Enhanced Robustness in Model Predictive Control of Water Quality Using Adaptive Estimation Methods
| Main Authors: | Eliades, Demetrios G., Kyriakides, Evripides C., Polycarpou, Marios M. |
|---|---|
| Format: | Proceeding Journal |
| Terbitan: |
, 2007
|
| Subjects: | |
| Online Access: |
https://zenodo.org/record/4306226 |
Daftar Isi:
- In this paper, the issue of designing algorithms for controlling the spatial-temporal distribution of chlorine residual in drinking water distribution networks is examined. The network dynamics are influenced by unknown time-varying water demands. A solution methodology is presented using model predictive control principles combined with on-line adaptive forecasting of the system hydraulics as they are driven by consumer demands. The periodic nature of water demands allows the use of Fourier series and Radial Basis Functions with adaptive coefficients for forecasting future demand patterns in the distribution network. The feedback control objective is to maintain chlorine residuals, which acts as a disinfectant for enhanced water quality, within certain lower and upper bounds. Simulation results on a real distribution network are used to illustrate the performance of the proposed feedback control algorithm under unknown water demands and the trade-offs involved in the robust model predictive control methodology.