The aim of this article is to present a new biological aerated filter (BAF) for nitrogen removal based on simultaneous nitrification and denitrification. Contrary to the systems which integrate both an aerated and a non-aerated zone to allow complete nitrogen removal in one compact or two different units (pre-denitrification and nitrification), this upflow BAF system is based on the principle of simultaneous nitrification and denitrification since the filter is completely aerated. The denitrification process is possible due to the diffusion effect which dominates biofilm processes. The real time aeration control allows us to maintain a low dissolved oxygen value (0.5 to 3 mg O2/l). In this case, the biofilm will not be fully (or less) penetrated with oxygen and denitrification will be carried out in a large part of the biofilm. Therefore, nitrification and denitrification is running simultaneously in different depths of the biofilm. By using 50% less air this BAF gave the same results (less than 20 mg TN/l) on pilot plant as a classical nitrification and denitrification BAF (Toettrup et al., 1994). Less recirculation was necessary to achieve the same denitrification.

Two new ways of controlling the aeration in nitrifying biofilters, by measuring on-line dissolved oxygen and ammonia concentration in the effluent of the filter, are compared to the standard control using time tables with constant air flow. Both on pilot and large scale, the new process control concepts proved to be very efficient in optimizing operation costs for aeration (up to 70% during rainy days) and to guarantee a demanded effluent quality.

The objective of this work is to provide an efficient method to control the dosing of chemicals in order to optimise the operation of the complete treatment train associating phosphorus precipitation in a lamellae separator followed by nitrogen removal in a Biostyr® filter with methanol addition. Four coagulation-flocculation tanks of 1.54 m3 total volume were used, followed by a lamellae primary tank of 0.65 m2 footprint and 5.4 m2 total settling surface. On-line measurements were used. The pilot was operated initially with no chemicals, at constant and variable flow. Different control strategies were sought. To help tune the controllers, step changes in chemicals dosing were made to identify the open-loop process transfer function at different flow rates. The system was then operated in closed loop. This proved to be an efficient means of operating the system. Future work will continue to test various control strategies.

No abstract