Exploring the cost of failure
Mark Smith and John Wood of WRc report on a recent project looking at reducing costs at unmanned WwTWs and improving procurement of low-cost monitoring technology
Treatment processes at small works are dominated by biological filters providing secondary treatment, but other processes including reed beds, activated sludge and septic tanks are also present. Since the 1960s increasing use has also been made of small packaged plants rather than custom-built works.
Operational costs are incurred by water undertakings for small, unmanned works in order to keep them performing effectively and avoiding the costs of ‘failure’.
In an age of efficiency and financial limits, methods of reducing operational expenditure are welcome. A failure at a works can arise from various causes including mechanical failure (e.g. pump breakdown), power failure, operator error (e.g. return liquors), blockage. Whatever the source, failure has significant financial implications for the undertaker, both internal (e.g. extra manpower costs for correcting a problem and cleaning up) and external (e.g. complaints, damage to the local environment, bad publicity and fines for breach of consent). Costs in excess of £50,000pa are not unusual for failures at small works. Clearly a balance needs to be maintained between the regular operational costs of avoiding a failure and the uncertain potential costs of the failure itself and achieving this balance is a business decision of the undertaking. Risk analysis is a process for identifying and quantifying the risks associated with operating and owning assets and there are many techniques available to assess risks and optimise costs. A typical approach to risk analysis is to identify:
- failure cause,
- failure consequence,
- failure frequency,
- failure costs.
Consideration of the results of such an analysis in relation to routine operational costs helps to make a judgement on the level of investment that is justifiable on failure prevention. This approach was used by WRc in a recent project examining the use of low-cost monitoring at unmanned works to reduce operational costs whilst still minimising the risk of failure.
Break from routine
The technology for monitoring and communication is developing rapidly, such that it is now possible to monitor works functionality at relative low-cost. Using this technology opens up the possibility of reducing the routine operating costs whilst maintaining the risk of failure, or maintaining the expenditure whilst reducing the risk, or some combination of both. The former is achieved through decreasing the frequency of routine visits to many, often very disperse, works and the latter by better information leading to an overall decrease in the number of failures as a result of earlier detection of potential problems resulting in reduced costs of dealing with them. More cost-effective operational expenditure is the goal.
The installation and operation of the instrumentation itself is obviously an increased cost which needs to be justified. The risk analysis confirmed that for the benefits of sensors to be worthwhile at small WwTWs it is essential the cost of the sensors themselves, their operation and costs of data transmittal must be low. Simultaneously, sensor reliability must be high and sensor lives long. Given the intrinsic nature of sewage and the severe demands it can place on equipment, non-contact sensors offer the most feasible approach, monitoring the following items:
- movement or motion,
Initially WRc tested three sensors which have potential application in monitoring aspects of the operation of processes at WwTWs - optical range sensors, microwave movement/flow sensors and rotation sensors. The sensors do not necessarily monitor the processes themselves, but aspects that show the performance is occurring and/or is likely to be satisfactory.
Optical range sensors are already integrated into commercial applications such as hand driers and work on the principle of the movement in the position of the detected (reflected) image from the target. The sensors do not work well on clear water, but could have useful application on settled or crude sewage and sludge. Power requirements are compatible with long-term battery-powered use and the sensor costs under £20. A successful, short field trial was carried out sensing levels of settled sewage to monitor siphon performance.
A standard microwave security sensor was tested in the laboratory to see whether it could be used to detect flow and movement. It proved possible to detect clean water flows down to 0.1m/s surface velocity without difficulty, but a difficulty was encountered in excluding movement from other sources. Security sensors are small, robust and available packaged for outdoor use as well as being cheap (circa £30). They are compatible with battery-powered use and would be particularly useful in a flow/no flow situation such as a storm overflow channel or RAS system.
Rotation sensors have particular application at biological filter works where this is the single most important functional check on process operation. Two broad classes of sensor are available - the proximity switch approach and others. The proximity switch approach is an established method which involves a fixed sensor detecting a rotating object attached to one of the filter arms. Possible alternatives are a gyroscopic sensor designed to monitor rotational speed and a magnetic field sensor using the earth’s magnetic field.
It proved possible to detect rotation using a low-cost (<£50) sensor, but sensor zero drift was found to be too great in relation to the typical rotation speeds anticipated for filter arms. More expensive sensors may not have this problem.
A magnetoresistive sensor was tested to explore if it could reliably measure rotation using the change in the earth’s magnetic field as the filter arms turned. The sensor costs below £20, is small and easily built into a compact module. However, care must be taken to locate the sensor outside iron masses (e.g. the central column of the filter assembly) which shields it from the earth’s magnetic field.
Low-power radio telemetry systems seem to offer a feasible methodology for transmitting the sensed data to a control room. The optimum design of a telemetry system depends on a number of factors including:
- number of sensors and their separation,
- the access requirements to the data,
- data download frequency,
- the site size and topography,
- availability of telephone links (e.g. landline, cellular coverage),
- power supply potential,
- housings for the equipment.
A potential difficulty is in integrating sensors, communications and appropriate software, as no one supplier currently provides all of these facilities.
In this initial project WRc has shown there clearly is potential for installing and operating low-cost monitoring schemes at small unmanned WwTWs. At the lower end of the small size range, worthwhile cost savings arise from reducing the number of failures rather than reducing the number of site visits. Potential works where savings are possible are those where the existing site visits consist mainly of checking rather than taking remedial action. Monitoring can then replace site visits.
Monitoring a known problem (e.g. blocking of siphons) also results in more cost-effective site visits and avoiding failures.
Care needs to be taken to design an optimum monitoring and transmission scheme and currently no readily available packages exist. Detailed work needs to be done on the practical integration of sensors and low-cost radio systems.
Typical costs for basic monitoring of a small WwTW have been estimated in this project as between £500-£1,600pa assuming a seven-year asset life for monitoring equipment. Potential cost savings on current operations are in the range of £500-£8,000pa. These costs need to be set in the context of avoiding a consent failure.
The project is continuing into a pilot study of a selected site to consider some of the issues raised in the initial study in more detail, to confirm the validity of the approach. These issues include integrated monitoring packages, lack of data handling software and other possible sensors.