Performance evaluation of the ISMLR package for predicting the next day's influent wastewater flowrate at Kirie WRP.

摘要: Soft-sensor applications for wastewater management can provide valuable information intelligent monitoring and process control above beyond what is available from conventional hard sensors laboratory measurements. To realize these benefits, it important to know how manage gaps in the data time series, which could result failure of sensors, errors measurements, or low-frequency schedules. A robust soft-sensor system needs include a plan address missing efficiently select variable(s) make most use information. In this study, we developed applied an enhanced iterated stepwise multiple linear regression (ISMLR) method through MATLAB-based package predict next day's influent flowrate at Kirie water reclamation plant (WRP). The increased retention 77% 93% achieved adjusted R2 up 0.83 by integrating with typical artificial neural network.