Operational reliability of urban drainage systems under uncertainties

Water quality risks from overflows have attracted significant research attention, and the reliability of urban drainage systems (UDS) is in urgent need of assessment and improvement. The overflow volume and concentration of critical pollutants are generally used as assessment indicators, which is quite time consuming and cumbersome especially under continuous rainfall. Simplifying the water quality risk assessment indicators for the UDS reliability is intractable. For this purpose, this study proposes the detention tank emptying time as a new reliability evaluation indicator, which greatly reduces the calculation burden by converting water quality risk into hydraulic risk. On this basis, the effects of rainfall, dry weather flow (DWF), actuators and their interactions on reliability are quantified by massive scenarios. It shows that the DWF affects the emptying process via weekly and daily seasonality and its interaction with rainfall is mainly responsible for unreliability. Further, the engineering facility linkage controlled by the actuator to cope with the interaction is the key. Particularly, the Prophet algorithm is innovatively applied to mine the patterns and generate the DWF series for the challenge of sparse DWF data. In conclusion, the indicator proposed expands the connotation of UDS reliability assessment, prompting a small investment in replacing actuators with better controllability and greatly improving reliability. It guides the engineering planning and enhancement from a new perspective of whole-chain optimization from the global to the detailed level.