MCerts: The fuller picture
The introduction of MCerts for self-monitoring of discharge flows focused attention on areas of flow measurement that had been long overlooked. But the story is far from complete. WRc's Andy Godley reportsIt is more than two years since the Environment Agency's MCerts scheme for self-monitoring of discharge flows became mandatory. This meant that those discharging treated water to the environment had to measure and report their discharged volumes in accordance with a series of requirements defined by the agency.
The scheme is policed by regular inspections of discharge measurement facilities by accredited inspectors who will certify that the facility is capable of meeting the agency's requirements and that appropriate procedures are in place to enable the performance to be maintained between the five-yearly inspections.
Introduction of the scheme focused attention on areas of flow measurement that had been long overlooked and triggered much work to better understand flow measurement in this application. But the story is far from complete, and this article looks at some of the issues that are still being worked on and some of the remaining unanswered questions.
One of the key principles in the Minimum Requirements document is uncertainty (discharges have to be reported to an uncertainty of +/- 8% total daily volume) and the use of an uncertainty budget to calculate the uncertainty for any individual installation. The uncertainty budget brings together the uncertainty of each component of the measurement that might give rise to an error and combines them in a root square sum. For example, Equation 1 shows the components of uncertainty that might be considered for a measurement using an open channel gauging structure.
Much of the work done so far, particularly that by WRc and Cranfield, has been concentrated on open channel flow measurement structures - specifically rectangular-bottomed long-throated flumes - to understand how the discharge co-efficients vary when the design parameters of the flume are varied. This was necessary as many in-service structures do not strictly conform with the design requirements described in BS3680 Part 4.
The cost of replacing all such structures would have been enormous and hence if it could be shown that a non-compliant structure could still meet the overall uncertainty required, then these costs could be avoided. This was acknowledged by the Environment Agency when it compiled the "derogations" (or allowable deviations from the Standard) which are contained in the Minimum Requirements. Though these were based on expert opinion, many practitioners felt that they needed some back up.
Much of the WRc and Cranfield's work was based on numerical modelling of structures using computational fluid dynamics (CFD) as a quick and relatively easy way of assessing the relative effects of a large number of parameters. Cranfield backed up some of their results with some limited experiments but there is now a call for further experimental evidence to support the results obtained from the numerical models, which in many cases, show that performance of a structure can be maintained outside the current design limits in the standard. Such evidence will enable development of the Standard and hence needs to be robust enough to be acceptable to the appropriate ISO and BS committees. Assuming that the experiments validate the CFD, it will also support the case for using numerical modelling to explore other types of structure.
Joint industry project
Support for further work on open channel flow structures has been announced through the DTI Flow Programme with funding allocated in the current programme. However, the DTI support is contingent on a degree of co-funding from the industry. A proposal for a joint industry project is currently being developed by the National Engineering Laboratory (NEL) with help from WRc, and is for release this autumn.
It is likely that, as well as reinforcing the existing work, the new project will address two further specific issues that require investigation, namely small flumes (i.e. less than100mm throat width, the minimum in BS3680) and performance at very low flows (less than 50mm head, the minimum in BS3680).
This is important as many small discharges which come under the MCerts scheme are operating at flows that are well below the minimum limits in the standard for flumes. V notch weirs could be used but often suffer from fouling and hence need frequent maintenance - often not practical or cost-effective on a small remote site.
Attention has also been focused recently on the role of the level sensor in an open channel flow measurement system. Typically, on account of their cost, ready availability, ease of installation and lack of maintenance, the majority of level sensors used in this application are ultrasonic devices mounted above the channel and firing down at the liquid surface. There is a view in some parts of the industry, including among some MCerts inspectors, that the uncertainty value being ascribed to this type of level sensor in the uncertainty budget is too low. This is particularly significant as there is a power law relationship between fluid depth and flow, any errors in the measurement of fluid level are magnified in the calculation of flow, for example for a flume where flow is proportional to level to the power 3/2, a 4% error in level gives rise to a 6% error in flow.
Their concerns are based around the possibility that there may be some poorly designed instruments available which do not adequately cope with variations in ambient conditions. It is well known that the speed of sound through air is dependent on the air temperature and manufacturers generally incorporate some form of temperature compensation. Typically there are three forms that this can take:
· Temperature probe integrated with the ultrasonic head
· Separate temperature probe mounted remotely from the ultrasonic head
· Reference bar - where the instrument measures sound speed directly by firing pulses at a target a known distance from the head
However, concerns have been raised with all three approaches. With the first, it is suggested that absorption of the solar energy from direct sunlight striking the sensor can raise the apparent temperature measured by the unit to higher than the surrounding air, thus giving errors in the compensation. These systems may also have a thermal lag, with the sensor heating up or cooling slower than the surrounding air. Some manufacturers now recommend sun shields to be mounted over the sensor to minimise any potential errors due to such effects.
With a separate temperature probe, there is the problem of probe location in order to ensure that the temperature measured is representative of the mean value between the fluid surface and the head. If it is too far into the channel, it could be in deep shade. If it is too close to the channel wall, it could be influenced by radiated heat from the wall.
Fouling on the bar
Finally, with the reference bar approach, there is the possibility of fouling on the bar from various sources including dew, ice, cobwebs and other debris giving rise to an incorrect reference measurement. If these concerns are as serious as some in the industry are saying, then many installations that currently have valid inspection reports and MCerts certificates could fail certification when they are next inspected. And a greater uncertainty is then allocated to the level measurement. A new project being launched under WRc's Portfolio research programme will address these issues.
The project will investigate the performance of level sensors in relation to temperature and ambient conditions through a rigorous experimental programme. The results will enable users to install and use ultrasonic level sensors in such as way as to minimise their uncertainty and reduce the risk of failing their MCerts inspections. Also minimising the uncertainty of one component in the uncertainty budget may allow more flexibility in other areas, as long as the overall requirement is still met.
This issue of compensation for sound speed variation as it affects ultrasonic level sensors was also considered when WRc helped the Environment Agency develop Version 2 of the MCerts product standard for flowmeters. A specific test has been put into the latest version of the standard, which has recently been released via the agency's MCerts website (www.mcerts.net). This will help identify those well designed instruments with good compensation for changes in the speed of sound.
The MCerts product certification process is also being streamlined to give manufacturers more control over the costs of certification and elements of the testing. While open channel flowmeters account for the majority of installed discharge meters, there is increasing interest in using closed pipe meters in such applications.
These are attractive in a number of cases, for example on small sites where traditional open channel systems have problems, as mentioned above, or on modern package plants where inlets and outlets are enclosed. They also offer good measurement performance and lower maintenance requirements.
WRc recently carried out a major study on behalf of a number of water companies under its Portfolio research programme to investigate the application of closed pipe meters to discharge flows. One of the main areas of concern for the project contributors was meter installation. Due to high land costs and planning considerations, many small plants are being fitted on to smaller and smaller sites which limits the space available for long pipe runs.
Modern electromagnetic meters are known to be reasonably tolerant of hydraulic installation conditions - most manufacturers now recommending 5D or less of upstream pipe after a wide range of fittings. But how close could a disturbance be such that the installation could still meet the requirements of MCerts? During the project, WRc test a meters of 100mm and 200mm bore directly coupled to a number of fittings, including bends, double bends, valves, incoming branches and reducers) as well as combinations of fittings. (See Fig 1 on page 38.)
The results showed that in most cases, the change in error was small - less than 2%. The exception was a half-closed gate valve, which gave significantly higher errors. This work helped the development of guidance soon to be issued to the MCerts inspectors on allowances for uncertainty for closed pipe flowmeters close to sources of disturbance.
Thus an inspector will not automatically fail an installation if the usual 5D upstream of the meter is not available but can put a sensible figure into the uncertainty budget to allow for the installation effect.
One significant area that still needs addressing is how to encourage acceptance of new flow measurement solutions. There are a variety of alternative devices that have become available in recent years for both open channel and closed pipe measurement. However, manufacturers are finding it difficult to get these installed on sites subject to the MCerts inspection regime. Site managers do not want to run the risk of failing inspections by installing something non-conventional that may be rejected by their inspector, and the inspectors may require a higher level of proof of performance for a device with which they are not familiar than for a conventional instrument.
The agency has frequently stated that one of the aims of the Mcerts product standards is to encourage innovation. As Version 2 of the flowmeter standard incorporates far greater flexibility for the field trial element of certification, perhaps there is a case for closer collaboration between the MCerts inspections and MCerts product certification to facilitate trials of new equipment.
Despite the fact that there are still a number of outstanding issues to be resolved, there is no doubt that MCerts for effluent discharge monitoring has spurred research and development on a number of fronts and generally improved the quality of discharge flow measurement.