Flow measurement – future directions

Citation

Sproston, J.L. (2002), "Flow measurement – future directions", Sensor Review, Vol. 22 No. 3. https://doi.org/10.1108/sr.2002.08722caa.002

Flow measurement – future directions

Flow measurement – future directions

John L. Sproston

Keywords: Flow metering, Spignal processing

Interest in the flow of liquids and its measurement can be traced back to early studies by the Egyptians, the Chinese and the Romans.

The Egyptians, for example, who employed large irrigation ditches and weirs for the systematic distribution of water to the fields, clearly demonstrated that they possessed a fundamental knowledge of flow measurement. Since these early times the science of flow measurement has undergone a massive change but during the last 25 years or so it has matured enormously. One of the principal reasons for this is that higher accuracies and reliabilities have been demanded by industry in the measurement of fiscal transfers and today there is vigorous interest in the subject from both the flowmeter manufacturer and user viewpoints. This interest is coupled with the development of advanced computer techniques in fluid mechanics together with the application of increasingly sophisticated electronics.

But the science of flowmetering has not yet reached the vicinity of its asymptote (as most sciences have not) and currently there are many examples of exciting developments in various types of flowmeter. Currently it is estimated that world-wide there is a flowmetering business worth approximately US$1 billion and that there is an increasing tendency towards the industrial application of the more sophisticated flowmeters such as electromagnetic, Coriolis and ultrasonic.

In process instrumentation relatively cheap microprocessors have been embedded into flow measurement devices to effect operation such as linearisation and compensation and these devices are commonly referred to as "smart". But such a device does not provide additional information on either the operational status of the device itself or of the process to which it is applied. It is to this latter area that current effort is now being directed to ascertain whether this information can be extracted from the flowmeter by analysis of the unconditioned sensor output and hence used in condition monitoring and fault diagnosis. There are two ways of achieving this – by employing either hardware redundancy, which involves additional elements for change detection and diagnosis, or software redundancy which involves extraction of information and judicious signal processing to achieve a comparison between, for example, a self-learning knowledge base and some estimated parameter.

It is clear that the acquisition of this additional information will not only increase operator confidence in the flowmeter reading but will also reduce the possibility of a catastrophic failure in the process.

Preliminary collaborative work carried out two years ago between the University of Sussex, Brunel, UMIST and Liverpool for the National Measurement Systems Policy Unit of DTI concentrated on the evaluation of the signals from electromagnetic, Coriolis and ultrasonic flowmeters. These were used in liquid flows and various signal analysis techniques were applied to identify information in the unconditioned signal outputs for different flow conditions including swirling, pulsating and two-phase (liquid/air) flows. The results of this study showed that all three flowmeters demonstrated the existence of separate independent spectral features (fault signatures) and were capable of detecting some or all of these flow conditions. Another important feature of the work stems from the fact that the output of the flowmeter when installed in a particular flow regime encountered in the process plant can be assessed and compared with that of the flowmeter when initially calibrated (under near ideal conditions). This would provide a measure of how far the performance of the flowmeter was short of its calibrated specification. In conjunction with the application of, for example, fuzzy logic and the more recent chromatic methodologies to the interrogation of the output signals this feature shows great promise for fault diagnosis.

Such fundamental studies on flowmeters have continued in these (and other) universities as well as in various research and industrial institutions and the future for signal extraction and utilisation now appears not only an assured but also an exciting commercial prospect.