Leak Test on High-Speed Separator
Master Thesis Worker: Yasaman Saffari
Industrial Supervisor: Mattias Nyqvist, Alfa Laval
KTH Supervisor: Prof. Mats Westermark, KTH (Royal Institute of Technology) KTH, School of Chemical Science and Engineering
June 2011
Abstract
High speed separators from Alfa Laval are widely in use for processing flammable and non-flammable liquids. The following work is focusing on the case of non-flammable liquid as the process liquid in case the working area around the equipment may contain quantities of explosive gases.
As stated by Alfa Laval documentation, the major risk is leaking of the explosive atmosphere into the separator from the surrounding environment which may result in producing zone 1 or zone 2 of hazardous area classification.
Zone 1: Area in which an explosive gas-air mixture is likely to occur for short periods in normal operation.1
Zone 2: Area in which an explosive gas-air mixture is not likely to occur, and if it occurs it will only exist for a very short time due to an abnormal condition.1 According to Alfa Laval design package, there is a need of continuous inert gas injection into the separator during the process in order to reduce the oxygen concentration and keep it in the safe level (inert gas purging) and this policy is aimed to meet the requirements of ATEX-directive 94/9/EC/2003.
The objective of the current thesis is a wish to have a better understanding of the potential risks, evaluating them and try to find ways to ease the process. The outcome can be useful to make a basic instruction for further tests and simplifications as well.
The separator GTN 50 is selected and hydrogen (1% concentration) is used to simulate the explosive atmosphere. The result of the tests indicates that the cooling down stage after normal operation is the only period in which hydrogen will leak into the separator, frame top part and it should be cleaned up before the next start up.
A number of recommendations -Ventilation to the fresh air, Water discharges, Pressurized air injection- are also being tested and discussed and ventilation to the fresh air and injection of pressurized air seems to be applicable A Standard Testing Flow chart is suggested and calculation on real case is considered. A number of additional ideas are also included in the last section.
Key words:
Leak test, Nonflammable liquid, Hazardous area, Centrifuge, Separator, Hydrogen, Hydrogen sensor
1
Table of content
1 Introduction & Background ... 6
2 Aim ... 7
3 Test Environment & Test Object: ... 8
3.1 Trace gas selection ... 8
− Helium ... 8
− Hydrogen ... 9
3.2 Hydrogen as the trace gas: ... 9
3.3 Test Object overview ... 10
3.4 Measuring instruments ... 13
− Calibration: ... 14
4 Pre-Tests ... 16
4.1 Guide to read the graphs ... 16
− Horizontal Axis: Time and Date scale: ... 16
− Vertical Axis: Graph Components scale ... 16
− Channel naming ... 17
4.2 Allowed Pressure loss ... 18
4.3 Sludge tank-leveling tank connection size ... 20
5 Testing Procedure: ... 21
5.1 Stand-Still test ... 24
5.2 Stand-By test ... 26
5.3 Normal Operation Test ... 28
5.4 Empty Leveling Tank Test ... 30
6 Results and Discussion ... 34
6.1 Testing Procedure flowchart: ... 35
7 Methane case study ... 36
7.1 Analyze ... 37
8 Recommendations ... 38
− Proper ventilation to the fresh air ... 38
− Water Discharges ... 45
− Injection of Pressurized air before start up ... 45
9 Further work ... 48
− Testing other machines: ... 48
− Flammable process liquid: ... 48
− Oxygen Sensor: ... 48
10 References ... 49
Appendix I: Hydrogen injection and Air properties ... 50
Appendix II: Sensors ... 51
Hydrogen sensor ... 51
Relative Humidity/Temperature Sensor ... 55
Appendix III: Detailed Graphs ... 56
Stand by test ... 56
Normal operation ... 57
Empty leveling tank ... 61
Detailed Graphs for “Recommendation” section: ... 62
Table of Figures
Figure 1: Application Selection for Trace Gas [7] ... 8Figure 2: Welded Connections ... 10
Figure 3: Inside the container, Overview ... 11
Figure 4: Sludge tank outlet ... 12
Figure 5: Water Lock ... 12
Figure 6: Frame top part measurment ... 13
Figure 7: GearBox measurment ... 14
Figure 8: Other sensors ... 15
Figure 9: Sludge tank-Leveling tank connection ... 20
Table of Drawings
Drawing 1: General Sketch ... 23Drawing 2: Empty Leveling tank sketch ... 31
Drawing 3: Open outlet 514, 642 & sludge tank ... 39
Drawing 4: Open outlet 514 & 462 sketch ... 41
Drawing 5: Open sludge tank sketch ... 43
Drawing 6: Water discharges sketch ... 46
Table of Graphs
Graph 1: Pressure Loss test ... 19
Graph 2 : StandStill test ... 25
Graph 3: Standby test ... 27
Graph 4: Normal Operation test ... 29
Graph 5: Empty Leveling tank test, Operation ... 32
Graph 6: Empty Leveling tank test, Over view ... 33
Graph 7: Open outlet 514, 462 & sludge tank ... 40
Graph 8: Open outlet 514 & 462 ... 42
Graph 9: Open sludge tank ... 44
Graph 10: Water discharges ... 47
Graph 11: Standby test, detailed ... 56
Graph 12: Normal operation test, detailed overview ... 57
Graph 13: Normal Operation test, cooling down, pressures ... 58
Graph 14: Normal Operation test, cooling down, temperatures ... 59
Graph 15: Normal Operation test, vibration ... 60
Graph 16: empty Leveling tank test, detailed ... 61
Graph 17: Open outlet 514, 462 & sludge tank ... 62
Graph 18: Open outlet 514, 462 ... 63
Graph 19:Open sludge tank ... 64
Graph 20: Water discharges, overview ... 65
Graph 21: Water discharges, Pressure ... 66
Graph 22: Water discharges, Pressure & Temperatures ... 67
1 Introduction & Background
Flammability limit is one of the most important parameters of explosive gas-air mixtures.
This information is vital to prevent explosion accidents during storage, transport, handling and processing explosive gases [1] [2]. According to many international and European regulations, a gas or gas mixture is classified as flammable if it has the chance of explosion in a special range in mixture with air at atmospheric conditions [1].
There are two flammability limits, Lower Flammability Limit (LFL) and Upper Flammability Limit (UFL). For explosive gas-air mixture with ratios between LFL and UFL, flame propagation is possible and there is a chance of fire and explosion [3]. It should be considered that flammability limit and explosion limit are interchangeable terms [1] and they are not pure physical properties of components. As a matter of fact, they are physical- chemical characteristic s of explosive gases and vapors from flammable liquids, which depend on several factors (ex. Heat losses by the flame) [3].
To develop a safe procedure for treating explosive gases and vapors, it is essential to know the flammability limits and try to stay out of the explosive region. A very common way to reduce the likelihood of fire and explosion is inerting the process by adding an inert gas to the explosive gas-air mixture. The inert gas addition dilutes the mixture and reduces the oxygen concentration below the limiting oxygen concentration (LOC).2 The inert gas is usually nitrogen or carbon dioxide and steam may also be used in some cases [2] [4].
Currently, nitrogen is the inert gas which is in use by Alfa Laval for its high speed separators, when they are installed in hazardous area and when they are processing flammable liquid. Chance of sparkles made by separator is the main concern because it is the third component needed to start a fire (explosive gas+oxygen+sparkle). As a result, inerting s is used to reduce oxygen concentration and keep the process in the safe level. However, this inerting system may affect the competitiveness of Alfa Laval in the market in a negative way by making the separators more expensive and complicated.
This work is a try to investigate the separators processing non flammable liquids while they are surrounded by hazardous environment classified as zone 1 and 2.3 The work is done by making a leak test.4
In the first place, the chance of leak into the separator is examined. In addition, the main problematic steps are founded. Lastly, a number of practical ways out of the problem also are tested.
Briefly, the leak is happening during the cooling down step, when the machine has been shut down after normal operation and the explosive gas leaks into the frame top part. The major risk is the next start up and the frame top part should be cleaned up from hydrogen.
This is possible with the injection of pressurized air.
Interestingly, separator will not suck in any explosive gas if it has a connection to the fresh air.
2 LOC is the minimum amount of oxygen needed to start a flame.
3 Definitions are given in Abstract section.
4 Leak detection is a non-destructive test performed to verify accordance of materials and components with the
2 Aim
The aim of the present work is to investigate to what extent the high-speed separators designed for ATEX environment are inherently safe when they have been surrounded by explosive environment. As an experimental setup, the amount of leakage from the surrounding atmosphere into the separator is to be studied for different operational conditions and the necessity of continuous inert gas injection should be discussed. A number of possible ways for leak prevention should be investigated and tested.
Flammable process liquid should not considered within this work.
3 Test Environment & Test Object:
As it mentioned in section “2 Aim”, the test should be designed to examine how much a flammable surrounding will leak into the separator while it is treating a nonflammable feed without having the inert gas injection system.
Obviously, the very beginning step is designing a suitable test environment for simulating the real case. Trace gas is playing a key role while its physical properties, since ease of penetration and detectability is affecting the selection of measuring instruments, concentration level and the test conditions as well. Moreover, designing an appropriate test environment and test object are of great importance while they can affect the test results directly.
3.1 Trace gas selection
For selecting a trace gas for the leak test, HVAC and Automotive industries have been investigated. Leak test is widely used in these two industries for checking vital systems like brakes, Engine’s final test, hydraulic valves, air conditioning unit, air bags and even steam turbines and condensers. [5][6]
However, the main focus for mentioned areas are mostly leaks from inside to outside environment, it sounds applicable to decide about the trace gas based on their principles.
In compare with traditional detection methods as water dunk, Pressure decay, etc; the trace gas detection could be assumed as the newest, most sensitive method. Helium and Hydrogen are two trace gases widely being used to perform leak tests. Figure 1 is giving a rough estimation for application selection.
Figure 1: Application Selection for Trace Gas [7]
− Helium
It is the traditional and more common trace gas then the hydrogen. The most important benefit is that Helium is a safe gas. It is nonflammable and the leak test is nondestructive.
However, it is the lightest inert gas and its natural concentration in air is low but even this small presence in the background gas may slow down the detection process. Furthermore, it is
an expensive gas (normally 100% Helium should be used) and the detection instruments are also costly. [6][8]
− Hydrogen
It is recently used for leak tests. With molecular weight half of Helium, very low concentration in air and higher diffusivity rate, it is a suitable gas for leak tests. The background concentration of Hydrogen is neglectable and will not intrude the measurements.
[9] A number of properties for Hydrogen and Helium are shown in Table 1. Adding to these points, Hydrogen gas and its sensors are cheaper in comparison with Helium which makes it strongly competitive. [6][7][8]
Based on all the information above Hydrogen has been selected as a trace gas for this test.
The explosive environment will be avoided by keeping the concentration of Hydrogen below LEL. Moreover, in case of any problems, being the lightest element and its fast diffusivity rate will result in rapid dissipation and dilution of Hydrogen.
Table 1: Physical Properties of Helium and Hydrogen Molecular
weight (g/mol)
Background concentration in air,
ppm [8]
Molecular velocity [10]
Flammability range (vol% in air)
[11] [12]
Gas price [13]
Helium 4 5 — Non flammable —
Hydrogen 2 0.5 2 times faster than
Helium (20m/s)
LEL: 4%
UEL: 75% 1/3 of Helium
In case of Hydrogen as the trace gas, it is vital to be sure that the test condition is always in the safe region and far enough from explosive situation. To achieve this, different standards has been reviewed to decide about the suitable hydrogen concentration for this test.
3.2 Hydrogen as the trace gas:
As it is shown in Table 2, concentration below 3.8% assumed to be nonflammable.
Table 2 Explosion Limits of Hydrogen Measured by Different Standards [14]
DIN 51649 EN 1839 (T) EN 1839 (B) ASTM E 681
LEL (H2-Air) 3.8 3.6 4.2 3.75
UEL (H2-Air) 75.8 76.6 77.0 75.1
On the other hand, according to Hydrogen safety data sheets, concentration higher than 10% of LEL will be dangerous in case of accidents and personnel exposure [15]. Regarding the facts that our test environment is firstly, a completely sealed area and secondly, located in an explosion-cell (with 0.5meter thick concrete wall), 1.0-1.1% of Hydrogen is selected for the test environment.
To provide the required Hydrogen concentration, gas code “FORMIER 10” by AGA has been used which is a mixture of 90%N2+10%H2. [16]
Detailed information about Atmosphere properties and the mixture being used for providing the required atmosphere are available in Appendix I.
3.3 Test Object overview
Undoubtedly, a sealed environment is needed to keep the concentration of Hydrogen constant during a test. A container with gaskets around its door is selected for this purpose and extra rubber seal strip is also added. Several connections to the outside in different sizes are predicted and carefully welded (Figure 2). The screw connections are all sealed using proper sealing glues.
The separator, the sludge tank and the leveling tank are located in the container (Figure 3).
The sludge tank is connected to the outside through a water column to prevent the machine sucking back fresh air during operation. In addition, a ∩-shape plastic hose is connected to the sludge tank (outside the container) makes it possible to see the water elevation in the sludge tank during a test (Figure 4).
There is also a fan coil for mixing purpose and providing a uniform concentration of Hydrogen all over inside the container (Figure 3). The fan can be used for cooling/heating the test surrounding in case it is needed.
The cables of sensors are connected to the outside by passing through a U-shape water lock in order to keep the sealing of the container (Figure 5).
Figure 2: Welded Connections
Separator
Sludge Tank
Leveling Tank Fan
Figure 3: Inside the container, Overview
Water column
≈20cm height
∩-shape plastic hose
U-shape water lock
Figure 4: Sludge tank outlet
Figure 5: Water Lock
3.4 Measuring instruments
Several thermocouples (type K), pressure gauges and hydrogen sensors are being used to control and monitor the test properties in the Separator frame top part, gearbox part and inside the container.
For the frame top part, a test pipe designed which is shown in Figure 6. A Hydrogen sensor, a Relative Humidity and Temperature sensor and a pressure gauge are mounted on it.
Furthermore, one thermocouple is reading the temperature through a direct hole on the frame top part (Figure 6).
Test Pipe
Direct Thermocouple
H2 sensor
Pressure Gauge
RH/T sensor
2 extra connections
Figure 6: Frame top part measurment
For the gear box, one hydrogen sensor, one pressure gauge, two thermocouples for measuring the temperature of top neck bearing, one thermocouple for measuring oil temperature and one thermocouple for air be used and they all are directly installed on the gearbox lid (Figure 7).
Apart from measuring instruments, 3 extra connections are also installed, two on the frame top part and one on the gearbox. One is used for injection of pressurized air and the other two is working as the air outlet; they are being used to clean up the separator from hydrogen after each test (Figure 6 & Figure 7).
Additional pressure gauges are connected to the outlet 514 (1) and sludge tank (2). Also, pressure gauges and thermocouples are installed for Feed flow and Product flow (3).
Vibration sensor (4) is monitoring the vibration of separator (Follow the numbers on Figure 8).
For monitoring the conditions inside the container, two sets of Hydrogen sensors along with thermocouples (5 & 6, one set close to the fan and another on the top of the separator) and one pressure gauge (7) are being used (Follow the numbers on Figure 8).
More detailed information about hydrogen sensor and RH/T sensor is available in Appendix II.
− Calibration:
Both types of the Hydrogen sensors and Relative Humidity sensor were received calibrated and ready to install. Pressure Gauges and Flow meter were calibrated in AlfaLaval Laboratory.
Gear Box
Pressure Gauge H2
Sensor
Extra connection Thermocouples
Figure 7: GearBox measurment
1
2 3
4 5
6
7
Figure 8: Other sensors
4 Pre-Tests
Before talking about tests and graphs, it is necessary to take a look on the scale of the graphs and its details first.
4.1 Guide to read the graphs
− Horizontal Axis: Time and Date scale:
The graph is based on “real time” scale which is located horizontally beneath the graph.
The following example makes it more understandable.
In this time scale, 18:00 means 06:00 pm at 11/05/2011 and it should not be mixed up with passing 18 hours of the test.
− Vertical Axis: Graph Components scale
Each graph may enclose different components based on the selection of user (hydrogen concentration, pressure, temperature, etc). The scale for each component is shown by a ruler located vertically on the left/right hand side of the graph.
Horizontal ruler, Zoom
Needless to say, in case that graph contains Pressure, Temperature…, other vertical rulers will be added to scale them.
− Channel naming
The following table contains the name of each channel and the exact parameter which it is measuring and the location of the measuring instrument. The naming here are the same as the graph parameters.
Table 3: Channel Naming system
Channel name on graph Measured parameter location
P measurment
AirUnderCapP Pressure, bar Pressure gauge on the test pipe GearBoxAir Pressure, bar Pressure gauge on the Gear Box lid FeedFlowP Pressure, bar Pressure gauge on the feed pipe ProductFlowP Pressure, bar Pressure gauge on the product pipe Flow514 Pressure, bar Pressure gauge for the outlet 514 InsideContainer Pressure, bar Pressure gauge in the container SlamtankP Pressure, bar Pressure gauge on the sludge tank
H2
concentration measurements
Hyd.Fan Concentration, % Hydrogen sensor located close to the Fan inside the container
Hyd.Centrifuge Concentration, % Hydrogen sensor located on the top of separator in the container
Hyd.TestPipe Concentration, % Hydrogen sensor on the test pipe Hyd.GearBox Concentration, % Hydrogen sensor on the Gear Box lid
m3, flow scale, belongs to blue line which shows the changes of flow
%, concentration scale, belongs to the green, light green and light blue lines which show H2 concentration
Table 3: Channel Naming system (continued)
Channel name on graph Measured parameter location
RH&T sensor HumidityUnderCap % Relative Humidity/Temperature sensor on the test pipe
AirUnCap Temperature, °C
Temperature measurements
AirGearBox Temperature, °C Thermocouple for Gear Box air
OilGearBox Temperature, °C Thermocouple immerged in the Gear Box oil NeackBearing1 Temperature, °C Thermocouple for the Neck Bearing
NeackBearing2 Temperature, °C Thermocouple for the Neck Bearing FeedFlowT Temperature, °C Thermocouple on the Feed pipe
airUndCapT Temperature, °C Thermocouple for the frame top part through the direct hole
InsideT1 (Alarm) Temperature, °C Thermocouple close to Fan in the container InsideT2 (Cent.) Temperature, °C Thermocouple on the top of separator inside
the container
ProductFlowT Temperature, °C Thermocouple on the Product pipe
Flöde m3 Feed Flow meter
Vib2 Vibration, mm/s Vibration sensor located on the separator body with a magnet
Pressures are Gauge Pressure.
4.2 Allowed Pressure loss
Separators provided for Atex environment should meet a Pressure Loss pre-test to be sure that they are tight enough.
According to AlfaLaval Document Number “596687, Testing Instruction”, the maximum allowed pressure loss for the separator under 400mm H2O pressure is 15% (60mm H2O) during the first hour. In other words, the pressure should not drop below 340mm H2O. In order to meet the worst case (ex. Used machines), 15% pressure loss in first 15 minutes was applied for the test object.5
All the outlets of separator and the connections to Feed and Product flows should be closed.
As it can be seen from Graph 1, the pressure is 340mm H2O after 16 minutes which fits in the requirements.
5
testm036(30) INTAB Interface-Teknik AB 0,050
0,045
0,040
0,035
0,030
0,025
0,020
0,015
0,010
0,005
0,000 bar
16:40 2011-04-29
16:45 2011-04-29
16:50 2011-04-29
2011-04-29 16:35:40 C1 2011-04-29 16:35:47 18,43 m 2011-04-29 16:54:16
AirUnderCapP = 0,040 bar
AirUnderCapP = 0,034 bar
GearBoxAir = 0,038 bar
GearBoxAir = 0,032 bar
16:37:30 16:53:15
Graph 1: Pressure Loss test
4.3 Sludge tank-leveling tank connection size
As it can be seen in Figure 9, a pipe is connecting the sludge to the leveling tank. When the pressure inside the separator exceeds 400mm H2O, separator will lose pressure via this pipe and air will bubble in the leveling tank. As a matter of fact, the size of this pipe is important because it affects the amount and speed of loosing pressure at the discharge time.
According to a number of tests that have been done, the size of the pipe is too big for this system. According to Alfa Laval experts6, the orifice with ≈5mm hole is suitable.
The case is a bit problematic for this test while there is no pump at the outlet of the sludge tank and the pressure loss becomes very important to control the level of water in the sludge tank and to prevent air suction into the separator from the sludge tank outlet. Subsequently, the size of open area being reduced by using a gate valve (Figure 9) and the proper opening value has been found by performing tests and observing the behavior of the system.
6
Figure 9: Sludge tank-Leveling tank connection
Metal pipe, Connection between two tanks
Gate valve
5 Testing Procedure:
The tests will be done in an explosion cell (with 0.5 meter cement wall and size of width=2.6, length=3, height=2.6) in order to provide safety in case of problem. The test object (GTN 50) is located in a container which is closed and filled with 1.0-1.1% Hydrogen during the tests. The Data of the test condition is monitored via INTAB software and is gathered through 24 channels, consists information on Hydrogen concentration (4 sensors), Temperature (9 thermocouples), Pressure (7 gauges), Relative Humidity and Temperature (1 sensor), Flow rate (1 flow meter) and one Vibration sensor.
The tests are generally laid in three categories:
− Stand Still test
− Standby test
− Normal operation test
Drawing 1 is applicable for these 3 series of tests.
In addition, another test has been done
− Empty Leveling-Tank
More detailed information on the test procedure for each test is available in the following sections.
The operational characteristics of the separator are constant for all the tests. The Feed flow fixed on 14-15 m3/hr And the product flow pressure around 1.5-1.6 bar. The discharge interval is 15 minutes and each operational test goes through 10-12 discharges.
The separator should be cleaned up from Hydrogen gas after each test which is done by injection of pressurized air into the frame top part for a couple of hours. The extra connections prepared on the frame top part, gearbox and sludge tank should be open during the air injection to let the hydrogen run out of separator parts.
Notes:
- The Fan in the container should work the entire test long in order to have homogeneous concentration of hydrogen in all parts of container. However, the hydrogen concentration is slowly decreasing during the test because hydrogen leaks out from the container.
- To have a clean-start up, the pressurized air injection for is recommended after all the tests even the ones which don’t show any Hydrogen leak.
- The machine should be cooled down completely in all parts before starting a new test to keep the test conditions constant and comparable.
- A small amount of water splashes out from the leveling tank during the first two-three discharges. So, it is necessary to check the water level in the leveling tank after each test and keep it constant in order to have a same pressure conditions.
- For all the tests, the hydrogen concentration in the container should be 1.0-1.1%
- The temperature in the frame top part and gearbox part is increasing during the test.
While the hydrogen sensors are sensitive to high temperatures, it is recommended not
to continue a test more than 70°C for the gearbox part which is the hottest part for these tests.
- The cooling water for gearbox oil should be open during the running mode.
- When the machine is off, a part of the hydrogen that leaks into it may stuck somewhere inside the separator and not being detected by the sensors. As a result, it is necessary to run the motor for a few minutes (ex. 6 minutes) to have a complete circulation both in the frame top part and gearbox and to let the sensors detect the whole amount of hydrogen.
- It is of importance to have a continuous supervision on the maximum water level in the sludge tank to be sure that it is not filled up by water. This is possible with the U-pipe provided outside the container. (Figure 4)
- Each test condition should be repeated at least two times for the certainty about the result.
- It is recommended to check the Hydrogen sensors after every two tests to be sure that they are still clean. However, no problem founded during these test series.
Drawing 1: General Sketch
T: Thermocouple P: Pressure gauge H: Hydrogen sensor
RH: Relative Humidity sensor F: Flow meter
Container
P5 T6 P6
T1 , P1 , F T2 , P2
T3,T4, P3 , RH , H1 T10 , P7, H4 T9 , H3
T5, P4 , H2
T7,T8
600 mm in water
400 mm in water
20cm water column
Leveling tank
5.1 Stand-Still test
This test is focusing on the chance of Hydrogen to leak into the separator while the machine is in Stand Still situation. In other words, this test is targeting the machine when it is not working.
To perform this test, the container filled up by hydrogen up to 1.0-1.1%, the separator is off and under stable conditions (normal pressure and temperature) and the fan is working to provide hydrogen mixing. The changes are monitored for ≈20 hours.
Graph 2 in the following page is showing the general outcome of test.
Result:
No hydrogen leakage into the frame top part and gearbox has been detected.
testm036(26) INTAB Interface-Teknik AB
2,0
1,5
1,0
0,5
0,0
-0,5
-1,0
-1,5
-2,0
-2,5
-3,0
-3,5
-4,0 bar
1,30 1,20 1,10 1,00 0,90 0,80 0,70 0,60 0,50 0,40 0,30 0,20 0,10 0,00 -0,10
%
18:00 2011-04-27
00:00 2011-04-28
06:00 2011-04-28
2011-04-27 15:06:46 C1 2011-04-27 15:10:58 18,79 t 2011-04-28 09:58:50
AirUnderCapP, bar
GearBoxAir, bar Hyd.Fan, %
Hyd.TestPipe, %
Hyd.GearBox, %
5.2 Stand-By test
This test is focusing on the chance of Hydrogen to leak into the separator in Standby situation.
To perform this test, the separator should be heated up first and this has been done with running the machine without feed for around 40minutes. Then the container filled up by hydrogen up to 1.0-1.1%, the machine turned on one more time for another 30 minutes and turned off until the end of the test. The hydrogen leak during cooling down has been examined.
The test result is shown in Graph 3 in the following page. More detailed graph is attached in Appendix III.
Result:
No hydrogen leakage into the frame top part and gearbox has been detected.
testm036(41) INTAB Interface-Teknik AB
0,070 0,060 0,050 0,040 0,030 0,020 0,010 0,000 -0,010 -0,020 -0,030 -0,040 -0,050 -0,060 -0,070 -0,080 -0,090 -0,100 bar
1,00
0,90
0,80
0,70
0,60
0,50
0,40
0,30
0,20
0,10
0,00
%
100 95 90 85 80 75 70 65 60 55 50 45 40 35 30 25 20 15
°C
12:00 2011-05-06
00:00 2011-05-07
12:00 2011-05-07
00:00 2011-05-08
2011-05-06 11:45:00 C1 2011-05-06 11:58:38 1,90 d 2011-05-08 09:45:00
AirUnderCapP, bar
GearBoxAir, bar
Hyd.Fan = 1,00 %
Hyd.TestPipe = 0,02 % Hyd.GearBox = 0,01 %
OilGearBox, °C
airUndCapT, °C
Second run First run
5.3 Normal Operation Test
This test is focusing on the chance of Hydrogen to leak into the separator during its normal operation and the cooling period after shutting off.
To achieve this, the container filled up by hydrogen up to 1.0-1.1%, the separator has turned on and works up to 11 discharges, separator has turned off then and cooled down for 17 hours.
Graph 4 belongs to this test is in the following page and graphs with more details are attached in Appendix III.
Result:
No hydrogen leakage detected during the normal operation while the machine is working with feed.
The hydrogen leaks into the frame top part 2 hours after separator shutting. It increases from 0,02% to 0,12% in 10 hours and become stable in that concentration.
Gearbox part shows no increase in hydrogen concentration.
Shutting down
H2 starts to increase in Hood
H2 concentration is almost stable
5.4 Empty Leveling Tank Test
For this test, the leveling tank is empty from water. In other words, the outlet 514 (a), 462 (b) and sludge tank (c) are open to the atmosphere of the container which includes hydrogen.
The container filled up by hydrogen up to just 0.6% because the gas capsule finished. The separator has turned on and works up to 11 discharges, separator has turned off then and cools down for 17 hours.
Drawing 2 in the following page belongs to this test and Graph 5 & Graph 6 are showing the outcome. Detailed graphs are attached in Appendix III.
Result:
Operation: As it was expected, hydrogen starts to leak into both frame top part and Gearbox while the machine is working. The hydrogen content in the frame top part is showing a small decrease with each discharge and increase again. The hydrogen content of gearbox is continuously increasing. (Graph 5)
Cooling down: The amount of hydrogen in the frame top part becomes stable and constant 1 hour after shutting down the machine but the gearbox hydrogen shows a slightly decrease.
(Graph 6)
Drawing 2: Empty Leveling tank sketch
T: Thermocouple P: Pressure gauge H: Hydrogen sensor
RH: Relative Humidity sensor F: Flow meter
Container
a
b c
P5 T6 P6
T1 , P1 , F T2 , P2
T3,T4, P3 , RH , H1 T10 , P7, H4 T9 , H3
T5, P4 , H2
T7,T8
20cm water column
testm036(43) INTAB Interface-Teknik AB
0,60 0,55
0,50 0,45 0,40
0,35
0,30 0,25 0,20
0,15 0,10
0,05 0,00 -0,05
%
15,0 14,0 13,0 12,0 11,0 10,0 9,0 8,0 7,0 6,0 5,0 4,0 3,0 2,0 1,0 0,0 -1,0 -2,0 m3
13:00 14:00 15:00
2011-05-09 12:18:00 C1 2011-05-09 12:19:04 3,51 t 2011-05-09 15:50:00
Hyd.Fan
=0,62 % Hyd.Fan
=0,58 %
Hyd.TestPipe
=0,02 %
0,35%
0,39 %
0,41 %
0,43 %
0,44 % Flöde, m3
Hyd.GearBox
=0,01 %
0,21 %
0,35 %
0,41 %
0,45 % 0,47 % 0,47 %
0,48 %
Hydrogen increase in the test pipe Hydrogen increase in gearbox
Decrease of H2 with each discharge
Normal Operation
Motor Shutting down
Cooling down
6 Results and Discussion
According to the outcome of the tests, cooling down time is the only time that hydrogen will leak into the machine, with the outlet 514 and 462 and sludge tank connection immerged into the water. There is a reduction in pressures at the shut off time. As it is shown in Graph 13, frame top part pressure reduces from 0 mbar to -25 mbar after shutting off and goes up to zero again after ≈4 hours. This can be a key point to explain hydrogen leak. In addition, temperature change is also important. Expectedly, it goes up in all parts during operation time and gradually comes down after separator shut off (Graph 14). This can make extra vacuum inside the separator.
As it can be seen in Graph 15, hydrogen leak is started around two hours after shutting off and the separator are completely stopped at that time (the vibration sensors shows almost zero 50 minutes after shutting off). This means that shutting off and cooling down period cannot be dangerous itself when there is no source of sparkles. The main problem can be assumed as the next start up time, when frame top part air contains both oxygen and hydrogen and starting up the separator may make the third component which is sparkle. As a result, the frame top part should be cleaned before starting up or, from another point of view, ways should be taken to prevent hydrogen leak into the separator.
Needless to say, in the case that outlets 514 and 462 and sludge tank are open to the contaminated atmosphere, separator will filled with hydrogen during the operation and this situation should be avoided and if it happens, proper cleaning up is needed for all parts.
In section “8Recommendations”, possible solutions are being discussed and tested to overcome this problem. The solutions are classified from two points of view. The first viewpoint is focusing on leak-prevention in which ways are selected and tested to prevent hydrogen leak into the separator. The second view point is trying to examine possible ways of cleaning up after hydrogen leak.
6.1 Testing Procedure flowchart:
The following flowchart provides an overview on the steps should be taken in order to test a separator to be used in explosive environment.
Notes:
1. In case of different design (ex. difference in gearbox part), separator may have different standards as well.
The exact testing procedure should be discussed.
2. Follow the test procedure under section 4.2
3. Follow the test procedure under section 5.3. Standstill & Standby test is not necessary.
4. Inert gas injection method or a new method from Recommendations section based on further decisions by AL.
Test Object
Same design as separators tested before?1
Possible to apply the current test procedure?
New test procedure maybe needed Run “allowed
pressure loss”
test2 Find & fix the
leaks5
Meet the requirements?
Run full test in the lab3
Is the result acceptable?
Decide about leak prevention method4
Ready to use Leaks were
founded and fixed before?
This separator may not be suitable for explosive
environment
This separator may not be suitable for explosive
environment Yes No
Yes
No
No
No No
Yes Yes
Yes
testm036(40) INTAB Interface-Teknik AB
0,20
0,18
0,16
0,14
0,12
0,10
0,08
0,06
0,04
0,02
0,00
%
16:00 2011-05-05
18:00 2011-05-05
20:00 2011-05-05
22:00 2011-05-05
00:00 2011-05-06
02:00 2011-05-06 2011-05-05 15:59:00 C12011-05-05 16:02:01 10,01 t 2011-05-06 02:03:00
0,04 % 0,05 %
0,06 % 0,07 %
0,08 % 0,09 %
0,1 % 0,11 %
0,12 %
0,02 %
7 Methane case study
The tests in this work are all done by the use of hydrogen with 1.0-1.1% concentration as the simulating gas. It is interesting to have a rough calculation based on real case. Methane seems to be a good example. It is classified as an explosive gas with 16.04 g/mol molecular weight, LEL=5% and UEL=15% in volume. [17][18]
The following graph is a focus on the cooling down stage of Graph 4 and it clearly indicates how hydrogen concentration is increasing in the frame top part.
Bellow is the data of hydrogen extracted from the graph:
Real time H2 concentration minutes Real time H2 concentration minutes
19:05 0.07
35mins
17:15 0.03 19: 40 0.08
20 mins 35 mins
17:35 0.04 20:15 0.09
25 mins 45 mins
18:00 0.05 21:00 0.10
30 mins 110 mins
18:30 0.06 22:50 0.11
35 mins 160mins
19:05 0.07 01:30 0.12
Total ≈ 8 hours
The concentration of 80% of methane in the surrounding has been selected for calculations.
The hydrogen concentration was 1.0% during the test which means that the leaked concentrations should be multiplied by 80.
CH4 concentration minutes CH4 concentration minutes
5.6
35mins
2.4 6.4
20 mins 35 mins
3.2 7.2
25 mins 45 mins
4.0 8
30 mins 110 mins
4.8 8.8
35 mins 160mins
5.6 9.6
Total ≈ 8 hours
7.1 Analyze
The question is “what does these numbers really mean?”
- According to “Normal Operation Test” section, leaking starts almost 2 hours after the shutting down time and it continuous more than 8 hours and it will stop then. During these 8 hours the presence of explosive gas in the surrounding is not favorable.
- Besides, the leaking process is not linear and it has an exponential shape, as it is clear from the graph. It is faster at the beginning and the rate is reducing until it becomes zero and concentration reaches a constant value.
- The concentration, the exact time that explosive gas is present in the surrounding and the length of this presence are 3 key points to decide if the situation is dangerous.
Suggested Methane concentration during the cooling down stage:
(explosive between 5-15%)
first 2 hrs: no leaks 6th hour: 8%
3rd hour: 4% 7th hour: 8%
4th hour: 5.6% 8th & 9th hours: 8.8%
5th hour: 7.2 % 10th hour & so on: 9.6%
Start
up Operation Shut
down
First 2hours
3rd, 4th, 5th & 6th hrs
7th, 8th, 9th &
10thhrs
Cooling down continued Cooling down
Safe steps, no leak detected
Leaking starts with
high rate
Leaking continued but with low
Safe, leaking
stops
Remember:
Hydrogen is the worth in case of leakage while its permeability is the highest.
8 Recommendations
As it mentioned in the “Results and Discussion” section, hydrogen is leaking into the machine during cooling down period and the problem is the next starting up. There is two general ways out of this problem.
1. Prevent hydrogen leak during the cooling down period.
− Proper ventilation to the fresh air 2. Clean up the Frame top part before starting up
− Water Discharges
− Injection of pressurized air before start up
− Proper ventilation to the fresh air
According to this theory, while the separator has ventilation to the fresh air, amount of pressure drop at the shut down time will be less and the separator will compensate this pressure loss with sucking in fresh air through that ventilation and hydrogen will not leak in anymore. To test the feasibility of it, tests are being done in 3 different modes:
1) Outlet 514, 462 and sludge tank are all open to the fresh air Drawing 3& Graph 7
2) Outlet 514 and 462 are open to the fresh air while sludge tank mmmmm connection is immerged in the water in the leveling tank
Drawing 4 & Graph 8
3) Sludge tank has ventilation to fresh air and outlet 514 and 462 are mmmmm immerge in the water in the leveling tank
Drawing 5 & Graph 9
Detailed graphs for each test are added in Appendix III, Graph 17 to Graph 19.
For these three tests, the container filled up by hydrogen up to 1.0-1.1%, the separator has turned on and works up to 10-12 discharges, separator has turned off then and cools down for several hours.
Result:
No hydrogen leakage detected for these tests.
Drawing 3: Open outlet 514, 642 & sludge tank
T: Thermocouple P: Pressure gauge H: Hydrogen sensor
RH: Relative Humidity sensor F: Flow meter
Container
P5 T6 P6
T1 , P1 , F T2 , P2
T3,T4, P3 , RH , H1 T10 , P7, H4 T9 , H3
T5, P4 , H2
T7,T8
20cm water column
Motor Shutting down
Cooling down
Drawing 4: Open outlet 514 & 462 sketch Leveling tank
T: Thermocouple P: Pressure gauge H: Hydrogen sensor
RH: Relative Humidity sensor F: Flow meter
Container
P5 T6 P6
T1 , P1 , F T2 , P2
T3,T4, P3 , RH , H1 T10 , P7, H4 T9 , H3
T5, P4 , H2
T7,T8
400 mm in water
20cm water column
Motor Shutting down
Cooling down
Drawing 5: Open sludge tank sketch
T: Thermocouple P: Pressure gauge H: Hydrogen sensor
RH: Relative Humidity sensor F: Flow meter
Container
P5 T6 P6
T1 , P1 , F T2 , P2
T3,T4, P3 , RH , H1 T10 , P7, H4 T9 , H3
T5, P4 , H2
T7,T8
600 mm in water
20cm water column
Leveling tank
Motor Shutting down
Cooling down
− Water Discharges
According to Graph 5, it can be possible to dispose off explosive gas with water discharges because each discharge is shooting out a part of gas in the frame top part. A test has carried out to examine the feasibility of this theory.
To perform this test, there should explosive gas in the top part first and it is known that the machine is sucking in hydrogen after shutting down, during its cooling period. So, the container filled up by hydrogen up to 1.0-1.1%, the separator has turned on and works up to 10-12 discharges, separator has turned off then and cools down for several hours to suck in some hydrogen.
After having hydrogen in the top part, the machine has turned on and goes through its normal operation mode with water as the feed. The number of discharges depends on the concentration of the explosive gas and it is not pre-decided for this test.
Data on this test is available through Drawing 6 & Graph 10 and the detailed graphs are attached in Appendix III, Graph 20, Graph 21 & Graph 22.
Result:
The machine sucks in around 0.18% hydrogen during the cooling down stage. The hydrogen flushed out from the frame top part with 7 discharges and the frame top part becomes clean after that. Consequently, it seems to be feasible to clean up the frame top part with water discharges.
The problem with this method is that the machine is under explosion danger for a while at the beginning because it takes a few numbers of discharges until the concentration of explosive gas comes down. In other words, we have to run the separator while there is an explosive environment (mixture of explosive gas and oxygen) in the frame top part.
Therefore, there will be a chance on presence of the third component, sparkle, with turning on the separator.
− Injection of Pressurized air before start up
According to the results, the problematic stage is the starting up. Pressurized air injection is a way to clean up both the frame top part and gearbox from explosive gas and needless to say, there is no need for continuous injection during the test.
Briefly, it is now clear that the explosive environment is leaking into the separator. The leak happens during the cooling down step and the separator is not working then. The ways to overcome the problem is listed below:
- The current injection of gas can be kept. The nitrogen will be replaced by pressurized air. Injection is necessary before each start up or during the cooling down step.
- If there is the possibility of having proper ventilation to the fresh air, there is no need to have the injection system.
Drawing 6: Water discharges sketch
T: Thermocouple P: Pressure gauge H: Hydrogen sensor
RH: Relative Humidity sensor F: Flow meter
Container
P5 T6 P6
T1 , P1 , F T2 , P2
T3,T4, P3 , RH , H1 T10 , P7, H4 T9 , H3
T5, P4 , H2
T7,T8
600 mm in water
400 mm in water
20cm water column
Leveling tank
testm036(47) INTAB Interface-Teknik AB
1,20
1,10
1,00
0,90
0,80
0,70
0,60
0,50
0,40
0,30
0,20
0,10
0,00
%
15,0 14,0 13,0 12,0 11,0 10,0 9,0 8,0 7,0 6,0 5,0 4,0 3,0 2,0 1,0 0,0 -1,0 -2,0 m3
13:00 14:00 15:00
2011-05-12 12:32:00 C1 2011-05-12 12:32:53 2,81 t 2011-05-12 15:22:00
Hyd.Fan = 0,98 %
Hyd.TestPipe
=0,13 %
0,18 %
0,14 %
0,1 %
0,07 %
0,05 %
0,03 %
0,02 % Hyd.GearBox = 0,01 %
Flöde, m3
H2 reduction with each discharge
Discharge
9 Further work
− Testing other machines:
As it is clear from “Testing Flow Chart”, it is necessary to perform tests on other machines as well. In order to make the testing process more convenient, a test box with hydrogen sensor, Pressure gauge and thermocouple can be designed and it is possible to connect it to the frame top part through regular plastic hoses because the frame top part is not experiencing very high pressure.
− Flammable process liquid:
This project is specifically focused on the case of “explosive environment around the separator” but it is applicable to partially extend Outcomes from “6Results and Discussion”
and “8Recommendations” sections to flammable process liquid case too, which HFO is the best example.
From this view point, air is the environment around the centrifuge. Eliminating oxygen from the frame top part is very important while it is one of the three components needed for an explosion.
Obviously, pressurized air injection is not applicable in case the process liquid is flammable. Nitrogen injection (inert gas injection) may still be needed but just before the start up time. With this suggestion, the inert gas injection system is still necessary but the amount of inert gas being used will be less while there it seems there is no need to supply nitrogen continuously during the normal operation and at the shutting down mode. This can be assumed as an early and simple stage to make the injection process more economical.
To make the inert gas injection even cheaper, the possibility of replacing nitrogen with steam can be investigated. Although, the available test results are a very good guide line for managing the HFO tests , it still needs practical tests before deciding about any change in the current system (inert gas injection).
− Oxygen Sensor:
For further tests on HFO, oxygen can be the targeted element to measure and control. A proper Oxygen sensor along with Humidity sensor can be extremely useful for monitoring and interpreting the results and comparing them with the suggested values from theoretical calculations.
10 References
[1] Flammability of gas mixtures, Part 1: Fire potential. Schroder, Volkmar and Molnarne, Maria. s.l. : Journal of Hazardous Materials , 2005, Vols. A121, 37–44.
[2] Carbon dioxide dilution effect on flammability limits for hydrocarbons. Chena, Chan- Cheng, Liawa, Horng-Jang and Wangb, Tzu-Chi. s.l. : Journal of Hazardous Materials, 2009, Vols. 163, 795–803.
[3] Approximation of flammability region for natural gas–air–diluent mixture. Liao, S.Y, Jiang, D.M and Huanga, Z.H. s.l. : Journal of Hazardous Materials, 2005, Vols. A125, 23–28.
[4] Nitrogen and carbon dioxide dilution effect on upper flammability limits for organic compound containing carbon, hydrogen and oxygen atoms. Wang, TzuChi, Chen, ChanCheng and Chen, HuiChu. s.l. : Journal of the Taiwan Institute of Chemical Engineers, 2010, Vols. 41, 453–464.
[5] Sensors. [Online] January 2005. http://www.sensorsmag.com/sensors/chemical-gas/leak- testing-with-hydrogen-624.
[6] MediVac Technologies. [Online] http://www.medivactech.com/heliumleaktestinginfo.htm.
[7] Morris, Daive. Testing Expo. Tracer Gas Leak Detection for the Automotive Industry.
[Online] An Alcatel-Lucent Company. http://www.testing-
expo.com/usa/08conf/pdfs/day_3/15_Alcatel_David%20Morris.pdf.
[8] Block, Matthias. NDT. Hydrogen as Tracer Gas for Leak Testing. [Online] Sensitor Technologies, Muehlheim am Main. http://www.ndt.net/article/ecndt2006/doc/Tu.2.6.1.pdf.
[9] Applied science. M. E. Reinders, ft. Schutten, J. Kistemaker. 1951, Vol. 1.
[10] US Fuel Cell Council. Hydrogen Safety. [Online]
http://www.hydrogenassociation.org/general/factSheet_safety.pdf.
[11] Air Liquid. Safety Data Sheet Library, Helium. [Online] http://www.msds- al.co.uk/search/doSearch/aGVsaXVt/1/.
[12] Air Liquid. Safety Data Sheet Library, Hydrogen. [Online] http://www.msds- al.co.uk/search/doSearch/aHlkcm9nZW4~/1/.
[13] Schoonovernic. Hydrogen Leak Tetection. [Online]
http://www.schoonoverinc.com/products/Leak%20Detection/Sensistor.htm.
[14] Schroeder, V. Holtappels, K. International conference on Hydrogen Safety 2007. Explosion Characteristics of Hydrogen-Air and Hydrogen-Oxygen Mixtures at Elevated Pressures.
[Online] Federal Institute for Materials Research and Testing (BAM), Berlin, Germany.
http://conference.ing.unipi.it/ichs2005/Papers/120001.pdf.
[15] Air Products. Hydrogen Safety Data Sheet. [Online] www.airproducts.com.
[16] AGA. [Online] 05 2011.
http://www.aga.fi/international/web/lg/aga/like35agacom.nsf/docbyalias/shielding_gases.
[17] Bjerketvedt, Dag, Bakke, Jan Roar and van Wingerden, Kees. Gas Explosion Handbook.
Gexcon. [Online] http://www.gexcon.com/handbook/GEXHBcontents.htm.
[18] Education & Training. Comenius - European Cooperation on School Education . [Online]
2004. http://cartwright.chem.ox.ac.uk/hsci/chemicals/methane.html.
[19] (BAM), Federal Institute of Materials Research and Testing. Public Deliverables.
SAFEKINEX. [Online] 2002.
http://www.morechemistry.com/SAFEKINEX/deliverables.html.
Appendix I
Appendix I: Hydrogen injection and Air properties
Injection of 10% H2+90%N2 mixture will slightly change the properties of the air by adding extra N2 to it. The following calculations are showing the extent of this effect.
Container volume: 2.6 2.6 3 20 20000
Air components7:
N2: 78%
O2: 21%
Others: 1.0%
Before injection:
N in container 15600 lit
O Others in container 4400 lit
During the injection:
1% 200
200 9 1800
After injection:
15600 1800 17400
4400 17400 21800 4400
% 17400
21800 100 79.8%
which means less than 2% increase in Nitrogen content of air.
The pressure of the container is not changing with this injection and remains constant. The temperature is almost constant as well in case the separator is not working (ex. Stand still test). However, it is gradually rising when the separator is on because the motor is producing heat.
7
Appendix II
Appendix II: Sensors
Hydrogen sensor
Hydrogen measurement is a vital point for this test. In the first place, HLS-440 sensor from AppliedSensor Company (www.appliedsensor), which is shown in the following picture, has been selected regarding to its good accuracy (±3000ppm), very fast response time (<2 seconds) and excellent operational temperature range (-40-110°C). The price was fairly reasonable in compare with the characteristics of the sensor. (Prices are available in
“Overview on several Hydrogen sensors” section)
Yet, difficulties showed off belonging to the sensor output signal, CANbus, and its connection to the software while this output seemed to be new at Alfa Laval. Concerning the shortage of time, second selection has been made.
The CC 28 Hydrogen Transmitter from GfG Instrumentation Company has been selected (www.gfg-inc.com).
This sensor has the ability of measuring 0-100% LEL (0-4% by volume) hydrogen with the accuracy of ±0.01% by volume and the response time <90 seconds. The sensor output is 4-20 mA and the expected life time is 2-3 years. Detailed information about this sensor is available in the following page.
