Dynamical modelling of a waste stabilisation pond

This paper is concerned with the dynamical modelling and the parameter identification of a waste stabilisation pond. First, a dynamical model of the pond is proposed, based on mass balances in the first basin. It involves a reaction network involving eight (bio)chemical reactions, and in particular the (chemical or biochemical) oxidation of H₂S. The height of the pond is divided into two layers: the upper layer (approximate depth: 0.8 m), and the lower layer (about 0.2 m). Three microorganism populations are considered: microalgae and aerobic bacteria (in the upper layer), and sulphate-reducing anaerobic bacteria (in the lower layer). The Droop model is introduced to emphasise the potential activity of microalgae when daylight has disappeared (sunset). The transport of organic matter between the two layers is also considered in the model. The derivation is based on collected data and intensive follow-up of a specific pond at the village of Rethondes in Northern France. The parameters of the model are then identified on the basis of these data by considering data in spring, summer and autumn. The calibration of the model parameters is a challenging problem because of the large number of parameters, the limited number of available experimental data and the model complexity. The objective in the identification procedure was thus limited to obtain the largest number of unique values for the parameters in the three instances.