Retour d'expérience des filtres d'apatite granulée en taille réelle pour la rétention du phosphore dans les filtres plantés de roseaux | Full scale experience of granulated apatite filters for phosphorous retention in treatment wetlands.

There are 20 full-scale Phosclean® filters for P removal being used together with treatment wetlands (TW) to treat wastewaters of small and medium-size communities, in France. Some of them are in operation for more than five years now, reason why some field experiences are being conducted under the project name of APPROVE (APatite for P Removal and Valorisation: an Evaluation) to assess the current performance and saturation conditions of full-scale filters. This paper will present fieldfeedback results from 4 different wastewater treatment plants (WWTP) with Phosclean® filters. WWTP A (filter A) has been monitored since its commission in 2017 to assess the effluent's pH evolution due to material dissolution in the very first stage of its life span. Three other WWTPs (B, C and D) with hydraulically different Phosclean® filters, in operation since 2012, are also monitored: vertical up-flow (filter B); vertical down-flow (filter C) and horizontal flow (filter D). Assessment of the treatment plants use regulatory surveys since the beginning of plants operations. As treatment plants have not the same capacity, the number of regulatory surveys is heterogeneous and not high enough to have a precise knowledge on saturation levels and retention kinetics. Consequently, additional measurement campaigns have been carried out over two years. Four-day campaigns have been carried out twice per year for each WWTP. In order to evaluate the P retention's performance, 24h-flow proportional samplings at the inlet and outlet of the Phosclean® filter were coupled to spotsamplings at different hydraulic retention times inside the filter to measure retention's kinetic. Online P-PO4 measurements at the outlet of the filter were also conducted to observe possible dynamic within a day. Mayor parameters as pH, conductivity, redox potential, TSS, BOD, COD, N, TP, P-PO4 3- as well as anions and cations were measured for these samples. For filter A, a flowmeter and an online pH probe at the outlet of the filter have been implemented for several months. To determine the nature of phosphate precipitates, SEM and DRX analyses were performed for solid samples taken at different points in the filters. Furthermore, fluorescein-tracer tests were carried out to precisely determine water retention times and possible short-circuiting, as well as dynamic penetrometer measurements to assess potential changes on the material density due to surface precipitation. Results of first campaigns show that P concentration at the outlet of filters C and D respect the maximal regulation limit for total phosphorous disposal after 5 years of operation. This is not the case for filter B, after 4 years of operation, which is over the regulation limit fixed at 2.5mgTP/L. Nevertheless, differences in the regulation limits, inlet P concentrations, hydraulic loads, operation and design, must be taken into consideration to analyze variations of filters performances. The evolution of TP concentrations along time will be presented for all filters according to saturation levels for comparison. Kinetics evolution as a function of saturation conditions will be also presented. Results show that filters B, C and D are at different saturation levels, being filter B the most saturated.