A methodology to predict the pollutant loads in combined sewer flow.

In the design of urban drainage systems, synthetic design storms are commonly
used to predict the peak flow rate in sewer systems and such storms are usually
based on local intensity-duration frequency curves or design storm profiles. To
estimate the quality of storm flow, the UPM Manual (1994) has highlighted the
development of detailed and sophisticated simulation models to estimate the
pollutographs, that is, the temporal variation in the concentration of pollutants in
urban drainage systems. The data requirements of these models are quite onerous,
and as a consequences implified models like SIMPOL have been developed. This
model predicts the BOD at 1 hour time intervals and is based on the
representation of the sewer system by a series of tanks. This approach may be
considered satisfactory for the prediction of accumulative pollution over an annual
series of events but for the prediction of acute effects, for example, the first foul
flush, the temporal variation in the concentration pollutants in sewer flow is
required. There is a need therefore to describe the change in pollution over a
much smaller time interval than that proposed in SIMPOL and this is particularly
so when consideration is given to the comparison of the design and control
options which may be proposed, for example, the real time control of storage
tanks to retain the first flush of pollutants.
The work outlined in this thesis presents an alternate simple methodology to
estimate the pollutographs corresponding to a particular storm event. The work is
based on the results of the measured pollutographs recorded on the WRc sewer
quality archive (1987) from two catchments at Great Harwood and Clayton-le-
Moors in the North West of England. The relationships for the shape of the
pollutograph were obtained by the direct comparison of the observed
pollutographs. The peak TSS concentrations were obtained by a detailed
regression analysis of the observed peak TSS concentrations, the antecedent dry
weather period and the hydrological parameters of maximum rainfall intensity,
average rainfall intensity and storm duration. These parameters were then related
to the shape of the pollutograph and the results of this methodology were shown
to satisfactorily reproduce results for the catchments considered. For practical
applications, the suggested procedure provides a methodology to calibrate the
design pollutographs for any catchment from a limited number of monitored storm events and to utilise these together with time series storms to assist in the
performancea ssessmenat nd selectiono f alternative design options. The work has
the limitation that it is catchment specific but as more information for different
catchments becomes available, it may be possible to establish standard
pollutographs for application to a wide range of catchment conditions.