Data-driven dam water level forecasting and intake optimisation models for proactive water treatment management
MetadataShow full item record
This paper describes a comprehensive raw water intake selection decision support tool which has been developed for a dual source drinking water treatment plant (WTP) in South-East Queensland, Australia. The WTP receives water from small Little Nerang (LND) dam by gravity and from the upper intake of Hinze dam (HUI) via electrical pumps. The core part of the optimisation model can predict treatment chemical dosages and costs given the quality of the raw water source, through a number of data-driven, chemical and mathematical models. This prediction tool was run over an historical data set and it was found that an optimal intake selection would imply an increased use of the LND source. However, WTP operators typically minimise the LND usage due to its limited storage capacity and thus high depletion risk in case of intensive withdrawal rates and unfavourable weather conditions (i.e. no rain). As a consequence, a probabilistic data-driven model was also developed, which predicts, 6 weeks ahead, the most likely volume of LND. The model is based on historical data correlations and takes as inputs the seasonal streamflow forecast from the Bureau of Meteorology, as well as prescribed raw water intake volumes. A Monte-Carlo approach is used to account for uncertainty. Hence, whenever LND is selected as the optimal source, WTP decision-makers can select the optimal intake volume in order to minimise treatment costs, but also the risks of wasteful dam releases or depletion of LND.
Proceedings of the 8th International Congress on Environmental Modelling and Software: Supporting Sustainable Futures
Copyright remains with the author(s) 2016. The attached file is reproduced here in accordance with the copyright policy of the publisher. For information about this conference please refer to the conference’s website or contact the author(s).
Water Resources Engineering