|
|
 |
Application notes |
 |
Sampling line size (AN-01)
It is a common practice in air separation plants to use ¼" O.D. line for sampling system. In older installation it is also common to have sample pressure regulator close to the analyzer sample inlet. This can lead to long lag time limiting the speed of response of an analytical system. For example, sample line of 100 feet of ¼" O.D. with .190" I.D. (typical copper line) has an internal volume of .02 cubic feet. If the sample flow is 1 SCFH (475 sccm) and we assume the line to be at atmospheric pressure, it will take 1 minute and 12 sec. to travel down this line. If the same line is made of 1/8" O.D. with .085" I.D., the internal volume becomes equal to .004 cubic feet. This is 5 times less volume. It will take 24 seconds to travel down the line with 1 SCFH. We have assumed a line pressure equals to the atmospheric one. If the line is pressurized at 1 atmosphere there will be twice the volume of gas into the line so, twice the time will be required for a sample to go through this line.
One may thought to increase the sample flow through the ¼" O.D.line to overcome this problem. If a bypass flow is set to 10 SCFH the time will be decrease by a factor of 10. But after one year of operation this result in 87600 cubic feet of gas thrown away. If this gas is pure argon, this gives around 97 cubic feet of liquid. This is also equivalent to 350 gas cylinders (250 cubic feet size cylinder). The older type of analyzer system for trace nitrogen measurement uses 2 to 4 SCFH of sample flow (for silent electric discharge type) or 1.48 cubic feet (700 sccm) for ion mobility type. Contrôle Analytique analyzer works with a default sample flow of 75 sccm. The flow can be set as low an 25 sccm if required.
In conclusion, we are recommending the use of 1/8" O.D. stainless steel line for sampling lines. Furthermore, 1/8" O.D. line are available in coil of 500 or 1000 feet for long run. There is no need for fittings or welding. Also, installation cost is minimum since 1/8" O.D. line are easily installed. A sample line of 100 feet made of 1/8" O.D. and .085" I.D. connected to a nitrogen source at 1 PSIG and vented to the atmospheric pressure will have a flow of 475 sccm (1 SCFH). Well enough to supply sample gas to a K4000NG. It will require 10 psig for 1000 feet long sample line.
The sample pressure regulator must be installed as close as possible from the sample connection point. The pressure will be adjusted to the minimum value required to have the proper flow into the analyzer. Such sampling system will have a faster response time, better leak integrity, less operation cost.
The Importance of Regular Purging (AN-02)
Here are some quick calculations to help you understand why it is so important to have some techniques to evacuate the air from pressure regulators when replacing calibration cylinders.
For example, let's take a pure argon cylinder of size 44 (i.e. 6m3 of gas). On this cylinder there is a double stage pressure regulator with two pressure gauges, CGA connector, and an outlet isolation valve. Lets assume that the internal volume of this pressure regulator is 100 CC ( 10%). When installing this pressure regulator on the cylinder, the internal volume is occupied by the atmospheric air i.e. 78.2% N2, 20.9% O2, 0.9% Ar, moisture, CO2, etc.
When the regulator is screwed in place on the pressure regulator, the air still is trapped inside the regulator. If you open the valve on the cylinder to pressurize the regulator, and there is no or little flow through the regulator, the air trap inside the regulator will diffuse inside the argon cylinder. The shock caused by the quick pressure build up inside the regulator helps to speed up the diffusion process.
So, assume no flow (worst case), we have the following situation:
100 CC of air and atmospheric impurities added to 6 m3 of pure argon (assuming perfect argon i.e. no impurities at all). This leads to the following calculation:
100 x 10-6 m3 (i.e. 100 CC) of Air
 |
= 16.66 x 10-6 |
| 6 m3 argon |
|
So the dilution ratio is 16.66 x 10 -6 and 16.66 x 10 -6 x 78.2% N 2 = 13 ppm of N 2
and
16.66 x 10 -6 x 20.8% O 2 = 3.5 ppm of O 2
So starting from a pure argon cylinder and just by a bad pressure regulator purging procedure, we've got an argon cylinder with 13 ppm of N2 and 3.5 ppm of O2. These impurities will be added to any other impurity in the cylinder. This situation makes it difficult or even impossible to get accurate calibration In some cases, we received phone calls from people claiming that the zero cylinder had higher readings than the span cylinder...
SO BE AWARE !!!!!
Improving argon recovery in air separation plants
with the use of proper process analytical tools. (AN-04)
The argon is produced by air separation plants. The air constituents are nitrogen (78.09 %) oxygen (20.94 %) and argon (.934 %). These constituents are not chemically bonded together but are moving freely. The distillation process can separate constituents of a mixture if their respective vapor pressure are different. This process is based on distillation columns where the most volatile exit at the top and the less volatile exit from the bottom of the column. The argon is took off from a low pressure column and introduced in a separate smaller distillation column called the crude argon column. The figure 1 shows typical curve for vapor pressure of N2, O2 and Ar. Since the vapor pressure of the argon is closed to the oxygen vapor pressure and between the nitrogen and oxygen, the argon will be extracted between these two constituents in the low pressure column.
A typical concentration distribution of N2 / O2 / Ar in the low pressure column is shown in figure 2. According to the curve in figure 2, it is clear that the argon should be extracted at the level where the concentration is maximum. However, at this point the nitrogen concentration is almost the same as argon and there is also a lot of oxygen. The curves are showing 14 % of argon, 14 % of N2, and 72 % for O2. It is not possible in this column to find a point where the argon is pure. In order to have a mixture which can be processed in a single distillation column, the low pressure column must be adjusted in such a way that the nitrogen concentration at the argon extraction point will be at the minimum. This way, the extracted mixture will be almost binary (i.e. " 10 % Ar / 90 % O2 / N2 < 2000 ppm).
This mixture is then fed to the crude argon column where it will be processed. In some plants, the process stops there, so the final product is crude argon. In some other plants, there is an extra cycle to produce pure argon, call the warm argon cycle. In this cycle the O2 in the argon will be reduced with H2. There is also today higher performance distillation column without the need to have the warm argon cycle column. In such column, packing is used instead of trays.
|
In order to have and maintain the optimum argon extraction efficiency, the argon draw-off mixture must be properly controlled. It is not an easy task, and there are two possible problems. First, if the column profile is too low, i.e. the nitrogen contents in the mixture draw-off from the low pressure column is high (> 2000 ppm), the crude argon column will stop working. At the limit, too much nitrogen will block the condenser of the crude argon column, eliminating the reflux. The liquid hold in the trays (essentially argon) will fall in the low pressure column. There will be a fast drop in O2 concentration in low pressure column. The result is lost of O2 and argon production. Many hours must be spend to restart the process.
Secondly, adjusting the low pressure profile too high i.e. O2 level is high, result in a loss of argon in the waste nitrogen. Furthermore, doing so increase O2 level in the crude argon column.
The challenge is to monitor the level of nitrogen in the argon draw-off from the low pressure column. Until now, the analytical tools available for this application were relatively complex custom built systems relatively complex, or chromatographic system with too long elution time and lacking of process control interface. So most of the time the plants are operated with a poor argon recovery efficiency by maintaining a low level of nitrogen in the crude argon to avoid to crash the plant.
|
The Contrôle Analytique's K4000 trace gas analyzer system can be configured to measure trace nitrogen in any mixture of oxygen and argon. The K4000 analyzer uses a separation column at the front end of the system to roughly isolate the oxygen of nitrogen. The detector is based on a plasma emission cell which is very selective to nitrogen. Refresh time less than 60 seconds is easily achieved. The K4000 comes with three operating ranges configured for the application. The most common ranges for low pressure distillation column control are 0-20 / 0-200 / 0-2000 ppm. The K4000 comes also with an isolated 4-20 mA output, three remote range identifications, dry contact outputs and two dry contact process alarm outputs. The K4000 is easily interfaced with any PLC, DCS, computer or other process control device. Automatic control of the argon draw-off is feasible. The system may come with an isolated serial communication port or with automatic calibration subsystem. The K4000 is designed to be operated by a non technical personnel. The system is friendly user, and is almost maintenance free.
When installed properly, it will operate many years without problem. When the analyzer is interfaced with the process control system, the plant may be operated at its optimum efficiency, resulting in a increasing argon recovery up to 5% in some situation. It is obvious that the payback is fast. Normally, the sample connection is made at the point where the argon mixture is extracted from the low pressure column. When the process is stable there is no problem to do so even if the level of N2 a little bit high. But on the same plants there are two big "desiccation bottles" to dry the compressed incoming air and to remove the CO2. One bottle is alternatively switched into the process when the other one is being regenerated. Before to bring back the newly regenerated bottle in process, this one must be pressurized. Pressurizing this bottle involve sudden change in pressure which can lead to an increased level of N2 in the crude argon up to the limit where the crude argon column may dump.
This event may happen in a very short period of time. To help to avoid such situation it is a good idea to monitor the low pressure column from a sample connection located physically higher than the argon draw-off point. This can be done just after the next tray section. Doing so will give the time to react when fast column upset occurs.
|
|
|
