Testing the grey matter
Cranfield's selection of technologies for grey water recycling and reuse
Grey water recycling systems have been established worldwide, principally in multiple-occupancy sites such as hotels, apartments and other large municipal buildings. However, a major barrier to the implementation of such technology in the UK is the lack of tested and approved systems for grey water treatment. Water quality standards applicable to in-building reuse are the subject of much debate. Two schools of thought predominate - those who favour stringent standards similar to those of the USEPA or those who advocate standards closer to the EU bathing water directive.
There are no UK national water quality standards for effluent destined for reuse and this additional hurdle further hinders uptake of reuse technology. In light of such debate, the appropriate selection of technologies for specific reuse applications is of utmost importance.
The School of Water Sciences at Cranfield University has its own onsite grey water collection and test facility. This comprises specialised plumbing in 18 student flats to facilitate the collection of grey water from baths, showers and bathroom hand basins.
The grey water is pumped from a communal sump to holding tanks located inside a secure container where systems to be tested can be connected directly to the grey water feed. A secure outdoor area opposite provides a location for testing constructed wetlands and similar extensive treatment systems, with piped connection to the grey water feed. This unique test facility has the immense benefit of providing a real grey water source for research and allows the water recycling team at Cranfield to trial different technologies in order to determine the performance envelopes for each.
Technology design and selection criteria for reuse of both grey and black water are now becoming established. Treatment processes must not only be capable of generating effluent of the desired water quality but should also be robust enough to perform under sub-optimal conditions.
Grey water is inherently variable in composition, particularly with respect to organic strength and this, along with variable flow patterns, poses major difficulty when treating grey water for reuse. Variability and susceptibility to shock loads can become even more pronounced at smaller scales. For example, in an apartment block, the activities of just one household allowing entry of a noxious or unusual substance (for example, disinfectant or oil) to enter the grey water stream could disrupt biological activity in a communal grey water treatment system. Likewise, flow patterns are more likely to be influenced by holiday periods and diurnal activities at the smaller scales. As treatment scale decreases from regional scale, to small developments, to single households, maintaining treatment performance becomes more challenging and consequently technology choice becomes increasingly important. Studies at Cranfield have revealed variation in the organic strength of grey water from different sources ranging from BOD5 of 39 in mixed grey water collected from the onsite grey water collection facility to in excess of 150mg l-1 observed in fresh shower water from a single individual.
A survey of the literature reveals this range is representative of that reported in the literature for grey water collected from different sources. Routine analyses of raw grey water from Cranfield's test facility also reveals substantial fluctuations within that source. The ability of a technology to cope with such variation can be measured using robustness indices. In the context of water treatment for reuse, 'robustness' refers to the sensitivity with which effluent parameters change as a function of either input or operational variations.
The concept is analogous to that of vehicle suspension, which is designed to provide a smooth, safe ride by absorbing the impact from variable road conditions (Coffer et al., 1998). Robustness can be visualised as a flat horizontal line on a graph of output (performance) versus input (influent or operating parameter) - see Figure 1. The graph depicts two different 'robust' systems (one and two) and a non-robust system (solid line).
Performance of robust systems varies little with changing operating conditions and it can be considered as an inherently safe technology. Robust system one provides the same level of performance as the non-robust system and represents technology selection that provides inherent process safety. The second robust system has reduced performance capacity representing sub-optimal operation of the process to increase its robustness. A classic illustration of this is during coagulation treatment where chemicals are dosed at above optimal concentrations to reduce the risk of process failure. The concept of robustness can be used to generate robustness indices for individual technologies or processes, allowing comparison among different treatment systems.
This will ultimately lead to a better informed selection of technologies for reuse applications. Robustness indices can be based on any standard water quality standard parameter (for example, BOD5, turbidity, etc). The performance at two percentiles (50th and 95th) is measured and compared to a target concentration as shown in equation 1. The first term in the index represents the uniformity of the performance, while the second term relates to the overall performance (how well the process is doing).
Low values of the robustness index indicate a treatment process that is meeting the water quality standard with relatively low variation. A value of one means the standard is being met with no deviation and numbers down to a minimum of 0.5 indicate the achieved water quality is lower than the target. Conversely, large index numbers mean the treatment is either not achieving its goal or the variability is high.
The concept of robustness indices was applied to a study at Cranfield, in which the suitability of a submerged membrane bioreactor (MBR) and a biological aerated filter (BAF) for grey water treatment were investigated. Robustness indices, based on each of the standard water quality parameters, were determined for steady state operation and performance during intermittent operation (cessation of air and/or feed) was also monitored. The submerged MBR consisted of two perspex chambers with a working volume of 0.035m3 each, subdivided into denitrification and nitrification units. Two 0.24m2 flat plate polysulphone microfiltration membranes (Kubota Corporation, Japan), with a pore size of 9.4m2 were immersed into the nitrification unit of each bioreactor.
Flow through the MBR was hydraulically driven to maintain a head of 0.6m above the top of the membrane modules. The BAF was a 0.165m diameter vertical perspex column containing 1.64m depth of 3-4mm plastic media (Lytag).
The feed was pumped counter-currently to the compressed air, which was supplied at 15 l/min. Flux was maintained by regular air scouring back washes. Mean loading rates are shown in Table 1. During steady state operation the MBR consistently met the most rigorous (USEPA) water quality standards. Mean effluent values of 1.1-1.6mgL-1 BOD5, 0.32-0.28 NTU (turbidity) and non-detectable total coliform bacteria were recorded.
However, separate operation of the membrane without connection to the bioreactor unit resulted in significant coliform breakthrough. The capacity of the BAF to remove organics was similar to that of the MBR, but performance was poorer in terms of turbidity and coliforms with mean effluent parameters of 4.3-4.1 mgL-1 BOD5, 3.2-8.9 NTU (turbidity) and 2x104-5x104 CFU/100ml total coliforms.
The robustness indices calculated for the MBR were 2.0, 1.8 and less than 1 based on BOD5, turbidity and total coliforms respectively, while for the BAF they were 2.3, 1.8 and 19.8 for the same parameters. The large value for total coliforms indicated the inability of the BAF to meet the microbiological effluent standard. During unsteady state performance, the robustness of the MBR was again evident. Unavailability of air, feed or both for up to eight hours had no effect on effluent quality or hydraulic performance. Consequently there was no recovery time required upon restart. Even after 25-day cessation of feed, effluent quality remained below 5mg L-1.
This inherent robustness is achieved through a combination of the membrane acting as a total barrier to suspended particles (above 0.1m2) and the biological breakdown of organic material in the bioreactor. Short-term unavailability of grey water feed or air had no effect on BAF effluent BOD5 or turbidity but longer periods of intermittent operation led to increased concentrations of effluent parameters and prolonged recovery times.
Cessation of air produced the most rapid deterioration of effluent quality with turbidity reaching a maximum level of 30 NTU when air was switched off for more than an hour. This sensitivity to air flow occurs because aeration is required to ensure maximum contact between the grey water and the biofilm for biological treatment to take place.
In addition to providing robust treatment solutions for the waste stream in question, technology choice must also address a number of other important criteria including appropriate foot print size and aesthetic properties. Such factors will vary from case to case, depending on population size and the requirements of the stakeholders involved.
This highlights the need for the process design to be carefully tailored for individual situations. Consequently, a range of technologies are under going further trials for waste water reuse at Cranfield's grey water testing facility. A comparison of leading contender technologies for grey water reuse is currently under way as part of the £2.5M EPSRC/industry/regulator-funded Water Cycle Management for New Developments (WaND) consortium project. Technologies under test include a membrane bioreactor which will be used as a benchmark technology (already proven to be a robust solution for grey water treatment both during the above trials at Cranfield and also through monitoring of numerous installations worldwide, particularly in Japan where it appears to be the technology of choice for grey water recycling). A new experimental chemical system incorporating treatment with titanium dioxide and UV radiation will also be tested and at the low-tech end of the scale, testing of reed beds running in vertical and horizontal modes is already under way. Data collected from the trials will be used to calculate robustness indices for each system and will establish the suitability of different types of treatment system for various re-use applications, site conditions and stakeholder requirements. Also under trial is the Green Roof Water Recycling system (GROW) designed by Water Works UK.
GROW is essentially a garden of low growing, flowering, native plants based on a stepped arrangement of shallow troughs. The troughs are designed to be placed on sloped roof tops in urban environments where space is scarce. Expanded clay and stone chipping media along with plant roots provide a surface for biological degradation of grey water by similar mechanisms to those employed in other constructed wetland systems. It has been Cranfield's experience and is indeed the consensus worldwide that biological treatments tend to be most appropriate for the generation of high-quality effluent for reuse. Biological treatment is particularly efficient at removing dissolved organic material from the wastewater - a step that is vital in order to reduce the disinfection demand of treated effluent and to prevent microbial re-growth. The configuration of the biological system will determine how robustly the process will meet standards.
Barrier processes such as the MBR provide high standards of robustness. Concerns about process failure deter developers, planners and other stakeholders from incorporating reuse technology as part of a sustainable water management strategy. Applying the concept of robustness and developing robustness indices for leading contender technologies for grey water treatment is a leap towards making useful comparisons and ultimately informed technology choices. If we want to see grey water recycling take off in the UK, measuring process robustness must be placed high on the agendaReferences: Coffer BM, Lang S, Green JF and Huck PM (1998). Quantifying performance and robustness of filters during non-steady state and perturbed conditions. IN: Proc. AWWA water quality technology conference, Denver, AWWA. Huck PM and Coffer BM (2002). Robust drinking water treatment for microbial pathogens - implications for Cryptosporidium. IN H H Hahn, E Hoffmann and H Odegaard (eds.) Chemical water and wastewater treatment VII, Proc. 10th Gothenburg Symposium, June 17-19, 2002, Gothenburg, Sweden, 165-182.