The drive towards closer monitoring of discharges, and the emergence of the Environmental Protection Agency (EPA) Consent to Discharge regulations, have created a new demand for flow measurement in pipes which do not run full.
By their nature, discharge flows can vary greatly on a daily, weekly or seasonal basis, and nearly always flow down hill from a processing works towards the water course. So the part-filled, gravity-flow pipe has always provided an ideal solution. However, retro-fitting a measurement system to meet new regulations – the problem now facing processors all over the country – can be tricky, especially when the pipe is underground.
There are a few technologies available for part-filled pipes, but they were mostly developed for niche applications. Each has its advantages and disadvantages. However, the increasing demand resulting from the emergence of the Environmental Protection Agency (EPA) Consent to Discharge regulations is forcing instrumentation companies to invest in finding more cost-effective solutions.
A partial flow can be calculated by measurement of velocity of the flow, and the liquid level. If the shape of the cross-section is known, then the flow can be calculated. Techniques for measuring the velocity of the flow include Doppler, ultrasonic transit time and, most recently, specially equipped electromagnetic meters for part-filled pipes. For these, ultrasonic level meters or pressure sensors can be used to measure level and so infer the wetted area.
Alternatively, the pipe can be engineered to run full for a short section, so that any of the full-bore measurement methods can be employed. Equally, it can be opened up into an open channel such that a flume system can be installed. These two options both provide good measurement solutions if they are designed and built into the system from the start. But as retrofit solutions to existing underground pipes, the cost of civil work and re-engineering is likely to be extremely high – and thus prohibitive.
Looking at retrofittable measurement technologies, Doppler for measuring velocity uses a small unit or `mouse’ fitted with ultrasonic transducers which sits on the pipe floor. The ultrasonic signal is emitted into the fluid, reflected off discontinuities in the flow, and received back. Difference in frequency between the emitted and received signals is an indication of velocity of the flow.
The technique is limited to liquids which are aerated or have fine suspended particles – such as effluent discharges. Its main disadvantage is that it takes its velocity measurement at a single point, and is not averaged across the pipe. Since the flow profile is likely to vary across the section and with the level of flow, accuracy tends to be unpredictable.
Moving on, the ultrasonic transit time method of velocity measurement, also known as time-of-flight or Aquasonic, uses a pair of transducers mounted on opposite sides of the pipe wall, one upstream of the other. Time taken for the ultrasonic signal to travel through the flow from one transducer, in each direction is logged. The difference between the two times is a measure of velocity.
For large pipes, more than one pair of transducers may be used for better flow averaging. The technique is already used widely on raw sewage applications, but is not suitable for liquids carrying fine gas bubbles, such as can result from cavitiation of a centrifugal pump, since the ultrasonic signal does not transmit through the bubbles.
Finally, the magmeter for part-filled pipes uses an external coil and electrodes mounted flush with the pipe, and is designed for easy retrofit. All the advantages of electromagnetic flow metering apply. They can be used with any conductive liquid, can cope with wide variations in flow rates, give excellent cross-pipe averaging, and offer long-term, low maintenance reliability.
The technology is particularly suitable for discharge monitoring because it maintains the integrity of the pipe and is bi-directional. Studies by the Water Research Council and anecdotal evidence indicate that a field accuracy of at least +/-5% can be expected – better than most flume installations. BS 3680 quotes typical uncertainty for flumes as 2-5% of reading. But since Standard conditions are rarely found, flume accuracy is often 5% at best.
Electromagnetic meters for part-filled pipes are considerably more expensive to purchase than those for full-bore measurement, and more expensive than Doppler or transit time meters. However, when the higher cost of installation and maintenance for other technologies is taken into account, the option often provides the most cost-effective long term solution.
* Author is with KDG Mobrey.