01 August 2003
Clean air laws inspire constant pressure
The measuring cell is the heart of a continuous analyzer.
By Stephen Jacobs, Frank Ruiz, and Steve Doe
For an analyzer to remain calibrated, it is imperative that the sample pressure arrives at the precise pressure at which technicians calibrated the analyzer.
There are many types of analyzers operating today in both industrial and non-industrial applications. Almost all of these analyzers fall into the major categories of continuous sampling and batch (discrete) sampling.
A continuous analyzer requires a constant flow of sample through the analyzer's measuring cell. This produces an ongoing measurement or analysis of the sample stream.
A batch analyzer, such as the gas chromatograph (GC), operates through a timed cycle. The sample enters the analyzer at the beginning of the cycle, and the actual analysis takes place in the remainder of the cycle. The cycle time varies from approximately one minute for a fast cycle to an hour or more. Typically cycles range from five to fifteen minutes.
Although the continuous analyzer provides faster and instantaneous results, there are trade-offs. The continuous analyzer can usually measure only one or two components in the sample stream, while the GC can easily measure 10 or more. Therefore, the user's needs and the system requirements dictate the choice of analyzer. We will consider the continuous sampling systems here.
Continuous gas analyzers operate in applications such as stack monitors to measure oxygen, carbon dioxide, carbon monoxide, and NOx. They serve in other industrial applications as process controllers, ambient air monitors, and environmental monitors.
These analyzers take many forms, but they have similar sample conditioning requirements such as pressure, temperature, and flow.
The most important parameter by far is to maintain a constant pressure in the measuring cell. The measuring cell is the heart of the analyzer where the actual measurement takes place.
Isolate these dynamics
At least two gases are required to calibrate a continuous analyzer. One is the zero gas and the other is the span or calibration gas. A typical calibration procedure is:
- Introduce the zero gas into the analyzer. Allow time for the analyzer to stabilize.
- Adjust the analyzer output to a zero reading.
- Introduce the known or span gas and allow time to stabilize.
- Adjust the analyzer's output to equal the value of the span gas. That is to say, if the known gas is 9.82% oxygen, simply adjust the analyzer to output a reading of 9.82% oxygen.
- Valve the process sample stream into the analyzer. The analyzer is now calibrated and in service.
Although this calibration procedure is relatively simple, analysis results will be erroneous if the analyzer's measuring cell is not operating at a constant pressure throughout the sequence.
Most systems today have the zero, calibration, and process streams set at different pressures. Without flow control, input pressure differences will affect the flow rate through the analyzer's measuring cell, resulting in cell pressure change.
The pressure is critical to the accuracy of the analyzer. While going through the calibration process, the technician must readjust the flow at each step to ensure a constant flow and pressure through the analyzer.
For the analyzer to remain accurate, it must calibrate and operate at the same pressure conditions. If the cell pressure changes during process stream analysis, the analyzer will not be accurate.
The sample system we will discuss is able to automatically adjust to a constant cell pressure regardless of variations in supply and outlet pressures.
Sample disposal presents a challenge to maintaining constant pressure within the analyzer cell. Generally speaking, if the user cannot return the bypass and sample stream to the process, disposal choices are limited to atmospheric vent or the flare header.
Although an atmospheric vent provides downstream pressure stability as constant as barometric pressure, compulsory and stringent Clean Air Act laws force many users to connect their disposal streams to the flare.
Oftentimes, such a change in process pressures can introduce pressure and flow fluctuations to the analyzer cell.
The flare header is likely to be the most dynamic and unpredictable system in a process plant. Pressure and flow fluctuations can develop back pressure at the analyzer cell, which exceeds the accuracy requirement.
To isolate these dynamics from the analyzer, a vent recovery system is useful. Several approaches have worked over the years, but all have limitations.
Vent recovery alternatives
A simple solution and the one most often employed to isolate the analyzer cell from downstream fluctuations is to install a back pressure regulator on the outlet of the analyzer.
This solution helps solve part of the problem, but we still see pressure fluctuations resulting from changes in flow. This is because a pressure regulator is a simple proportional-only controller.
A proportional-only controller changes its valve position in proportion to the deviation of output pressure from set point. In other words, the regulator will adjust flow rates only after there is a droop or rise in output pressure.
The amount of change corresponds to the gain of the regulator. This gain is inherent to the design of the regulator and is not adjustable. Regardless of manufacturer, the outlet pressure of any regulator will change with the flow rate. This is the droop characteristic of a regulator.
Therefore, a vent recovery system must incorporate a flow controller to maintain a constant flow through the regulator so it can maintain a constant outlet pressure.
Vent recovery system panel
We installed this new vent recovery system recently on a hydrogen system at Eastman Chemical Company in Kingsport, Tenn. The hydrogen process stream undergoes analysis for low levels of carbon monoxide, with the return sent to a common collection vent.
The previous sample conditioning system consisted of two single-stage pressure-reducing regulators with back pressure regulation of approximately 30 pounds per square inch, gauge (psig). The system technician manually set pressure and flow regulation for zero and span calibration gasses, a practice that was both time consuming and subject to human error.
The new Parker Hannifin sample-system panel incorporates stream switching, sample filtering, pressure regulation, and flow control.
Stream switching is the result of a three-stream switching system. This system provides air-operated, double block and bleed stream selector valves for calibration, zero, and sample streams.
The double block and bleed arrangement ensures that the sample does not mix with leak-through from other streams—cross contamination. Each stream can activate using a three-way solenoid valve remotely controlled through the analyzer via Ethernet communication.
The stream switching system can also have an optional integral bypass filter. The term bypass means that most (approximately 90%) of the inlet flow bypasses the filter element and exits the filter bowl, which serves three purposes:
- It reduces the transport time of the sample stream from the process line to the analyzer.
- The high bypass flow provides a continuous flushing action of the filter element.
- The filter life is longer because only a fraction of the total flow passes through it, and only when that particular stream is operating.
After selecting and filtering the stream, the flow routes to the inlet of a point-of-use pressure-reducing regulator. The point-of-use regulator is a very sensitive pressure-controlling device, with a large surface area metallic diaphragm.
Regulator outlet pressure range selection is critical, as the analyzer cell pressure setting must always be higher than the flare header's worst case scenario. For optimum system performance, the cell pressure should always be at least 2 pounds per square inch higher than the system's outlet pressure.
The flow controller is composed of two components: a sensitive back pressure regulator and a flow-restricting device such as a needle valve. The back pressure regulator is set up to function as a differential pressure regulator. It controls the pressure across the needle valve at a constant differential; therefore, if the differential across the valve is constant, the flow is constant. The needle valve adjusts to deliver different flow rates, but the pressure differential across the valve will always remain constant.
The amount of differential adjusts by turning an adjustment screw on the bottom of the back pressure regulator. Once the system tunes for an application, there is no reason for readjustments.
The pressure-reducing regulator provides pressure isolation from varying upstream pressure fluctuation, and the back pressure regulator provides flow control and isolation from downstream pressure fluctuation.
The combination of pressure regulation and flow regulation provides the required pressure stability within the analyzer cell.
Continuous analyzer venting to atmosphere
When an analyzer is vented to atmosphere it is very stable; however it is subject to atmospheric pressure fluctuations. It is not uncommon, usually over an extended period of time, for the barometric pressure to change from 31-inch Hg down to 29-inch Hg.
This 2-inch Hg change will produce an error of 6.5% in the analyzer reading. For example, if an analyzer is reading 1% O2, a 2-inch Hg drop in barometric pressure will cause the analyzer to read 0.93% O2 , a relatively small deviation not likely to cause a problem.
However, if an analyzer is reading pure oxygen (100%), this 2-inch Hg drop in barometric pressure will cause the analyzer to read 93.5% O2 , a significant difference. The vent recovery panel will accept a barometric pressure compensated reference pressure.
This reference pressure would connect to the dome of the pressure-reducing regulator, enabling the analyzer to operate at a constant absolute pressure that is completely independent of barometric fluctuations, eliminating the error.
Certainly for now, and far into the foreseeable future, the analyzer engineer will always look first to the atmospheric vent for continuous analyzer sample disposal, because it is constant, always available, and the most cost effective.
However, for each system that can vent to atmosphere, there will be a system that, due to environmental requirements, cannot. The vent recovery panel handles the vast majority of these recovery applications.
The unit installed at the Kingsport site replaced a system that simply monitored carbon monoxide in the hydrogen stream, where variability was too great to accomplish anything but manual intervention upon high alarm or observing trending rules. P
Behind the byline
Stephen Jacobs, Ph.D., is a development associate chemist at Eastman Chemical Company. Write him at email@example.com. Frank Ruiz is a consultant and the principal at Ruiz & Associates. Contact him at firstname.lastname@example.org. Steve Doe is the analytical market manager at Parker Hannifin. E-mail him at email@example.com.
Qualify the system
Engineers performed several laboratory tests on this new system prior to installation in the field. The tests simulated various up- and downstream dynamics to determine pressure stability at the analyzer cell.
In the testing phase, the analyzer was a digital gauge with 0.01-pounds per square inch resolution. In the first test, the analyzer pressure set to 10.00 pounds per square inch, gauge (psig), and the inlet pressure varied from 20.00 psig to 40.00 psig, then back to 20.00 psig.
The analyzer cell pressure stability was 0.06 psig.
To determine the percent analysis error, the cell pressure fluctuation divides by the cell set pressure (psia = psig + patmosphere) and then multiplies by 100.
0.06 / 24.73 x 100 = 0.24% = the percent analysis error
Analyzer cell pressure curve inlet
A condition that frequently occurs in flare headers is high positive pressure fluctuations. To prevent these high-pressure spikes from affecting the analyzer, the analyzer pressure can adjust to operate 2 psi higher than the highest possible header pressure.
If the flare header pressure rises to closer than 2 psi of the cell pressure, a pump or vacuum ejector must cut in. The sample pump can easily discharge into line pressures exceeding 15 psig while producing a vacuum at the outlet of the vent recovery system.
Tests involving varying outlet pressures from 1-inch Hg (one inch of mercury or about 0.5 psi) to 14-inch Hg then back to 1-inch Hg took place. These tests simulated a sample pump sucking on the outlet of the vent recovery system.
Negative downstream pressure swings had little impact on the analyzer cell pressure.
Analyzer cell pressure curve outlet fluctuation