Available on CMS information serverCMS IN 2000/040
CMS Internal Note
The content of this note is intended for CMS internal use and distribution only
August 21, 2000
Experimental Tests with a Standard Closed Loop Gas Circulation System
A Prototype Study
LHC Gas Working Group:
L. Besset, F. Hahn, S. Haider, C. Zinoni
Abstract
This document summarises tests made on a complete prototype of a Closed Loop Gas Circulation System that was built according to the standard layouts as proposed for the CMS Muon gas systems in the Technical Design Report.
The design is based on a pressure regulated closed loop circuit as used also on some of the LEP gas systems. One of the new features at the LHC is that there are several sub-distribution systems in parallel and in some cases, hydrostatic pressure differences due to the height of the detectors have to be taken into consideration.
This report provides also a valid record of practical experience on a number of crucial gas system components that can be used for the LHC.
Introduction......
Experimental Set Up......
Circulation Loop......
Description of Specific Components
Pipe-Runs......
Experimental Measurements and Results......
Closed Loop Operation:......
Pressures and Flows for Freon134a and N2 Operation......
Start up and Partial System Stop......
Compressor Unit......
Safety-Back-Up System......
Filling and Purging......
Pressure Regulation......
Effective Gas Replacement......
Component tests......
Control Valves......
Pressure Regulators......
Conclusions......
Operational aspects......
Component Tests......
Acknowledgement......
Introduction
Because of the large detector sizes at the LHC most of the experimental gas systems are closed loop circulation circuits that split into multiple sub-distribution systems near the chambers. Usually large-area ionisation detectors cannot sustain large mechanical forces and require a pressure regulation with a stability of approximately 1mbar above atmospheric pressure. To achieve adequate pressure accuracy at the LHC detectors, one has to take into account hydrostatic pressures that are generated by gases heavier than air depending upon the position of the chamber inside the experiment.
Table 1 Basic Prototype Parameters
Total Volume / 1m3Chamber size / 0.25 m3
Number of fake Chambers / 4
Operating Chamber Pressure / 0.5 - 1 mbar
Max. Hydrostatic p / 4 mbar
Operating Gas / Freon134a
Purge Gas / N2
Figure 1 Vertical cross section of the prototype installation. The gas racks with the pressure regulation systems are situated at floor level, the bottom and top chambers are at 3.3 and 12.5 m. On the containers themselves the gas inputs are near the bottom, the outputs are at the top.
This document describes tests made on a full closed loop gas circulation prototype (Table 1) that was built according to the design proposed to all four LHC experiments. The distribution sub-systems consisted of three gas racks: two for the top and bottom sub-distribution systems and one for the pump unit (Figure 1). Four barrels, each of 0.25m3, were used as mock-up chambers. To generate noticeable hydrostatic pressures the barrels are placed at 3.3 and 12.5 m height.
Table 2 Gas parameters of the operating and purge gases.
Gas / Density atT=20oCkg/m3] / Viscosity
[poise] / Relative Hydrostatic p
[mbar/m]
Freon134a / 4.45 / 1.20 * 10-4 / 0.3
N2 / 1.09 / 1.78 * 10-4 / -0.01
The motivation for this prototype circulation system is to get experience and information on the following aspects of the future gas systems:
Validate the proposed pressure regulation system and test the operation with two sub-systems situated at significant height differences. Special emphasis is given to the output control valves, the stability and reliability of the regulation system at various flow rates.
Verify filling and purging procedures for chambers sitting at different heights above the gas rack level and using an operating gas of high density (Freon134a[1]) and a purge gas (N2) of a density comparable to air (Table 2).
Check operational stability and the mutual influences of two sub-distribution modules if one sub-system is running and the other one is not.
Quantify the dynamic range (minimal, nominal and maximal flows) at which a typical system can run once the component sizes are fixed.
Test a maximum number of crucial system components in real conditions; e.g. control valves, pressure regulators, pressure sensors, etc.
Figure 2 Component layout of the gas circulation system
Experimental Set Up
Circulation Loop
The prototype consists of a full closed loop gas circulation system (Figure 2) comprising two distribution sub-systems equipped with fake chambers placed at large height differences (
Figure 1); a membrane pump recompresses the output gas for recycling to the inlet.
Each distribution sub-system is connected to two barrel-shaped containers that simulate the chambers. The pressure is regulated before and after the mock-up chambers allowing an accurate control of the chamber pressure. The flow rates can be adjusted by choosing appropriate P between input and output. The input pressure is controlled by a mechanical pressure regulator (2); a pneumatic control valve (10 and 12) driven by a PID with an input signal from a pressure transmitter regulates the chamber pressure at the outlet.
Whenever the gas circulation is stopped, the distribution sub-systems are isolated by electro-valves, and a mechanical safety system at the outlet ensures that the chambers can follow changing atmospheric pressure. In this prototype two different safety systems are tested and compared: in both cases the pressure relief is done by a bubbler (P=1 to 5 mbar), the filling is provided either at constant flow rate by a rotameter (15) or by a pressure regulator (Zero-Governor (3)) that injects gas only if the atmospheric pressure is rising. The second system is in particular useful for chambers with large volume it allows the atmospheric pressure to be followed, but it adds only the minimal amount of gas needed.
A control valve (9) in parallel to the pump regulates the pump inlet pressure so low that both sub-systems can readily return their gas. A pressure regulator (5) placed after the gas inlet ensures a constant supply pressure to the sub-distribution units and helps to stabilise the flows in the sub-systems. To calm down pressure fluctuation from the membrane pump, two small buffer volumes are added on each side of the pump, mechanical filters absorb similar fluctuations before each pressure transmitter used for the regulation.
For closed loop running a moderate fresh gas flow is injected by a mass flow controller (17), an equivalent amount of gas is exhausted controlled by a back pressure regulator (6) sitting behind the compressor. Filling and purging of the prototype is done in single pass mode at much higher rates using the rotameter (16) at the input, and a manual three-way valve at the end of the loop.
Table 3 List of components used in the prototype setup.
No / Component / Location / Trademark / Reference / Characteristics / Comments1 / Pump / KNF / N 145 / Membrane pump
2a / Pressure regulator / Sub-distribution system inlet / Jeavons / J125-S / 1 to 10 mbar / Used if press. contr. input
2b / Pressure regulator / Sub-distribution system inlet / Dungs / FRS 503 / 2.5 to 9 mbar / Used for tests
2c / Rotameter / Sub-distribution system inlet / Voegtlin / 0-100 l/h N2 / Used if const. input flow
3 / Pressure regulator / Safety system / Jeavons
Zero-Governer / -1 to +1 mbar
4 / Pressure regulator / Safety system / Dungs / FRS 207 / 2.5 to 9 mbar
5 / Pressure regulator / Upstream from both sub-systems / Rombach / 143-31 HP / 70 to 280 mbar
6a / Back pressure regulator / Exhaust / Conoflow / GH 30 / 0 to 1050 mbar / Used in normal operation
6b / Back pressure regulator / Exhaust / Bachofen / LPS / 15 GP 200 / P1= 200 mbar
Seat : Kalrez
Membrane:PTFE / Used for tests
7 / Non return valve / Nupro / CH 16 / Opening pressure : 70 mbar / steel 316
8 / Filter / Before press. transmitters / Sagana / 4F-FL-02-SS / Porosity : 2 micro / Sintered steel 316 L
9 / Regulating valve / Compressor bypass / Kämmer / 3114 / Kvs= 1.2 ;
“equal-%”
10 / Regulating valve / Output bottom distribution / ARCA-
von Rohr / Kvs= 0.25 ; linear
12 / Regulating valve / Output top distribution / Kämmer / 37047 / Kvs=0.074;
“equal-%”
Profibus / replacing
SAMSON 3020)
12 / Regulating valve / Output top distribution
(later replaced) / Samson / 3020 / Kvs=0.1 ;
“equal-%”
13 / Ultra-Sonic flowmeter / Spirocell / Plastic
14 / Va flowmeter / Sub-distribution system inlet / Krohne / H 250 / N2 : 0-700 Nl/h
Freon 0- 320 l/h
15 / Rotameter / Safety system / Voegtlin / V 100-80 / 0 to 1200 Nl/h / Glass
16 / Rotameter / Fast fill inlet / Voegtlin / V 100-80 / 0 to 10 Nl/h / Glass
17 / Mass flow controller / Fresh gas inlet / Bronkhorst Hi-Tec / F-201C-FB-33-P / Freon flow
0 to 100 Nl/h
18 / Mass flow meter / Exhaust / Bronkhorst Hi-Tec. / F-201C-FB-33-P / Freon flow
0 to 100 Nl/h
19 / Pressure transmitter / Pump outlet / Haenni / ED 510/314.461/035 / 0 to 1 bar / Output voltage
20 / Pressure transmitter / Upstream from both sub-systems / Jumo / 4 AP-30-242 / 0 to 1 bar / Output current
21 / Pressure transmitter / Sub-distribution system inlet / Jumo / 404327/412-412-504-28-61 / 0 to 50 mbar / Output voltage
22 / Pressure transmitter / Regulating valve inlet / Sensortec / BTEL5P10D1A / -15 to +15 mbar / Output voltage
23 / Pressure transmitter / Regulating valve inlet / Effa / GA Ex 63 C 20 P S / -20 to +20 mbar / Output current
24 / Pressure transmitter / Chamber / Effa / GA Ex 63 B 20 P S / 0 to 20 mbar / Output current
25 / PID Controller / Honeywell / UDC 3300
Description of Specific Components
For the prototype tests, in total four different control valves (items 9,10, and 12 in Figure 2and Table 3) were used. The rate of flow through a valve is defined by its Kv parameter. If P< ½P1 the Kv value is calculated by:
T / : / absolute temperature at valve inlet; / : / gas density at STP
Q / : / gas flow at STP;
P / : / pressure drop : P1-P2.
The Kvs parameter is the Kv value that corresponds to a valve opening of 100%. The relationship between the flow rate through the valve and the valve travel (as the valve opens from 0 to 100%) is called the inherent flow characteristic.
In this set-up, the Arca valve had a linear flow characteristic; the three other valves were of type “equal- percentage” [2]. Depending on the installation points they had to be of different Kvs values which allows only a partial comparison of performance parameters. However, a competitive evaluation is of interest because two of the valves tested were using very modern so-called “intelligent positioners” while the other two had mechanical positioners.
Moreover a couple of low range pressure regulators (items: 2 to 6 in Figure 2 and Table 3) were tested and compared in this prototype study.
To understand the operation of the prototype, several sensors have been installed to monitor flows and pressures at “strategic points”.
- Pressure Transmitters: All transmitters used in the prototype (items 19-24 in Figure 2 and Table 3) read relative pressure; their precision is typically 1 % of full scale[*]:
- JUMO (type 404327, range 0 to 50mbar): one electrode of a plate capacitor is used as membrane that is displaced by pressure, so the plate distance changes as a function of pressure, the changing capacity is used to measure the pressure.
- Sensor Technics (type BTEL5P10D4A, range –10 to + 10 mbar): Pièzo-resistive measurement, a thin layer of silica-gel transmits the pressure onto a resistor made of doped silicon that acts as a strain gauge, deforming the silicon leads to a change in resistance.
- Flowmeters:
- GILL: Three specially made ultrasonic time-of-flight flowmeters (13) are used, allowing to measure gas flow with good accuracy (±5 % of measured value) independently from the gas composition.
- Krohne: (type: H250/RR/M9/ESK, range: 32 to 320 l/h Freon134a) A variable area type flowmeter with magnetically coupled remote signal is used for the bottom sub-distribution system.
- Hydropneu: (type DMTF range 37 to 375 Nl/h of freon134a) also variable area type meter with remote signal that was used for some tests on the top sub-distribution system. (This meter was taken off later, since it was introducing oscillations into the flow and the reading was unusable.)
Pipe-Runs
The prototype set-up has gas pipe length built-in to reproduce as near as the future conditions at the LHC gas installations (Table 4). The supply and return lines from the sub-distribution systems has a lengths about 75m, the lines to and from the chambers span a distance of 40 and 51 m. For the prototype all gas pipes were made in copper.
Table 4 Pipe parameters and calculated pressure drops for straight pipes.
At Max. FlowSection / Length
[m] / Internal Pipe Diam.
[mm] / Nominal
Flow
[l/h] / Max. Flow
[l/h] / Gas
velocity
[m/s] / Reynolds
No. / P/
pipe
[mbar]
Circulation Loop
Common Input Line / 1 / 75 / 12 / 120 / 360 / 0.88 / 3714 / 4.2Top Sub-System
Inside Distribution Rack / 2 (top) / 3 / 10 / 30 / 90 / 0.32 / 1114 / 0.0Input Coil / 3 (top) / 25 / 10 / 15 / 45 / 0.16 / 557 / 0.2
Vertical tube to chamber / 4 (top) / 23 / 8 / 15 / 45 / 0.25 / 696 / 0.3
Vertical tube from chamber / 5 (top) / 23 / 10 / 15 / 45 / 0.16 / 557 / 0.1
Output Coil / 6 (top) / 25 / 10 / 15 / 45 / 0.16 / 557 / 0.2
Inside Distribution Rack / 7 (top) / 2 / 8 / 30 / 90 / 0.50 / 1393 / 0.1
Total
/ 0.9Bottom Sub-System
Inside Distribution Rack / 2 (bottom) / 3 / 10 / 90 / 270 / 0.95 / 3342 / 0.2Input Coil / 3 (bottom) / 25 / 10 / 45 / 135 / 0.48 / 1671 / 0.6
Vertical tube to chamber / 4 (bottom) / 12 / 8 / 45 / 135 / 0.75 / 2089 / 0.8
Vertical tube from chamber / 5 (bottom) / 12 / 10 / 45 / 135 / 0.48 / 1671 / 0.3
Output Coil / 6 (bottom) / 25 / 10 / 45 / 135 / 0.48 / 1671 / 0.6
Inside Distribution Rack / 7 (bottom) / 2 / 8 / 90 / 270 / 1.49 / 4178 / 0.5
Total
/3.0
Circulation Loop
Common Input Line / 8 / 78 / 16 / 120 / 360 / 0.50 / 2785 / 1.1Experimental Measurements and Results
Closed Loop Operation:
The closed loop circulation system consists of three gas racks: one rack for the compressor unit and two independent sub-distribution systems. Each distribution module regulates the gas flow and pressures for two mock-up chambers, which are sitting at the same height (see Figure 1).
Pressures and Flows for Freon134a and N2 Operation
Table 5 Nominal freon134a gas flows.
Fresh Gas Flow / 6 to 12 l/h(5 to 10% of total Flow)
Nominal
Circulation
Flow / Top / 30 l/h
Bottom / 90 l/h
Total / 120 l/h
Volume Replacement at Nominal Flow / Top / 17 h
Bottom / 6 h
The design flows of the circulation system are given in (Table 5). To make the situation more rigorous the sub-system flows are unequal, the rates are equivalent to one volume replacement after ~17h for the top chambers and 6h for the bottom ones. The P’s for the purge gas N2 are very different due to changed gas properties. Although it is not really necessary, we have defined a nominal N2 flow that is nearly a factor two higher than the rates with freon134A. As can be seen in Table 6, it is not a problem to reduce the purge flow rates if required. In this document we refer frequently to nominal flows as defined in Table 6 for freon134a and for N2, minimal and maximal rates are multiplied by 1/3 and 3 respectively.
Table 6 Measured pressures for N2 and Freon134a at nominal, minimal and maximal gas flows.
N2 / Freon134 aBottom / Top / Bottom / Top
Nominal flow / 150 / 60 / 90 / 30
Pressure distribution / In / Cham. / Out / In / Cham / Out / In / Cham / Out / In / Cham / Out
1.9 / 0.45 / -0.3 / 1.35 / 0.9 / 0.1 / 2.1 / 0.2 / 0.7 / 5.6 / 0.9 / 5.3
Minimum flow / 50 / 20 / 30 / 10
Pressure distribution / In / Cham. / Out / In / Cham / Out / In / Cham / Out / In / Cham / Out
1 / 0.8 / 0.1 / 1.0 / 0.9 / 0.8 / 1.6 / 0.32 / 1 / 5.3 / 1.4 / 5.3
Maximum flow / 450 / 180 / 270 / 90
Pressure distribution / In / Cham. / Out / In / Cham / Out / In / Cham / Out / In / Cham / Out
5.6 / 0.5 / -2 / 2.6 / 0.7 / 0.0 / 5.4 / 0.3 / -0.5 / 6 / 0.6 / 5
The actual pressure drops across the chambers at nominal flows were measured to be smaller then 0.5 mbar for the top system and smaller then 1.5 mbar for the bottom system. Although it is comfortable to have a small p in the outlet pipe, it was felt not very helpful if the inlet pressure was very low; a larger P span allows a better adjustment of the flow. The minimal flow rate is in reality limited by the lowest pressure one can adjust at the inlet. Note that a hydrostatic P cancels if the pressures are measured at the same height.
Table 7 summarises recommended P’s for the pipes and valves in the distribution sector, adding a reference pressure and a possible hydrostatic P enables to calculate the supply and return pressures required for the distribution modules.
Table 7 Recommended pressures drops for distribution pipes and valves. The outlet control valve has to compensate hydrostatic P between sub-systems before they join into the common return.
P [mbar]At Nominal Flow / P [mbar]
At Maximal Flow
Inlet Pipe / 1 to 2 mbar / 2 to 4 mbar
Chamber Inlet (i.e. orifice) / 1 to 2 mbar / 1.5 to 2.5 mbar
Outlet Pipe / 0.5 mbar / < 2 mbar
Output Control Valve / 2 to 3mbar + PHydrostatic / 3 to 5mbar + PHydrostatic
An additional complication occurs due to hydrostatic pressure effects i.e., the return pressure from the top chambers is 4 mbar higher then the return pressure from the bottom chambers; before joining the two lines, this must be compensated by the control valve at the outlet of each distribution module (see Figure 3 and Figure 4).
The circulation flow is enriched with about 5 to 10 % of fresh gas. The fresh gas rate is monitored permanently and compared with the output gas that is exhausted to vent. The average difference is a direct and sensitive measure of the total leak rate of the circulation loop. An example of this can be seen in Figure 5: until about 15:21h the fresh gas rate is about 15 l/h and the average return is about 13 l/h, the difference of 2 l/h is due to small loss that occurred that day in the bottom sub-system. At 16:00h the bottom system was stopped and the flow difference vanished.
Figure 3 Pressure evolution at the various points of the Top distribution unit for nominal and maximal freon134a flow. The pressure at the inlet and outlet sensors (situated in the gas rack) is increased by 4 mbar, due to the weight of the heavy freon134a gas.
Figure 4 Pressure evolution at the various points of the bottom distribution unit for nominal and maximal freon134a flow. The hydrostatic p between chamber and regulators is only 1 mbar, the dynamic P between input and outlet increases from 1.3 mbar at nominal flow to 6 mbar at maximal flow. The chamber pressure is kept in the desired range by choosing an appropriate output pressure.
Figure 5 Comparison of fresh input gas flow and output flow. The difference is a sensitive measure of the gas loss in the circuit. The plot shows until 15:50 a loss of ~2l/h independent of the input flow; after the bottom system was switched off the loss returns to zero.