A receding horizon data-driven based control for short term air quality management

In this work, the implementation and test of a tool for the definition of air pollution optimal short term control to support the definition of policies by Local Authorities is presented. The methodology is based on a receding horizon approach where an autoregressive model provides information about the air quality dynamic in the selected time horizon. The model has been identified starting from measured data (concentration and meteorological variables) and estimated information (emission levels) over the area under study. Every time step, the resulting optimization problem has been solved through a genetic algorithm. The system has been tested for the control of NO2 concentrations in the municipality of Milan. The results show that the control system can be a valuable asset to aid Local Authorities in the selection of suitable air quality plans.