Meeting an increasing demand for sewer flow monitoring
With ever more strict legislation, sewer networks are increasingly turning to flow monitoring, writes Steve Russell, WRc's senior instrument engineerSewer networks operate with practically no installed monitoring of any kind. This simple picture is beginning to change for two main reasons. Firstly, consumers expect ever higher standards from service providers of all kinds, and sewerage undertakers are being challenged by this. Secondly, environmental legislation is demanding much higher performance from wastewater collection and treatment systems than was envisaged during their design.
Sewer flow monitoring could have a useful role in at least three main areas; asset/risk management, regulation, and operations.
Under the risk-based approach of the Common Framework, improvements to the condition of the sewer network will be funded only if it can be shown the sewer poses a risk to service to customers, or where it is necessary to meet new environmental requirements. Unless whole-life cost considerations force expenditure, low-risk sewers will therefore be operated to failure. For parts of the network at high risk however, due diligence requires water companies to deploy available technology at reasonable cost to minimise the risk.
An obvious possible application for flow metering is in monitoring high-risk sewer rising mains, where flows at the beginning and end of the pumped length can be monitored, and alarms raised if there is a discrepancy. Flow monitoring of rising mains can also provide early warning of pumping station failures, where emergency overflows often pose a high risk to the environment.
By contrast, on high-risk gravity sewers, there would be little mitigation of the risk by measuring open channel flow at each end. Compared with closed pipe flow, the achievable accuracies are very much poorer, and large exfiltration or infiltration could occur undetected.
It may, however, be worth measuring the levels at each end of a high-risk section. Intelligence could then cross-check the levels and alarm if the difference between them is greater than usual, or if individually they are outside the normal range.
The risk-based approach depends on the operator knowing the risks of the sewer network. Flow or level measurement could be used to monitor for changes to risk due to increased loading.
Level measurement would let operators measure the frequency of surcharging, flow would track rising loads. Provided the monitoring was cheap enough, this could be an economic means for the sewerage operator to be up to date on the risks due to loading. A judgement could then be made on pre-emptive action, to avoid possible damage to property and the environment.
At a recent Water Industry Flow Club meeting, James Blackmore of the Environment Agency suggested that storm sewer overflows will in future be regulated by monitoring the discharge and treatment flows, rather than through the engineering parameters of the structure such as storage volume and design flow to treatment. The timescale was not given, but if this goes ahead as a part of self-monitoring, it will generate a significant new market for sewer flow measurement. For new build CSOs, flumes or V-notch weirs may be included to make these flow measurements. Beyond just monitoring however, this development would open the door to control of these flows as part of the agreed management regime for a catchment. Such a scheme would place much greater responsibility on water utilities, but they are in a unique position to carry out active protection of the environment during a storm event, and it would be rational to encourage them to do this well.
In WRc's recent work on sewer instrumentation for operational requirements, the focus has been on measuring levels in sewerage rather than flows. Knowledge of levels provides some information about how much sewage there is at points in the network and therefore indirectly some information about flows. Much of the level monitoring is for CSOs, either for regulated discharge, and/or to monitor to prevent pollution incidents. However, in the wider sewer network, WRc is currently working with water utilities on affordable systems for monitoring sewer levels, both to prevent flooding and to schedule maintenance.
UKWIR currently has a desk study project on the possible opportunities for integration of sewerage and sewage treatment. This could be the beginning of a second look at real-time control (RTC) in UK sewerage, which was seen by many in the industry as an expensive toy in the 1980s. The bad press of that time was for large complex automation schemes which were unlikely to deliver net benefits.
More modest schemes which use control of sewer flows to improve the management of network and treatment should have a place. Perhaps a new acronym is needed, Integrated Network and Treatment Management, INTM? The potential benefits are well known:
- Utilisation of the storage capacity of the sewer network to control the flow to treatment and minimise discharges from CSOs
- Deferment of sewage treatment works upgrades
- Controlling flows in the network to prevent the build-up of sediment loads during dry weather
There is a trend in water treatment SCADA systems toward creating greater integrity in the process by online cross-checking of data. If sewers do begin to operate with flow and level data at several points in the network, such cross-checking could be carried out and used to generate alarms for abnormal patterns, and to track infiltration/exfiltration. An untapped source of sewage flow information is to utilise potable water flow data, most of which reappears in the sewer. Low-cost flow data is also available from pumping station controllers, which can provide flow data from the rate of rise or fall of the level in wet wells. By pooling all the data available, some valuable information about sewers could be generated.
The likely timing of these different sewer flow applications is shown in the timeline. While there will be schemes requiring sewer flow measurement up to 2009, most of the applications will be for level measurement and flow monitoring on rising mains. Beyond 2009 the WFD programme of measures (POMs) will make an impact and flow monitoring of gravity sewers is likely to have an important role in integrated catchment schemes. To meet this need, there have been some significant advances in open channel flow measurement since the 1980s.
By far the commonest sewer flow monitor is the ultrasonic mouse (see diagram). The first devices were developed in the late 1970s to meet the need of sewer hydraulic modellers for a low-cost survey instrument. It is a velocity area method, so measures volumetric flow by measuring flow velocity and deriving the cross sectional area of the flow from a measurement of liquid depth. The velocity is measured using an ultrasonic Doppler sensor. And in the original systems, the depth was measured by pressure transducer. Later developments have provided additional options of ultrasonic pulse echo within the flow, or air-fired ultrasonics to measure the depth and hence the cross section of flowing sewage.
The major uncertainty of the basic ultrasonic mouse has always been the estimation of the average velocity of the flow. In shallow flows, the strongest Doppler signals generally come from the liquid surface. But as the flow deepens, the ultrasonic Doppler signals can come from almost anywhere within the velocity profile. Under good practical conditions, an accuracy of ± 20% can be achieved and packages with a logger cost about £2,000. An important advance has been the development of ultrasonic sensors, which can measure the flow profile such as the Nivus OCM Pro active. This uses correlation techniques on the returned signals to infer the range to the source of the signal. By combining signals from several ranges, a flow profile can be fitted to the data and thus obtain a reliable estimate of the average velocity. This removes the major source of inaccuracy of the single channel device, and a big improvement in performance. An accuracy of 0.5% to 2% is claimed at a price of £5,000-6000 for a system.
Ultrasonic time of flight has also been applied to sewer flow measurement, either with the transducers held on rings in the sewer as with the Doppler mouse or with the transducers embedded in the walls of the sewer. Time of flight offers better estimation of average velocity than the single channel Doppler method and the Sarasota 200 using multipath time of flight quotes typical accuracies of 2-5%. There are also open-channel electromagnetic flowmeters, for example the ChannelMag from Flow Measurement, which offers velocity measurement accuracies of 5% at velocities of 0.6m/sec to 6% at 0.2m/sec. In these sensors, the field coils are embedded in the channel requiring significant engineering work. To date their use has mostly been restricted to sewage works sites where there are important reasons to avoid the head loss of a flow structure.
The workhorse of open channel flow on sewage treatment plants has long been standard flumes and weirs, with level measured by air-fired ultrasonics and the flow derived using standard relationships. There has been some recent work by WRc and Cranfield University to extend these methods to non-standard structures and the Water Industry Flow Club plans to complete this important task as part of its new programme. However, flow structures are not normally used in sewers.
The most interesting technical development of the past decade has been the Flodar from Marsh McBirney. It uses microwave Doppler from above the flow to measure surface velocity and radar to measure level. This is priced at about £10,000 for a claimed accuracy of 5%. It can be affected by rain if installed in the open, but this is not normally a problem for sewer flow.
There have been a number of attempts at applying optical cross correlation to sewer flow measurement going back to the late 1970s. More recently, WRc built a laser-based cross correlation flow sensor as part of a possible all-optical package to monitor polluting load in the Loadmon project. Good results were obtained at flow velocities above 0.2m/sec, but at low flows it was difficult to get a clear correlation peak and under certain conditions surface ripples could produce large errors.
In Germany, Diehl & Siebeneck have invested in a cross-correlation patent DE19820849 and may develop a commercial product. While optics are, on the face of it, the wrong technology for sewers, sensors mounted well away from the flowing surface could be engineered to be reliable. The accuracy of open channel systems is often quoted for a near-ideal channel. Real sewer conditions are often far from ideal and accuracy is consequently degraded.
Closed pipe flowmetering
On many sewer networks, there are opportunities to monitor flow at pumping stations using full-bore electromagnetic flow meters on the rising main and this will deliver accuracies of 0.5% to 1% for £1,500.
It is three years to 2009 when sewer flow monitors could have a crucial role in schemes under the Water Framework Directive Programme of Measures. User requirements for such a flow monitor would probably include:
- Very low maintenance, and this probably means non-contact; able to deliver 10% accuracy on a real life open channel and to maintain good performance at low flows
- Able to provide at least level data if the sewer surcharges priced below £4,000 with wireless communications on board; ATEX zone 1 and IP68
- Easy and therefore cheap to install
© Faversham House Group Ltd 2006. edie news articles may be copied or forwarded for individual use only. No other reproduction or distribution is permitted without prior written consent.