showing that fast-acting protection devices are available. The

burst discs ruptured in 1.9-10 msec and this is in line with (and therefore supportive of) the

values used in the IP study, which formed the basis of the IP Guidelines.

The study also shows that the pop-action of the RVs occurred in 2.5-4 msec but this finding, in

particular, should be viewed with extreme caution and should not be taken out of context. We

believe it would be premature if these values were taken as typical and applied across the

industry in general, for all sizes of device and all operating conditions.

Overall, our reservations are as follows:

Firstly, the study implies that RVs are faster than a metal burst disc but this, potentially

misleading finding, has arisen by comparing a vastly oversized (8") disc against a severely

undersized, (2in) RV.

Secondly, the study has given the unexpected finding of very fast opening times for the RVs,

(4msec) almost 100 times faster than some of the values quoted by others. Although a quick

response can be expected from the test case (the valve is very small and it is the pop-action

type) this may not be typical in the field

Further work is therefore essential to determine whether the action of the RV is affected by

size, pressure, onsite variables (such as the settings of blowdown rings) etc.

5.2 PRELIMINARY APPRAISAL

Introduction

Our preliminary appraisal of the test results raised questions that might affect the mathematical

modelling, the test program and/or the test procedures that were being used. The first section of the report therefore reproduces a document which was issued to Sheffield University with the

aim of raise these questions and thereby increasing the likelihood that, between us, we could

identify and eliminate (or control) the phenomenon that was initially presenting as 'rogue' (i.e.

non-standard) behaviour.

Test Data

This review uses the basic numbering system from the University tests.

Table 5.2.1

Summary of Test Numbers and Conditions

Type 4mm

Orifice

8mm Orifice 15mm Orifice

Test No. 1 Test No. 2 Test No. 3Open

Tube Test No. 1b Test No. 3b

Sample Results

From our basic experience and from our involvement in a similar study in the past, we would

expect the results to show a basic surge pattern,11

with test-specific effects superimposed. But

11

For reference, the basic wave and the variations are outline in Appendix A

7/27/2019 oto00130


15

comparison between the predicted and measured results for K2 during Test 3b (Figure 5.2.1)

shows several departures. Most notably:

the period of the measured wave is longer than the predicted one

the measured wave shows an inflection - an early decrease and subsequent recovery12

0

10

20

30

40

50

60

70

0 0.005 0.01 0.015 0.02 0.025 0.03

Basic Test

Sheffield Test 3b - 15mm orifice and film


Pressure (bar)

2. 6

Standard Model

Max = 59.87

Min =-0.03

Figure 5.2.1 Idealised and Measured Pressures, High Pressure Test

Other features we noted are:

the wavespeed is not apparently constant down the tube - the transmission time between K1/K2

and K2/K3 gives a wavespeed of between 1500-1200m/s but there is a step change between

K3/K4 down to a minimum of 350m/s the apparent friction loss in the tube (indicated by the pressure offset between K1 and K4 as the pressure wave decays) is significantly higher than

expected

Wavespeed Changes

To investigate this further, the mathematical model was arbitrarily calibrated to match the

measured wavespeeds and the results immediately mimicked the M-shape wave.

However, we had no reasonable hypothesis to support this calibration:

in the absence of any air bubbles in the water, the wavespeed in this test rig would be constant

because it is a solid steel tube

the wavespeed would vary if the water contains a typical distribution of free air (as occurs whentaking water from the mains, for example). But, under these circumstances, the wavespeed

would vary with pressure (as the bubbles compress) rather than distance and so this would not

give the sudden step change between K3/K3

Localised Phenomena

In view of the fact that the difference appeared to lie between K3 and K4 we investigated the

potential impact of two possible options, namely a discrete air bubble or pipe distension. These

12 This can be visualised as an M-shape

7/27/2019 oto00130


16

were localised at the site of the orifice/baffle13

because we were unaware of any other suitable

location on the tube.

The results show that, as with the calibrated wavespeed, the approach of using a short section of

distensible pipe gives the right trends at K2 but K3/K4 are poor.

In contrast, the inclusion of a small air pocket gives fair correlation at all of the transducers,

replicating both the initial M-shape and also the amplitude and phasing as the wave decays

(Figure 5.2.2).

-10

0

10

20

30

40

50

60

70

80

90

0 0.01 0.02 0.03 0.04 0.05

Air Bubble Test

Sheffield ST3 with Air at Baffle Site (0.2 litre)


Pressure (bar)

8. 6

Standard Model

Max = 76.35

Min =-.91

Figure 5.2.2 Effect of Air Pocket on High Pressure Test

We did not attempt to optimise the correlation and so, as shown on Figure 5, the prediction over-

estimates the second part of the M-wave. But the overall trend is sufficiently close that we feelit clearly illustrates the phenomenon.

Additional Data

We also assessed the additional data (Test 2 and the re-tested results, Test 1b and 3b) and

although these show some differences, the overall trends are materially unchanged; most

significantly, both the M-shape wave and the wavespeed variations still remain.

For completeness, the additional data shows:

more conformity between tests 1b and 3b, differing as expected only by test-specific factors. In

contrast, the pressures originally recorded at K3 and K4 during Test 1 had been very different

from the pressures at the same sites in Tests 2 and 3. But this disparity is eliminated by the re-testing (i.e. Test 1b)

some difference between the pressures at K1 in both Test 1b and 3b compared against the

original testing. But this is to be expected - this site is the one that is most likely to be affected by

blowby from the driver section, one of the reasons for re-testing

Summary

In summary therefore, the results suggest that the new testing procedure has eliminated one

source of unpredictable variation but a further source still remains.

13 We understood that there had been a baffle in the tube for the initial JIP testing

7/27/2019 oto00130


17

We felt that the presence of a small air pocket was the most likely cause of the 'non-standard'

behaviour, firstly because of the correlation shown above and secondly because the volume of

air that gives good results differs between the tests. Sheffield University therefore expended a

great deal of time and effort on identifying the 'rogue' phenomenon and eliminating it with revised

test procedures.

5.3 OPEN TUBE TESTS

Introduction

Our preliminary appraisal of the initial test results suggested that small air pocket(s) were the

most likely cause of the 'non-standard' pressure and flow transients. And, as a result, a

significant amount of work was undertaken by Sheffield University to isolate and then eliminate

the air.

This section therefore compares the final set of measured results for the Open Tube tests with

the mathematical modelling results.

Test Data

Table 5.3.1

Test Summary

Type 4mm Orifice 8mm Orifice 15mm Orifice

Open Tube Test No. 39p - Test No. 37p

Calibration of the Hydraulic Model

The hydraulic model of the shock tube was firstly configured with all the project data (i.e. tube

length, diameter, driver pressure, orifice diameter etc.) to produce an accurate mathematical

representation of the system. However, some parameters, such as the internal hydraulic

roughness of the tube and the amount of free air in the test-water are not unique data items and

can vary between systems. The model was therefore calibrated to provide good correlation

between the measured and predicted results.

Using the methodology outlined in the Preliminary Appraisal, the measured data was reviewed

for consistency, smoothed14

and then the wavespeed and friction losses were calibrated.

The low pressure test (39p) was then simulated and, using transducer K2 as an example, Figure

5.3.1 shows that a reasonable correlation was obtained:

The pressure rises at the same timeThe phasing (periodicity) of the waves is the same

The peak pressure is similar

14

A limited amount of data smoothing (5 point average) was used to eliminate the most severe of the

measured oscillations

7/27/2019 oto00130


18

0

5

10

15

20

0 0.005 0.01 0.015 0.02

Test 39p - Calibrated


Pressure (bar)

1. 6

Sheffield Model - Calibrated from Test 39p

Max = 18.14

Min = 0

Figure 5.3.1 Preliminary Correlation for Open Tube, Low Pressure Test

Air in the Tube

Although Sheffield University had made significant changes to their test procedures and thereby

reduced the amount of air that was trapped within the tube, they noted that some could remain

on the downstream side of the filling outlet, on the tube soffit. A small volume was therefore

included in the hydraulic model and this immediately introduced the type of pressure inflections

seen on the measured traces.

The volume was therefore calibrated at about 0.025 litres (under initial, atmospheric conditions)and the resulting correlation is given on Figure 5.3.2. For comparison, the high pressure test

(37p) was also run on the same model (Figure 5.3.3).

-5

0

5

10

15

20

0 0.005 0.01 0.015 0.02

Test 39p - Calibrated with Air Pocket and Free Air


Pressure (bar)

6. 8


Max = 16.90

Min = -0.7

Figure 5.3.2 Correlation for Open Tube, Low Pressure Test

7/27/2019 oto00130


19

0

10

20

30

40

50

60

70

80

0 0.01 0.02 0.03 0.04 0.05

Test 37p - Calibrated from 39p


Pressure (bar)

7. 8


Max = 62.47

Min = 0

Figure 5.3.3 Correlation for Open Tube, High Pressure Test

Criteria for Acceptance

Our criteria for acceptance of the accuracy of the correlation are based only on the first wave.

This generates the pressure conditions that opens the relief device and so the subsequent

pressure oscillations are not relevant.

Additionally, we have taken position K3 to be the most important for this particular correlation.

Although K4 is closer to the end of the shock tube (and therefore closest to the relief device) it

provides very little information in the Open Tube tests because the duration of the pressure wave

at this point is too short. K3 is therefore used this correlation.

Review of Results

The hydraulic model used to generate Figure 5.3.2 and Figure 5.3.3 was therefore accepted for

the second phase of the study (i.e. the analysis of the relief device tests) on the basis that:

The initial pressure rises are coincident with the measured tests

The initial rates of pressure rise are the same as the tests

The duration of the pressure waves is the same as the tests

The magnitude of the pressure waves is similar

Summary

Our analysis of the hydraulic modelling and measured results for the Open Tube tests shows that

a high level of correlation can be obtained for both of the available tests. Some air remained in

the shock tube under test conditions but the volume was now very small and could be adequately

accommodated by calibration.

We were therefore confident that the calibrated model could be used for the analysis of the test

results obtained from the relief devices.

7/27/2019 oto00130


20

5.4 GRAPHITE BURST DISC TESTS

Introduction

This section of the report describes the next phase of the modelling study, investigating the

response times of relief devices; in this case two sets of results from the 4in graphite burst disc

tests are studied.

Test Data

Table 5.4.1

Test Summary


4in Graphite

Burst Disc

Test No. 51p - Test No. 49p

The nominal burst pressure for the discs is 15.4 barg.

Test Characteristics

As noted earlier, the tests are completely different in simple terms and this study shows that

these differences have a significant impact on the action of the relief devices:

The 4mm orifice givers a relatively low pressure test. In the Open Tube tests, the peak

pressures were only a few bar above the nominal burst pressure of the discs (15.4 barg)

The 15mm orifice gives a high pressure test with the disc subjected to a significant, rapidly

applied pressure which is far higher than the rupture pressure

It is also interesting to note that the university researcher observed a physical manifestation of

these differences; the graphite discs did not always shatter completely with low pressure tests,

occasionally leaving an graphite annulus at the edge of the disc holder.

Mathematical Model of the Burst Discs

Under steady state conditions, the mathematical model used by PSI validates against API RP

520 and manufacturers' catalogue data. The aim here was therefore to calibrate the opening

time and thereby confirm the model under dynamic conditions.

Trend Study for High Pressure Test (49p)

The 15mm orifice test was investigated first because this is the 'high' pressure test which

subjected the disc to a significant pressure wave. This means that the disc ruptured completely

with no likelihood of a residual annulus.

Firstly, the hydraulic model of the shock tube was re-configured from the Open Tube tests to

include a burst disc. Then, a trend study was undertaken to investigate the burst time of the test

disc. In this, all the test parameters remained the same with the exception of one (i.e. the overall

burst time of the disc) which was changed systematically.

The burst time was taken as the time for the disc to shatter completely. And the results on

Figure 5.4.1 and Table 5.4.1 show the effect of parameter on the pressure at K4. The dark line

7/27/2019 oto00130


21

is the measured result and the 3 lighter ones are the simulated result with a disc burst time of 1, 2

and 3 msec, respectively.

Table 5.4.1

Peak Pressure at K4 from Trend Study

Disc Burst

Time (msec)

Peak Pressure (Barg)

1 29.99

2 41.93

3 51.60

0

10

20

30

40

50

60

0.004 0.0045 0.005 0.0055 0.006

Test 49p Trend Curve for 1,2, 3 millisec


Pressure (bar)


Figure 5.4.1 Trend Study for Graphite Disc, High Pressure Test

Optimising the Burst Time

From these results, the optimum burst time of 1.9 msec was selected and the high pressure test

(49p) was re-simulated with this burst time.

The results are given for K4 and K2 (Figures 5.4.2 and 5.4.4) and, to demonstrate the high level

of correlation, K4 is also repeated at a very small time scale (Figure 5.4.3). The dark line is the

simulated result.

7/27/2019 oto00130


22

0

5

10

15

20

25

3035

40

45

0 0.002 0.004 0.006 0.008 0.01 0.012



Pressure (bar)

1. 10


Max = 40.87

Min = 0

Figure 5.4.2 Correlation for Graphite Disc, High Pressure Test

0

5

10

15

2025

30

35

40

45

0.004 0.0045 0.005 0.0055 0.006

Standard



Pressure (bar)

1. 10


Max = 40.87

Min = 0

Figure 5.4.3 Correlation at K4 over Short Time Scale

7/27/2019 oto00130


23

0

10

20

30

40

50

60

70

80

0 0.002 0.004 0.006 0.008 0.01 0.012



Pressure (bar)

1. 3


Max = 67.30

Min = 0

Figure 5.4.4 Correlation for K2

Discussing the High Pressure Test (49p)

The correlation between the measured and simulated results is exceptionally high for Test 49p,

giving a high level of confidence that the mathematical model accurately reflects the action of

the bursting disc when a 1.9msec burst time is adopted.

However, a detailed appraisal of the results shows that, in fact, the duration of the wave at K4 is

only about 0.3 msec from the time when it starts to rise until the time that the pressure peaks.

This means that, in this particular test, the results are only sensitive to the disc action for the first

15% of the burst time. Thereafter, the disc must have continued to shatter but, with the pressurewave already decaying, this further increase in the relief capacity had no impact.

Low Pressure Test (51p)

The same burst disc model was adopted for the low pressure test i.e. with a burst time of 1.9

msec and the results are given on Figure 5.4.5.

Initially, these results show good correlation but the modelled pressure wave then decays almost

as soon as the burst disc ruptures whereas the measured wave takes about 2 msec to decay.

(The dark line is the simulated result).

7/27/2019 oto00130


24

0

5

10

15

20

0 0.002 0.004 0.006 0.008 0.01 0.012

Test 51p Linear 1.9 ms Rupture

K4 to Burst Disc Pressure Time (seconds)

Pressure (bar)

3. 12

Sheffield Model - Re-Calibrated for Test 49p

Max = 17.31

Min = 0

Figure 5.4.5 Graphite Disc, Low Pressure Test

A trend study was therefore undertaken, to determine whether the correlation would improve

with a longer burst time but in fact there was less overall agreement, (Figure 5.4.6, where the

dark line is measured).

For example, the duration of the simulated wave is similar to the measured one if the burst time

is 10 msec, but the peak pressure is over 30% higher than the measured one.

0

5

10

15

20

25

0 0.002 0.004 0.006 0.008 0.01 0.012

Test 51p Trend Curve for 5, 10, 20 millisec


Pressure (bar)

3. 12


Figure 5.4.6 Trend Study for Graphite Disc, Low Pressure Test

Effect of Incomplete Rupture

In the light of this poor correlation, the measured and simulated results were completely re-

appraised. This indicated that the burst disc model we were using (which had proved highly

suitable for the high pressure test) did not reflect the performance of a graphite disc under

marginal, low pressure conditions which were likely to result in incomplete rupture.

7/27/2019 oto00130


25

Obviously, we were not in a position to develop a new model based solely on one set of results;

nor do we feel that we would use such a model very often. However, we interrogated the

measured results diagnostically to gain insight into the possible performance of the disc.

Firstly, this review showed that the burst disc is able to provide some relief capacity almost

instantly (i.e. within 0.1 msec). This capacity only represents about 2% of the overall capacity

(and may therefore be provided by the initial cracks) but, interestingly, this was enough to limit

the pressure rise in this particular low pressure test.

Subsequently, for our own interest, we developed a bespoke capacity-model for the low pressure

test, giving the correlation shown on Figure 5.4.7. This was interesting as a correlation exercise

but overall, did not provide any further information about the way in which a graphite disc

shatters in normal conditions.

0

5

10

15

20

0 0.002 0.004 0.006 0.008 0.01 0.012

Test 51p


Pressure (bar)

3. 12


Max = 17.52

Min = 0

Figure 5.4.7 Bespoke Correlation for Graphite Disc, Low Pressure Test

Summary

Our analysis of the hydraulic modelling and measured results for the Graphite Burst Disc tests

again shows that a high level of correlation can be obtained. When the disc was subjected to a

significant, rapidly applied pressure then the (overall) burst time was 1.9 msec. However, results

are only sensitive to the disc action for the first 15% of the burst time. Thereafter, the disc must

have continued to shatter but, with the pressure wave already decaying, this further increase in

the relief capacity had no impact.In contrast, the findings from the low pressure test were not conclusive. The results suggest that

the disc provides a small relief capacity almost instantly; although we have no proof, we feel that

this is possibly in the form of the initial cracks. In turn a small relief flow developed and this was

sufficient to reduce the pressures. Thus, for most of the test, the pressure and flow changes in

the shock tube were insensitive to the performance of the burst disc.

Overall these results suggests that further work would be beneficial to investigate the 'marginal'

conditions, i.e. low pressure cases where the natural system pressures are only slightly higher

than the setting of the relief device. But this is probably low priority when put in context with the

aims of the study.

This project arose from the investigation into tube rupture within an industrial heat exchanger

and, in that context, the rapidly applied high pressure wave is far more important than the

7/27/2019 oto00130


26

marginal, low pressure wave. For such high pressure cases, the results obtained from the high

pressure test could be adopted, within an overall burst time of 1.9 msec.

5.5 STAINLESS STEEL BURST DISC TESTS

Introduction

This section continues the examination of the burst discs, this time looking at the stainless steel

discs.

Test Data

Table 5.5.1

Test Summary


4in StainlessSteel Test No. 41p - -

8in Stainless

Steel

- - Test No. 46p

As noted previously, the tests are completely different in simple terms:

The 4mm orifice givers a relatively low pressure test

The 15mm orifice gives a high pressure test

The nominal burst pressure for the discs is:

14.7 barg for the 4in stainless steel disc14.8 barg for the 8in stainless steel disc

Mathematical Model of the Stainless Steel Discs

There is a significant difference between the metal discs and the graphite ones in the way in

which they rupture and this difference is incorporated in the mathematical models:

the graphite discs will continue to shatter even if the upstream pressure subsequently drops away

with the exception of inertia effects, there is no mechanism to continue opening a metal disc if

the pressure on the upstream face is falling and is less than the downstream pressure. As seen

by the researchers, the end results is that a metal disc can finish only partially open

As noted for the graphite burst discs, the mathematical model used by PSI for the stainless steel

discs validates against API RP 520 and manufacturers' catalogue data under steady state

conditions. Additionally, the dynamic model was upgraded to include the feature noted with the

graphite tests i.e. that a small relief capacity develops almost instantly, within 0.1 msec.

Trend Study for High Pressure Test with 8" Disc (46p)

Firstly, the hydraulic model of the shock tube was re-configured to reflect the physical changes

needed to study an 8" disc e.g. the expansion piece was added. Then, a trend study was again

undertaken with the metal burst disc model, to investigate the burst time of the test disc.

7/27/2019 oto00130


27

The results on Figure 5.5.1 and Table 5.5.2 show the effect of the burst time on the pressure at

K4. The dark line is the measured result and the 3 lighter ones are the simulated result with a

disc burst time of 5, 10 and 15 msec, respectively.

Table 5.5.2


Disc Burst

Time (msec)

Peak Pressure

(Barg)

5 18.95

10 19.44

15 21.50

0

5

10

15

20

25

0 0.002 0.004 0.006 0.008 0.01 0.012

Test 46p Trend Curve for 5, 10 and 12 millisec


Pressure (bar)


Figure 5.5.1 Trend Curve for 8" Metal Disc, High Pressure Test

Optimising the Burst Time

From these results, the optimum burst time of 10 msec was selected and the results are given for

K4 (Figure 5.5.2). The dark line is the simulated result.

7/27/2019 oto00130


28

0

5

10

15

20

25

0 0.002 0.004 0.006 0.008 0.01 0.012

Test 46p 10ms Rupture


Pressure (bar)

5. 10


Max = 19.44

Min = 0

Figure 5.5.2 Correlation for 8" Metal Disc, High Pressure Test

7/27/2019 oto00130


7/27/2019 oto00130


30

Position K4 for 4in Graphite and SS Discs

-5

0

5

10

15

20

0 0.002 0.004 0.006 0.008 0.01 0.012

Time (s)

P r e s s u r e

( b a r )

K4, 51p, 4in

Graphite

K4 41p, 4in SS

Figure 5.5.4 Measured Test at K4 - Graphite and Metal Discs

Figure 5.5.4 shows the low pressure tests for both 4" discs and hence are directly comparable

but their performance is different:

The capacity15

of the graphite disc is high enough that the device offered no restriction to the

flow in the shock tube after it had ruptured. This means that the pressure at K4 dropped to

atmospheric pressure by about 8 msec

The metal disc only opened partially and offered significant restriction to the flow in the shock

tube. This means that the pressure at K4 remained at almost 10 bar. Subsequently, the pressurethroughout the shock tube dropped, as discussed in Appendix A

Summary

We have two different models for burst discs, one for graphite discs (in which the relief capacity

continues to increase to the maximum once it starts to rupture) and one for metal discs (which

only opens when there is a positive driving force). And the this study has confirmed the need for

two different models.

The results also show that the metal discs opened more slowly than the graphite discs (nominally

in 10 msec compared with only 1.9 msec for the graphite discs). And, subjectively a slower time

is expected because of the difference in the disc material and the manner in which theyopen/shatter. However, the discs studied in the high pressure tests

16are also different sizes (the

slower disc is also the larger) and so it is not possible to state whether the difference in opening

time is only attributable to the material or whether it is also a function of size.

Additionally, we have some reservations about the accuracy of the burst time for the metal disc

(10 msec). As seen with the graphite disc, the pressure wave was relieved within about 10% of

the opening time and this means, subsequently, the test became insensitive to the disc capacity.

15

Although it is impossible to say whether the graphite disc shattered completely (or whether a small

annulus remained), either way, it gave materially the full capacity

16 The high pressure tests give a more reliable burst time

7/27/2019 oto00130


31

In practice therefore the overall burst time represents an effective rupture period of about 1

msec, extrapolated to give an overall estimate of 10 msec.

Overall, the study again shows good correlation between the measured and simulated results and

we have a high level of confidence in the model. But we suggest that further work is undertaken

to examine this in more detail:

with a wider range of tests that are directly comparable,

under tests conditions that sensitive to the disc capacity and hence give a more accurate

estimation of the overall opening time

5.6 SPRING LOADED RV, HIGH PRESSURE TEST

Introduction

This section continues the examination of the relief devices, starting to look at the spring loaded

relief valves (RVs).17

Test Data

As noted previously, the 15mm orifice gives a high pressure test. The equivalent low pressure

test is discussed later in the section.

Table 5.6.1

Test Summary


2in Spring

Loaded

See Section

5.7

- Test 57p

The set pressure for the RV is 15 barg.

Mathematical Model of the Spring Loaded RVs

As noted for the burst discs, the mathematical model used by PSI for the spring loaded relief

valve validates against API RP 520 and manufacturers' catalogue data under steady state

conditions. However, the dynamic model also incorporates the unique characteristics of the valve

action, described in the manufacturers' catalogues with the specific terms 'pop open' and

'blowdown'.

In brief the cycle of an RV is:

ClosedValve starts to 'simmer' when local pressure exceeds set pressure

Valve pops open at popping point (i.e. a pressure slightly above the set pressure). When the

valve has popped open, the rated capacity is in accordance with the international codes i.e. at

110% of set pressure

Further lift is proportional to pressure to give slightly more capacity

In the initial closing phase, lift is proportional to pressure

17

The simple difference between a burst disc and a relief valve (as defined in API RP 520) is the fact that

the burst disc is a non-closing device whilst an RV is designed to automatically re-close and prevent the

further flow of fluid.

7/27/2019 oto00130


32

Then the valve blows down, with a smaller pop action, down to closure at a lower (blowdown)

pressure. The cycle is now complete with the valve re-closed.

7/27/2019 oto00130


33

Opening Time for RVs

As noted above, the 'rated' capacity18

of an RV is reached when it pops open and this is the

most important feature of the valves. We have therefore defined the opening time for RVs as

the time taken to reach the rated capacity i.e. the capacity for 110% of the set pressure, or 10%

over-pressure (10% OP).


Firstly, the hydraulic model of the shock tube was re-configured to incorporate the RV and then,

a trend study was again undertaken to investigate the opening time.

The results are given on Table 5.6.2 and Figure 5.6.1, the dark line is measured.

Table 5.6.2


RV 10% OP

Time (msec)

Peak Pressure

(Barg)

1.5 87.0

3.0 97.2

4.5 102.1

6.0 104.9

0

20

40

60

80

100

120

0 0.002 0.004 0.006 0.008 0.01 0.012

Test 57p Trend Curve for 1.5, 3 4.5 & 6millisec


Pressure (bar)


Figure 5.6.1 Trend Curve for Spring RV, High Pressure Test

Optimising Opening Time

18 The 'rated' capacity is in accordance with international codes

7/27/2019 oto00130


34

From these results, the optimum opening time of 4 msec was selected and the high pressure test

was re-simulated with this time.

The results are given for K4 and K3 (Figures 5.6.2 and 5.6.3, the dark lines are the simulated

result) and these again show very good correlation.

-20

0

20

40

60

80

100

120

0 0.002 0.004 0.006 0.008 0.01 0.012



Pressure (bar)

5. 8


Max = 100.75

Min = 0

Figure 5.6.2 Correlation for Spring RV, High Pressure Test

-20

0

20

40

60

80

100

120

0 0.002 0.004 0.006 0.008 0.01 0.012



Pressure (bar)

5. 12


Max = 101.28

Min = 0


Evaluation of the Results

Our evaluation of the tests results also highlighted two other features:

RV remains fully opening for the test period. As shown on Figure 5.6.4, the pressure in the

shock tube remains well above the set pressure of the RV (15 barg)

As might be expected, the capacity of the 2in RV is far lower than the capacity of the 4in burst

disc. Additionally, the burst disc reacts more quickly. For example, in comparable tests, the

7/27/2019 oto00130


35

graphite burst disc limited the peak pressure to about 41 barg and reduced the pressure at K4

towards atmospheric pressure. In contrast the peak pressure is over 100 barg with the 2in RV

and the pressure at K4 remains above 60 barg for the remainder of the test.

2in Spring Relief - 15mm

0

20

40

60

80

100

120

0 0.05 0.1 0.15

Time (s)

P r e s s u r e

( b a r g ) K4

Figure 5.6.4 Measured Test at K4 - Spring Loaded RV

Summary

The mathematical model for the RV is far more complex than the burst disc model as it needs to

reflect the complete performance cycle of pop open and then blowdown. But despite this we areagain able to show good correlation when with an opening time of 4 msec. However, it must be

remembered that this opening time is defined very specifically for RVs as the time taken to

reach the rated capacity i.e. the capacity for 110% of the set pressure, or 10% over-pressure

(10% OP).

5.7 SPRING LOADED RV, LOW PRESSURE TEST

Introduction

This section discusses the findings of the low pressure test on the spring loaded RV.

Test Data

The test and number is that used by Sheffield University. As noted previously, the 4mm orifice

gives a low pressure test.

Table 5.7.1

Test Summary


2in Spring

Loaded

Test 59p - Section 5.6

7/27/2019 oto00130


36


7/27/2019 oto00130


37

Evaluation of Test Results

The main feature of this low pressure test was immediately apparent when the test results were

initially appraised. Unlike the previous tests, where the pressure in the shock tube dropped after

the relief device is active (see Figure 19, for example) this test showed unstable pressure

oscillations, commonly termed valve chatter (Figure 5.7.1). This behaviour shows:

A pressure frequency of about 80 cycles per sec

The initial pressure peak was not the highest; a later pressure peak is 36% higher

2in Spring Relief - 4mm - Smoothed

0

10

20

30

40

50

0 0.05 0.1 0.15

Time (s)

Pressure

(barg)

K4

Figure 5.7.1 Measured Test Showing RV Chatter

RV Chatter

Chatter is associated with over-sizing of an RV. In the context of the system, a relatively high

relief flow develops when the valve pops open and this tend to reduce the inlet pressure. But the

pressure can drop below the blowdown pressure if the valve is too big and the relief flow is too

high. This means that the valve will close, generating a surge pressure rise when the relief flow

stops suddenly. And so a repetitive cycle can develop where opening is followed by immediate

closure, resulting in the chatter effect shown on Figure 5.7.1.

Low Pressure Test

To date, the mathematical modelling of the shock tube and the relief devices has shown very

good correlation, being able to take into account the differences in the devices themselves and

also re-configuration of the shock tube (with the expansion piece for example). However, the

accurate study of a chattering RV was well outside the scope of this project.

As a matter of interest, we therefore took a simplistic approach and looked only at the first

opening phase; we used an opening time based on the cycle frequency.19

The results are shown

19

The valve frequency was 80 cycles per second i.e. 12.5 msec and so an opening time of 6 msec was used,

slightly slower than the opening time in the high pressure test

7/27/2019 oto00130


38

on Figure 5.7.2 and Figure 5.7.3 and the overall correlation is good. But it already under-

estimates the oscillation seen at K4, i.e. near the RV.

Summary

The low pressure test on the spring loaded relief valve showed all the characteristics of valve

chatter, consistent with the performance of an over-sized valve. The detailed study of this

phenomenon was well outside the scope of this study and so only the initial pop-open phase was

studied. And this suggested that the RV opened slightly slower in this test than in the high

pressure test (6 msec compared with 4 msec). However, some pressure oscillation is still

apparent, even within this opening stage.

-10

0

10

20

30

40

50

0 0.002 0.004 0.006 0.008 0.01 0.012



Pressure (bar)

6. 8


Max = 27.38

Min = 0

Figure 5.7.2 Preliminary Correlation for RV, Low Pressure Test

-10

0

10

20

30

40

50

0 0.002 0.004 0.006 0.008 0.01 0.012



Pressure (bar)

6. 12


Max = 27.39

Min = 0


7/27/2019 oto00130


39

5.8 PILOT OPERATED RELIEF VALVE TEST

Introduction

This section discusses the final set of results, namely for the pilot operated relief valve.

Test Data

As noted previously, the 4mm orifice gives a low pressure test and the 15mm is a high pressure

test.

Table 5.8.1

Test Summary


2in Pilot RV Test 66p - Test 64p


Valve Characteristics

Although the pilot operated RV is nominally the same size as the spring loaded RV discussed in

the preceding section, the capacity of a pilot operated valve is slightly greater. It is also worth

noting that, physically, this pilot valve is more compact and most importantly the moving parts of

the valve are smaller than the spring loaded RV. And, overall, these factors mean that we

expected the pilot operated valve to react slightly more quickly than the spring loaded valve.

Additionally, the pilot configuration offers either pop-action or modulating-action; there is also a

combined option of pop-open and modulate-closed which was the configuration studied in these

tests. And again, this means that the test valves were the fastest available.Opening Time for RVs

As noted earlier, the 'rated' capacity20

of an RV is reached when it pops open and this is the

most important feature of the valves. We have therefore defined the opening time for RVs as

the time taken to reach the rated capacity i.e. the capacity for 110% of the set pressure, or 10%

over-pressure (10% OP).


As before, the hydraulic model of the shock tube was re-configured to incorporate the pilot

operated RV and a trend study was again undertaken to investigate the opening time.

The results are given on Table 5.8.2 and Figure 5.8.1, where the dark line is measured.

Table 5.8.2


RV 10% OP

Time (msec)

Peak Pressure

(Barg)

2.0 81.04

3.0 86.66

20 The 'rated' capacity is in accordance with international codes

7/27/2019 oto00130


40

4.0 90.04

0

20

40

60

80

100

0 0.002 0.004 0.006 0.008 0.01 0.012

Test 64p Trend Curve for 2, 3 & 4millisec


Pressure (bar)


Figure 5.8.1 Trend Study for Pilot RV, High Pressure Test

Optimising Opening Time

From these results, the optimum opening time of 2.5 msec was selected and the high pressure

test was re-simulated with this time.

The results are given for K4 (Figure 5.8.2, the dark line is the simulated result) and again this

shows very good correlation.

-20

0

20

40

60

80

100

0 0.002 0.004 0.006 0.008 0.01 0.012

Standard

Test 64p 10% Capacity in 2.5ms


Pressure (bar)

7. 8


Max = 84.07

Min = 0

Figure 5.8.2 Correlation for Pilot RV, High Pressure Test

Low Pressure Test

Examination of the test results for the low pressure test shows that, within 2 msec of the valve

opening, the local pressure drops back below the set pressure. This means that the valve will pop

7/27/2019 oto00130


41

open and then immediately start to modulate closed. However, this is a function of the type and

configuration of the pilot system and the accurate study of its performance was well outside the

scope of this project.

Again, as a matter of interest, we therefore took a simplistic approach and looked only at the

first opening phase; the opening time was taken as 4 msec, based on earlier findings of this study

that the devices tend to act more slowly in the marginal, low pressure tests, compared with the

high pressure ones. The results are shown on Figure 5.8.3, where the dark line is simulated.

-5

0

5

10

15

20

25

30

0 0.002 0.004 0.006 0.008 0.01 0.012



Pressure (bar)

8. 8


Max = 22.11

Min = 0

Figure 5.8.3 Correlation for Pilot RV, Low Pressure Test

RV Performance: Spring Loaded versus Pilot

Pilot operated RVs are not widely used in industry. As stated in API RP 520, "since the main

valve and pilot contain non metallic components, process temperature and fluid compatibility limit

their use. In addition, fluid characteristics such as a susceptibility to polymerisation, fouling,

viscosity, the presence of solids and corrosiveness may affect pilot reliability".

Despite this, as shown on Figure 5.8.4, their modulating action may by beneficial. The severe

chatter exhibited by the spring loaded RV is completely eliminated from the low pressure test

and replaced by stable performance.

7/27/2019 oto00130


42

Position K4 Relief Valve

-5

0

5

10

15

20

25

30

3540

45

50

0 0.05 0.1

Time (s)

P r e s s u r e (

b a r )

K4, 66p, Pilot

K4 59p, Spring

Figure 5.8.4 Measured Test at K4 - Spring Loaded and Pilot RV

Summary

The study shows that the 2in spring loaded RV popped open in 4 msec compared with 2.5 msec

for the pilot operated valve. And this finding, was unexpected. Subjectively, a slower response

was expected. However, we note that:

The pilot operated RV and the spring loaded RV are not directly similar valves. The physical

mass of the moving parts is smaller for the pilot operated valve and so the inertia effects would

be lower We do not have details of the pilot system for the valve but the findings of the study suggest that

it is the type that is characterised by pop action, followed by a modulating action.21

This means

that the opening performance will not be adversely affected by the pilot

The study also shows that the modulating action is particularly beneficial in eliminating the severe

valve chatter seen previously with the spring loaded RV. However, this is not the only way of

eliminating chatter. It can also be avoided by other means, for example, by using dynamic

simulation methods to ensure that the spring loaded RV is correctly sized in the design stage.

21 This is supported by the valve performance in the low pressure test

7/27/2019 oto00130


43

6. CONCLUSIONS AND RECOMMENDATIONS FOR FURTHER

WORK

The study was initiated as a result of concerns over the failure of pressure vessels holding liquids

when subjected to internal pressure pulses whose amplitudes and durations were above certain

limiting values. Whilst pressures higher than yield pressure may be tolerated for durations whichare short compared to some characteristic time of the structure, the effect of delayed pressure

relief is expected to extend the period of higher pressure exposure and hence may move the

transient into a hazard range.

The study emphasises a few key features associated with the pressurising source and

subsequent wave propagation, and these can be summarised as follows :

1. following gas release there is a constant pressure which develops at the liquid/gas interface

and this pressure propagates at sound speed in the liquid, being experienced by the structure

following its passage. The constant pressure will depend on the rate of flow of gas into theinterface region and the compressibility of the downstream liquid. This is characterised by a

plateau region on the upstream transducers (K2, K3) in the study.

2. On arrival at the relief device, the pressure wave opens the device after a certain time and

this then propagates an expansion wave upstream which cancels the high pressure. For

instantaneous opening, the duration of the high pressure exposure upstream is solely

dependent on the distances involved and the sound speed in the liquid. For water filled vessels

this corresponds approximately to 0.7msec/meter pathlength, where the pathlength is the

round-trip distance from any point to the relief device.

3. Any delay in the opening of the device will result in the reflection of a pressure wave of

twice the amplitude upstream and therefore the exposures which should be of concern are the

initial plateau pressure, the reflected wave pressure and the duration associated with each. The

plateau duration has been referred to above. The reflected wave duration will depend on the

opening time of the device, and a number of data traces show this effect.

4. The analysis has indicated that for the devices tested, times to full opening are 1.9 - 10 msec

for the bursting discs and 2.5 - 4 msec for the RVs. The differences in the relative scale of

these two families of devices has already been emphasised and is reiterated here.

5. Whilst the study used oversized discs and undersized RVs, the quoted times should strictly be

applied to pressurising systems for which the devices would represent the correct sizes.

6. It is considered that the timescales for opening are significant and are comparable with the

response times for structures as they move toward yield when exposed to a step pressure

rise. A transient analysis is therefore considered a worthwhile exercise during the design of

low pressure vessels which may be exposed to high pressure transients, whose relief is

7/27/2019 oto00130


44

governed by RV or bursting disc behaviour. In summary, the opening times to full relief capacity

for a number of the devices tested are collected together in Table 6.1.

Table 6.1Measured relief times for selected relief devices

Relief device Conditions Relief time - msec

Graphite 4" high pressure 1.9

Stainless steel 8" high pressure 10

Stainless steel 4" low pressure 10

Relief valve high pressure 4

Relief valve low pressure 6

Pilot operated relief valve high pressure 2.5

Pilot operated relief valve low pressure 4

We recommend that further work is undertaken to determine the dependency of these findings

on pressure conditions. Although we have nominally undertaken a detailed analysis of two sets

of results (high and low pressure), only the high pressure tests were effective in terms of the full

opening of the devices. We therefore suggest that we take advantage of the fact that Sheffield

University already have another set of measured data available (8mm tests) and repeat the

detailed analysis on this information.

To determine whether the apparently slow opening of the metal disc is due predominantly to its

size we also suggest that we analyse the existing test data for the 4" metal disc.

Turning to the RVs, we feel that there are two areas of investigation available. The first, and

potentially the most conclusive, would be to undergo further physical testing on different valvesizes and types. The second would combine the skills of detailed analysis and literature review to

explore the basis for the longer opening times for the RVs that were given in the IP guidelines

and in other research papers. The work should focus on points of comparison e.g. establishing

whether the valves were the same type (pop action) and size.

7/27/2019 oto00130


45

7. APPENDIX

7.1 PHOTOGRAPHS

A number of photographs illustrate the key components used in carrying out the experimental

study.

7/27/2019 oto00130


46

Plate 1. Working area for shock tube control and data collection

7/27/2019 oto00130


47

Plate 2 Aluminium driver diaphragm controlling driver operating pressure - before and after rupture

Plate 3. Driver diaphragm in place after firing - driver section on the right.

7/27/2019 oto00130


48

Plate 4. The three discharge orifices used in the tests

7/27/2019 oto00130


7/27/2019 oto00130


7/27/2019 oto00130


51

Plate 8. Expansion section for 8" stainless bursting disc.

7/27/2019 oto00130


52

Plate 9. Graphite disc holder in place at outlet end of shock tube.

7/27/2019 oto00130


53

Plate 10. Samples of graphite and stainless bursting discs after firing

7/27/2019 oto00130


54

Plate 11. 2" relief valve body used in valve tests.

7/27/2019 oto00130


55

7.2 DATA CHARTS

The data charts represent up to 50 msec of pressure data following driver disc rupture.

Full records are actually 250msec in duration.

7/27/2019 oto00130


56

-2 -1 0 1 2 3 4 5 6 7 8 9 10

-5

0

5

10

15

20

25Test 39 4mm orifice, plastic

film at outlet

P r e s s u r e

- b a r

msec

K1 K2K3 K4

CHART A.1

-2 -1 0 1 2 3 4 5 6 7 8 9 10

-10

0

10

20

30

40

50

608mm orifice, plastic

film at outlet

P r e s s u r e

- b a r

msec

Test 38

K1 K2 K3 K4

CHART A.2

7/27/2019 oto00130


57

-2 -1 0 1 2 3 4 5 6 7 8 9 10

-20

0

20

40

60

80

100

12015mm orifice, plastic

film at outlet

P r e s s u r e

- b a r

msec

Test 37

K1 K2 K3 K4

CHART A.3

-2 -1 0 1 2 3 4 5 6 7 8 9 10-5

0

5

10

15

20

25

304mm orifice, 4"graphite disc

P r e s s u r e

- b a r

msec

Test 51

K1K2 K3

K4

CHART A.4

7/27/2019 oto00130


58

-2 -1 0 1 2 3 4 5 6 7 8 9 10

-10

0

10

20

30

40

50

60

708mm orifice, 4"

graphite disc

P r e s s u r e

- b a r

msec

Test 50

K1 K2K3

K4

CHART A.5

-2 -1 0 1 2 3 4 5 6 7 8 9 10

-20

0

20

40

60

80

100

12015mm orifice, 4"

graphite disc

P r e s s u r e

- b a r

msec

Test 49

K1 K2

K3

K4

CHART A.6

7/27/2019 oto00130


59

-2 -1 0 1 2 3 4 5 6 7 8 9 10-5

0

5

10

15

20

25

304mm orifice, 6"

graphite disc

P r e s s u r e

- b a r

msec

Test 55

K1

K2

K3

K4

CHART A.7

-2 -1 0 1 2 3 4 5 6 7 8 9 10

-40

-20

0

20

40

60

808mm orifice, 6"

graphite disc

P r e s s u r e

- b a r

msec

Test 54

CHART A.8

7/27/2019 oto00130


60

-2 -1 0 1 2 3 4 5 6 7 8 9 10

0

20

40

60

80

100

12015mm orifice, 6"

graphite disc

P r e s s u r e

- b a r

msec

Test 53

K1 K2K3 K4

CHART A.9

-2 -1 0 1 2 3 4 5 6 7 8 9 10 11 12-5

0

5

10

15

20

25

304mm orifice, 4"

stainless steel disc

P r e s s u r e

- b a r

msec

Test 41

K1 K2 K3 K4

CHART A.10

7/27/2019 oto00130


61

-2 -1 0 1 2 3 4 5 6 7 8 9 10

-10

0

10

20

30

40

50

60

70

80

P r e s s u r e

- b a r

msec

Test 42 8mm orifice, 4"


K1

K2 K3 K4

CHART A.11

-2 -1 0 1 2 3 4 5 6 7 8 9 10

-20

0

20

40

60

80

100

120

P r e s s u r e

- b a r

msec

Test 40 15mm orifice, 4"stainless steel disc

K1 K2K3 K4

CHART A.12

7/27/2019 oto00130


62

-2 -1 0 1 2 3 4 5 6 7 8 9 10-10

0

10

20

30

40

50

P r e s s u r e

- b a r

msec



K1 K2K3

K4

CHART A.14

-2 -1 0 1 2 3 4 5 6 7 8 9 10-10

0

10

20

30

40

P r e s s u r e -

b a r

msec



K1 K2K3 K4

CHART A.13

7/27/2019 oto00130


63

-2 -1 0 1 2 3 4 5 6 7 8 9 10

-20

0

20

40

60

80

100

120

P r e s s u r e

- b a r

msec



K1 K2 K3K4

CHART A.15

-10

12

34

56

78

910

1112

1314

15

-20

-10

0

10

20

30

40

504mm orifice, 2"

relief valve

P r e s s u r e

- b a r

msec

Test 59

K1K2 K3 K4

CHART A.16

7/27/2019 oto00130


64

0 5 10 15 20 25 30 35 40 45 50

-20

-10

0

10

20

30

40

504mm orifice, 2"

relief valve

P r e s s u r e

- b a r

msec

Test 59

CHART A.17

-10

12

34

56

78

910

1112

1314

15

-10

0

10

20

30

40

50

60

708mm orifice, 2"

relief valve

P r e s s u r e

- b a r

msec

Test 58

K1K2 K3 K4

CHART A.18

7/27/2019 oto00130


65

0 5 10 15 20 25 30 35 40 45 50

-10

0

10

20

30

40

50

60

708mm orifice, 2"

relief valve

P r e s s u r e

- b a r

msec

Test 58

CHART A.19

-10

12

34

56

78

910

1112

1314

15-20

0

20

40

60

80

100

120

P r e s s u r e

- b a r

msec


relief valve

K1 K2 K3K4

CHART A.20

7/27/2019 oto00130


66

0 5 10 15 20 25 30 35 40 45 50

-20

0

20

40

60

80

100

120

P r e s s u r e

- b a r

msec


relief valve

CHART A.21

-10

12

34

56

78

910

1112

1314

15

-10

0

10

20

30

40Test 62 4mm orifice, 2"

bellows relief valve

P r e s s u r e

- b a r

msec

K1K2 K3

K4

CHART A.22

7/27/2019 oto00130


67

0 5 10 15 20 25 30 35 40 45 50-10

0

10

20

30



P r e s s u r e

- b a r

msecCHART A.23

-10

12

34

56

78

910

1112

1314

15

-20

-10

0

10

20

30

40

50

60

70

P r e s s u r e

- b a r

msec



K1 K2 K3 K4

CHART A.24

7/27/2019 oto00130


68

0 5 10 15 20 25 30 35 40 45 50-20

-10

0

10

20

30

40

50

60

70

P r e s s u r e

- b a r

msec



CHART A.25

-10

12

34

56

78

910

1112

1314

15

-20

0

20

40

60

80

100

12015mm orifice, 2"


P r e s s u r e

- b a r

msec

Test 60

K1

K2 K3 K4

CHART A.26

7/27/2019 oto00130


69

0 5 10 15 20 25 30 35 40 45 50

-20

0

20

40

60

80

100

120Test 60

15mm orifice, 2"


P r e s s u r e -

b a r

msecCHART A.27

-10

12

34

56

78

910

1112

1314

15

-10

0

10

20


pilot assisted

relief valve

P r e s s u r e

- b a r

msec

K2 K3K4

K1

CHART A.28

7/27/2019 oto00130


70

0 5 10 15 20 25 30 35 40 45 50-10

0

10

20

304mm orifice, 2"

pilot assisted

relief valve

P r e s s u r e

- b a r

msec

Test 66

CHART A.29

-10

12

34

56

78

910

1112

1314

15

-10

0

10

20

30

40

50

60

P r e s s u r e

- b a r

msec

8mm orifice, 2" pilotassisted relief valve

Test 65

K1 K2 K3

K4

CHART A.30

7/27/2019 oto00130


71

0 5 10 15 20 25 30 35 40 45 50-10

0

10

20

30

40

50

60

P r e s s u r e

- b a r

msec


Test 65

CHART A.31

-10

12

34

56

78

910

1112

1314

15

-20

0

20

40

60

80

100

P r e s s u r e

- b a r

msec


Test 64

K1K2 K3 K4

CHART A.32

7/27/2019 oto00130


72

0 5 10 15 20 25 30 35 40 45 50-20

0

20

40

60

80

100

P r e s s u r e

- b a r

msec


Test 64

CHART A.33

7/27/2019 oto00130


73

7.3 VIDEO TAPE RECORD

The following is a brief review in sequence of the content of the 13 minute video tape

accompanying the report.

1. Panning view of shock tube with driver section to the left and water filled section at theright. At the right hand end can be seen the conical expansion and holder for the 200 mmstainless steel bursting disc.

2. Close-up of the driver aluminium diaphragm after firing followed by one of the Kistler transducer amplifier boxes.

3. Close-up of 200 mm expansion section showing the bursting disc in situ.

4. Close-up of burst diaphragm region showing 'petalled' diaphragm after firing, the water inlet

used for tube filling, the driven section with a discharge orifice in place and the surfactantreservoir.

5. Panning view of high speed camera set-up and associated computer for image downloading.

6. A panning view of the control area and pressure data acquisition system.

7. A sequence of views showing the materials used in tube firing and including :aluminium diaphragms before and after firinggraphite bursting discs before and after rupture100 mm and 200 mm stainless steel bursting discs before and after rupturethe set of 3 discharge orifices used at the entrance to the water filled section.one of the relief valves used for testing showing the outlet flange of the valve.

8 Two sequences follow showing the operation of the 200 mm stainless steel and 150 mmgraphite bursting discs at normal video rates.

9 The final sequences show the same discs bursting but taken at a framing rate of 4000 frames/sec.

7/27/2019 oto00130


7/27/2019 oto00130


oto00130

Documents