Flowmeter selection for improved gas flow measurements: a comparison of DP and thermal dispersion technologies

As the costs of fuels and consumables continue to rise, the ability to accurately measure the amount used in a process becomes significant in controlling costs and determining bottom line profits. Therefore, it is important to implement a strategy of adding cost-effective, accurate, gas flow measuring devices to heaters, boilers and cogeneration equipment.

Once we’ve decided to add gas flow measurement, it should be relatively easy to select a flowmeter that will meet our needs. Differential pressure with primary flow elements, magnetic, ultrasonic, turbine, venturi, rotameter, coriolis, vortex shedding, thermal dispersion and several other technologies exist.

All of us should understand by now that there are advantages and disadvantages associated with any type of technology used in process measurement. Just the type of fluid that we are trying to measure can limit the options available. Fluids come in the form of liquids, slurries, gases and steam. There are fewer concerns associated with the flow measurement of a liquid or slurry, given that they are considered incompressible and, if homogeneous, have a constant density. Gases require more consideration, given that they are compressible, which results in a density that varies with changes in process pressures and temperatures. Steam presents its own set of complications since not only is it compressible, but it also has a high moisture content at relatively high temperatures. A proven method for measuring a liquid does not necessarily translate into a good solution for measuring a gas.

Consideration should be given, but not necessarily limited, to the following items when selecting a flowmeter for gas measurements: accuracy, turndown ratios, pressure drops, process temperatures, additional sensor requirements and process connections. To help develop a method that will allow us to effectively compare technologies, we are going to look specifically at how these factors are addressed by differential pressure (DP) and thermal dispersion technologies.

Differential pressure technology

The most common method of measuring liquid flow is to use a DP transmitter with a sharp-edged orifice plate. The square root extraction of the pressure drop across the orifice is directly proportional to the volumetric flow rate in the pipe. Other primary flow elements used to take similar measurements with DP transmitters are pitot tubes, averaging pitot tubes, v-wedges, and v-cones.

Volumetric versus mass flow rate Assume a volumetric air flow rate of 5000 ACMH at standard conditions of 0 °C and 1.01325 Bara Actual process temperature = 40 °C Actual process pressure = 3.5 Barg To determine an equivalent flow rate at process conditions, we would calculate it as follows: Note: Temperatures and pressures must be converted to absolute values: °K (273 + °C) and Bara (1 + Barg). Since air is compressible, the mass flow rate can be represented as a volumetric flow rate of 19.368 NCMH.

These same instruments are often selected in gas flow measurement based upon maintaining commonality of instrumentation throughout a facility. While this makes sense from a maintenance and inventory standpoint, our real objective is to improve the gas flow measurement of the process. Since we are now trying to measure a compressible gas, we have to recognise that knowing the mass flow rate is more beneficial than the volumetric flow rate (see sidebar). Without taking into account that the density of a gas will change with variations in process temperature and pressure, a volumetric flow reading will not be an accurate representation of actual gas consumption in a process.

This limitation of volumetric flowmeters in gas applications can be overcome. The addition of pressure and temperature transmitters can provide the data required to compensate for changes in gas density under process conditions. Sending the flow, temperature and pressure readings into the PLC or DCS will allow for the calculation of the mass flow rate. We have now added complexity, extra sensor expense, and extra installation expense to our gas flow measurement (Figure 1). When working with flow elements like an orifice or averaging pitot tube, the use of a multivariable transmitter would definitely simplify our installation.

Figure 1: Mass flow measurement.

Several factors come into play when determining the actual accuracy of a DP transmitter being used with a primary flow element. If you look closely at the manufacturer’s specifications, the accuracy could vary with the span ratio (turndown), percentage of flow rate being measured, long-term drift, temperature effects and static pressure effects. Best-case conditions may provide accuracy better than ±1%, but the true accuracy can be ±5% or greater under actual process conditions. We have yet to take into account the additional inaccuracies associated with the additional pressure and temperature transmitters required because we are trying to determine the mass flow rate. We may further degrade the accuracy if the gas has particles that may build up around the edges of the orifice or plug the small openings in a pitot tube over time.

When using a DP transmitter with an orifice, the turndown ratio would be close to 10:1, maybe 20:1, depending upon the transmitter. This could become a significant issue when the required gas flow is high for one process and very low for another. Without adequate turndown, we may end up with a meter that is only capable of accurately measuring on the high end of the flow range. It is a common practice to ‘stack’ meters of varying ranges to take readings from the same primary flow element in order to increase the measured flow range, further increasing the cost and complexity of our system.

The use of a sharp-edged orifice or any other type of primary flow element is intended to create a measurable pressure difference. Although pressure drop is not critical in all gas applications, it does impact the efficiency of a process in the form of wasted energy. For an orifice plate, this loss could be significant over the life of our process. Averaging pitot tubes or v-wedges can limit those losses by reducing the size of the obstruction in the flow line. In the case of an orifice, this loss can be in the neighbourhood of 125 mbar in a 100 mm line for a flow rate of 5000 NCMH. Given the same conditions, that value may be less than 50 mbar for an averaging pitot tube and v-wedge.

When performing mass flow measurements, we must also take the actual process temperatures into consideration. There are applications in heating and cogeneration systems in which the temperatures can be quite high. Most DP transmitters are rated for temperatures up to 120 °C at the point of the measuring cell. For applications that will be significantly higher than this, say 260 °C or so, it will be necessary for us to use impulse tubing or a process (chemical) seal in order to dissipate the extra heat from the process. The use of impulse tubing or process seal is not a major concern, but we should be aware that it will slow the response time of the meter, add cost and complexity to our installation and require adequate elevation if there is condensation in our lines.

Proper installation of a DP transmitter to a primary flow element adds to the complexity of our installation. In order to ensure accuracy, DP transmitters do require periodic calibration checks. In order to perform these checks with the system operating, installation of a manifold between the flow element and transmitter is common - which is another item that will add to our installed costs.

Thermal dispersion technology

Another technology often utilised in gas flow applications is the thermal dispersion flowmeter. This particular instrument makes use of two high-precision RTDs. A reference RTD measures the process temperature and an active RTD is heated to a known value to create a differential temperature between the two sensors. When there is no flow, the differential will be at its greatest. As the gas begins to flow, the active RTD begins to cool and decreases the differential between the two sensors (Figure 2). This is an oversimplification of the operating principle, but provides a basic understanding. Thermal technology is advantageous because it also takes into account the density, absolute viscosity, thermal conductivity and specific heat of the gas being measured. The end result is a very accurate mass flow reading that requires no additional instrumentation or calculations.

Figure 2: Operating principle of thermal dispersion.

The accuracy of a thermal mass flowmeter is very straightforward. It is commonly broken into two components: a percentage of reading and a percentage of full scale. These instruments are immune to long-term drift, are commonly compensated for broad temperature ranges and the effects of pressure changes are negligible. So, to really understand how technologies compare, we have to look at our worse case process conditions and run the calculations.

For most applications, we can expect to achieve a turndown ratio of 100:1 with a thermal meter. This allows us to maintain a high level of accuracy over the entire flow range without having to ‘stack’ multiple instruments.

The pressure drop across a sharp-edged orifice versus the drop across a single-point thermal flow element can be in the magnitude of 5 to 10 times greater. The most significant difference can be observed when we are operating 70 to 100% of the maximum flow range. Using our example of a flow rate of 5000 NCMH in a 100 mm line, the pressure drop is in the neighbourhood of 35 mbar for a thermal meter versus 125 mbar for a DP meter and orifice.

Thermal meters are inherently suited to high-temperature applications. Since we are literally dealing with RTDs in thermowells, standard temperature capabilities of these meters run to about 175 °C. With modifications to the design of the flow element, some manufacturers offer variations suited to process temperatures as high as 260 to 450 °C that require no added installation considerations.

The installation of a thermal flow element is simple. In the case of an inline meter, the elements can be provided with either threads or flanges. With insertion type elements, it is common to install the units with a threaded compression fitting. Unlike a DP transmitter, periodic calibration of a thermal meter is not required, but manufacturers may recommend that a calibration check be performed every 12 to 18 months. In the case of processes that run continually or involve dirty gases, the use of a packing gland and ball valve assembly is recommended, with an insertion meter for extraction of the flow element for either calibration or inspection and cleaning.

Like any other instrument, thermal dispersion technology does have limitations and is not ideal for certain applications. First and foremost is that these instruments are not suitable for measuring the flow of liquids, slurries or saturated steam. Thermal technology is best suited for the measurement of dry gases, gases with limited moisture or superheated steam (no water vapour).

We must also keep in mind that thermal meters are normally calibrated for a specific gas composition. For instance, this can be a single gas such as air, hydrogen or oxygen, or a composition like natural gas (methane and ethane). If the composition changes, the mass flow reading will remain repeatable but it will no longer be as accurate. The use of a correction factor may improve the accuracy to acceptable limits for some processes.

If our process has condensation in the lines, a thermal meter may provide false readings due to the cooling of the active RTD that is not directly related to the flow rate. In some cases, appropriate positioning of the flow element in the pipe can reduce or eliminate this effect. Other cases might require the use of condensation (knock-out) pots or filters to reduce the moisture content to acceptable levels.

Ideal versus actual flow conditions Another factor that impacts accurate gas- flow measurement is the upstream and downstream (straight-run) piping conditions. For line sizes up to 150 mm, it is normally accepted that a straight run of 20 pipe diameters (ie, 20ø) upstream and 10 pipe diameters (ie, 10ø) downstream from the metering point is required for a fully developed flow profile. The acceptable requirements for lines over 150 mm are 15ø upstream and 7.5ø downstream. Although it is realistic to be able to find a 2.3 m straight run of 75 mm pipe, it is more difficult to locate an appropriate 6.9 m run for a 300 mm pipe.

Figure 3: Common types of flow conditioners.

We need to therefore understand the flow disturbance created by our actual piping conditions. Knowing this will help in selecting a conditioning device that will properly address our needs. Common types of flow conditioners are perforated plates, screens, vanes, tube bundles and tabs (Figure 3). These are all simple, mechanical devices that are installed in the process piping before the metering point. Perforated plates and screens do an adequate job of generating a measurable velocity profile but have limitations when it comes to swirl. Tube bundles and vanes provide better conditioning for swirl, but allow a good portion of a distorted velocity profile to move on to the metering point. Only the tab design has been shown to eliminate the effects of both a distorted velocity profile and swirl by generating a very repeatable and measurable velocity profile. In the case of thermal flowmeters, some manufacturers have integrated flow-conditioning devices into their flow element (Figure 4). This is a great benefit to us, since not only have we reduced the amount of straight run required for our installation, we now have an instrument that has been fully calibrated to our process conditions and will provide us with a very high level of accuracy in our measurement.

Figure 4: Thermal flowmeter with integral flow conditioner.

Fluid Components International LLC
www.fluidcomponents.com
AMS Instrumentation & Calibration Pty Ltd
www.ams-ic.com.au