Rational basis for optimal design of sequencing batch reactors with multiple anoxic filling for nitrogen removal N. Artan * , R. Tasli, D. Orhon Environmental Engineering Department, Istanbul Technical University, ITU Ins ¸aat Fakultesi, 34469 Maslak, Istanbul, Turkey Received 20 May 2005; received in revised form 31 October 2005; accepted 15 November 2005 Abstract The paper defines the rational framework for understanding the mechanism of nitrogen removal in SBRs with multiple anoxic fill phases. Mass balance between denitrification potential and nitrate nitrogen introduced in each anoxic period sets the basis for optimal design. System evaluation shows that additional anoxic filling primarily avoids oversized reactor volumes as it reduces the internal recycle requirement. Low effluent nitrate levels can be achieved with dual anoxic systems with optimal hydraulic retention times. Higher denitrification potentials obtained by increasing the duration of the total anoxic period within the process phase (higher T DN /T P ) can best be utilized by introducing multiple anoxic filling. Such systems may be optimized with an intermittent aeration continuous filling mode, which may be operated either with denitrification potential or available nitrate limitation. Model simulation of SBR behaviour also supports the results of optimal system design derived from process stoichiometry. # 2005 Elsevier Ltd. All rights reserved. Keywords: Activated sludge; Multiple anoxic filling; Nitrogen removal; Pre-denitrification; Sequencing batch reactor 1. Introduction The most striking feature of the sequencing batch reactor (SBR) is the system flexibility offered with a very simple physical structure. The operation alternatives inherently associated SBR with has been the basis of its promotion [1]. This flexibility of operation, if well understood and interpreted in terms of governing biochemical process may prove very useful. The past experience however has developed as a blindfolded exercise of trial and error where different operation options have been experimentally tested without so much emphasis on process kinetics and stoichiometry. When SBR technology is selected for nitrogen removal, the system flexibility is seriously challenged as internal recycle cannot be handled and controlled separately from sludge recycle and this restriction: (i) may negatively affect the reactor volume and therefore the hydraulic retention time of the system; and (ii) does not allow full denitrification using dual anoxic periods such as the 4-stageBardenpho process. This is a major limitation for the SBR compared to continuous- flow activated sludge, which offers independent control of internal recycle. One of the remedial actions to overcome this limitation has been the introduction and testing of SBR operation with multiple anoxic periods: Oles and Wilderer [2] operated an SBR system with a two and three anoxic filling and simulated the experimental results obtained by means of ASM1 model. Andreottola et al. [3] provided the simulation results based on ASM1 of an SBR with three sub-cycles (three anoxic periods) and a single short duration filling at the beginning (dump filling); they optimized the ratio of aerobic/ anoxic periods for a selected effluent total N level. Andreottola et al. [4] reported the results of system optimization through on line monitoring and control of pH, ORP and DO for an SBR operated with triple anoxic filling phases of equal duration. Tilche et al. [5] tested an SBR with five subcycles, each with 2 h anoxic and 2 h aerobic periods for the treatment of piggery wastewater with a very high nitrogen content. Puig et al. [6] also described a successful implementation in the SBRs using 4–6 filling events with a sequence of anoxic-aerobic phases treating synthetic and urban wastewater. The information provided through these studies, although useful, is empirical and specific for selected cases. Only a recent study gives the mechanistic basis for the selection of additional anoxic periods for a SBR system treating a strong wastewater, where the available nitrogen for denitrification is limiting [7]. www.elsevier.com/locate/procbio Process Biochemistry 41 (2006) 901–908 * Corresponding author. E-mail address: nartan@ins.itu.edu.tr (N. Artan). 1359-5113/$ – see front matter # 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.procbio.2005.11.009