Ultrasound abilities assessed
Alick MacGillivray on an NEL ultrasonic flowmeter performance study.Although it is a mature technology, ultrasonic flow measurement has recently become increasingly popular in the industrial flowmeter market. The Ultrasonic Flowmeter World-wide Outlook, a market analysis published by the ARC Advisory Group, forecasts double-digit growth in sales of these devices until 2005. The first generation meters, introduced in the 1970s, had problems with repeatability and linearity and the technology as a whole gained a reputation for unreliability.
Thirty years later, however, measurement techniques have been perfected. Technology services company NEL has conducted a series of tests on its National Standard water flow measurement facility to investigate the behaviour of these meters in a range of pipe sizes and materials, in different pipe configurations and in fluids of different viscosities and temperatures.
The two most popular types of ultrasonic flowmeters use either transit-time or Doppler technology. Transit-time meters use the time difference for a sonic pulse to travel a fixed distance, both with and against the direction of flow. They have two different operating modes: time domain and frequency domain. Both modes transmit pulses from a transducer to a receiver and back again through the fluid. Time domain meters use the time difference between the upstream and downstream transit times to calculate the fluid's velocity. Frequency domain devices convert time into frequency. Again the difference in frequency between the flow and counter-flow directions is proportional to the flow velocity. Some meters also include temperature compensation to convert volumetric to mass flow rate.
Most Doppler meters use a continuous transmission of a single ultrasonic frequency. Bubbles, entrained solids or eddies act as reflectors in the flow scattering the ultrasound back to a receiver. Since these reflectors are moving, the reflected ultrasound is Doppler (or frequency) shifted from the transmitted signal. The difference between the transmitted and received frequencies is proportional to the flow velocity.
Ultrasonic meters offer a unique capability for measuring flow with transducers that are mounted on the external surface of a pipe. The potential advantages of this non-invasive method include:
- zero additional pressure loss,
- ease of installation and maintenance without flow interruption,
- protection from fouling or impact damage from entrained debris.
Another advantage of ultrasonic meters is that relative costs decrease as pipe diameter increases because the component parts do not need to be scaled-up. Furthermore, diagnostic capabilities such as signal quality and velocity of sound measurement can be utilised for both meter health and process condition monitoring.
Clamp-on ultrasonic meters are commonly used in the water industry primarily as a means of flow verification. However, because of the recent improvements in their performance, their use as a primary measurement device is increasing and they now act as a credible alternative to insertion and electromagnetic flowmeters for large- bore applications.
The National Measurement System Directorate (NMSD ) recently identified that the need for quality data to aid users and support the development of standards is acute in relation to clamp-on ultrasonic meters. The main reasons for this are the wide range of process conditions under which the meters are used and the commercial confidentiality involved in their development. To identify the most important factors that affect design and accuracy a series of tests were conducted at NEL on a selection of commercial clamp-on ultrasonic meters manufactured by several different companies.
Meters were obtained from four companies which were selected as providing good coverage of the range of products available. The calibration work was carried out using the 100, 200 and 600mm nominal bore test lines of the NEL National Standard water flow measurement facility. The meters tested in the project provided output in the form of a current or frequency. When recording a current output, the average value during the test period was determined and converted to volumetric flowrate for comparison with the reference. Test velocities ranging from 0.1-5m/s were selected. For the pipe material tests, three commonly utilised industrial pipes were chosen: a naturally corroded mild steel, a medium density polyethylene (MDPE) pipe and a cement lined ductile-iron pipe.
Each of these were obtained in 200mm nominal bore for good comparison with the baseline 200mm (stainless steel) calibration results. The reproducibility of the meters' performance under the condition of removal and reapplication of the transducers was investigated. This involved the complete removal of the transducers, coupling medium and fasteners on the 200mm nominal bore stainless steel pipe.
The performance of the meters when subjected to various installation effects was also investigated. Results were obtained at three nominal distances downstream of four pipe components, namely a contraction, an expansion, a single 90° bend and two 90° bends out of plane.
In the description of the results, repeatability is used as a measure of the random uncertainty in measurements. Non-linearity is derived from the maximum difference between errors within a given velocity range based on the average of a number of repeats at each point.
In general, performance improved with increasing pipe size. This is illustrated in Figure 1 which demonstrates linearity improves with increasing bore size and the effects are more significant at lower velocities.
Application on certain pipe materials produced greater errors but for the range of materials and conditions tested measurements could always be made. Reproducibility was shown to be affected by the placement of the transducers. The consequence of misplacement of transducers can lead to inaccuracy or lack of repeatability or both. Installation effect tests demonstrated pipe components installed upstream of the meters produced deviations which varied in magnitude by up to 7-6%. Results obtained show that with sufficient care in selection and application clamp-on ultrasonic meters, are capable of performance within ±5% uncertainty over a range of conditions when velocities are greater than 1m/s.