Probabilistic Intraday Wastewater Treatment Plant Inflow Forecast Utilizing Rain Forecast Data and Sewer Network Sensor Data

Abstract

Forecasting of wastewater treatment plant inflow dynamics constitutes an enabler technology for wastewater treatment process optimization using model predictive control. However, accurate inflow prediction is still challenging, especially for strong rainfall events, where complex system dynamics and missing information on future rainfall represent limiting factors. We propose a seasonal probabilistic time series model for modelling the short-term wastewater inflow accurately while providing quantification of forecast uncertainty. To ensure suitability for practical implementation, the unconstrained parameters of the predictive distribution are modelled as linear functions of the input variables in the framework of Generalized Additive Models for Location Scale and Shape. Non-linear effects are approximated by Rectified Linear Units (ReLU), accounting for buffering within the sewer network and flow-dependent catchment response time. In addition to water level measurements from within the sewer network and rain rate measurements, rain forecasts are incorporated as exogenous regressors, where historical rain forecasts are used for model calibration. The model performance is evaluated on historical data from a German wastewater treatment plant using deterministic and probabilistic scoring rules. We benchmark against probabilistic time series models (SARX) and LSTMs. Our results show that the proposed model unites the benefits of high prediction accuracy of the LSTM and enhanced intelligibility of the SARX model, but accurate real-time rain forecasts are mandatory for successful real-world implementation.

Probabilistic Intraday Wastewater Treatment Plant Inflow Forecast Utilizing Rain Forecast Data and Sewer Network Sensor Data

“…The inflow, level measurement and rain measurement time series data used for model calibration and evaluation in the study are available at Zenodo via https://doi.org/10.5281/zenodo.69926944 with Creative Commons Attribution 4.0 International license (Sonnenschein & Ziel, 2022c). The time series data of spatially aggregated rain radar forecasts used for model calibration and evaluation in the study are available at Zenodo via https://doi.…”

“…The time series data of spatially aggregated rain radar forecasts used for model calibration and evaluation in the study are available at Zenodo via https://doi. org/10.5281/zenodo.6992694 with Creative Commons Attribution 4.0 International license (Sonnenschein & Ziel, 2022b). The model implementation and the code for the generation of the results, tables and figures is available at Zenodo via https://doi.org/10.5281/zenodo.6997701 with (Sonnenschein & Ziel, 2022a).…”

“…org/10.5281/zenodo.6992694 with Creative Commons Attribution 4.0 International license (Sonnenschein & Ziel, 2022b). The model implementation and the code for the generation of the results, tables and figures is available at Zenodo via https://doi.org/10.5281/zenodo.6997701 with (Sonnenschein & Ziel, 2022a). Rain radar raster data is available on the Open Data Server of Deutscher Wetterdienst via https://opendata.dwd.de/climate_environment/CDC/grids_germany/5_minutes/radolan/reproc/2017_002/bin/ (Wetterdienst, 2022).…”