BACKGROUND The study aimed to set forth a model simulation approach for an accurate calculation of denitrification potential (N-DP) and selection of appropriate anoxic volume (V-D) in biological nitrogen removal. Simulations utilized the characteristics of a typical sewage and the model coefficients associated with a model structure that included all relevant microbial processes. They were carried out first for a fully anoxic reactor sustained with excess nitrogen and then for a pre-denitrification system operated with different anoxic volume ratios (V-D/V-T) of 0.1-0.4 when the influent nitrogen (TKN) level was set as 57 and 85 mg N L-1. RESULTS Without nitrate limitation, generated N-DP could reach 37 mg N L-1 for V-D/V-T = 0.1 and gradually escalated to 56, 75 and 86 mg N L-1 when V-D/V-T was raised to 0.4. Results for the pre-denitrification system showed the effect of nitrate limitation on both substrate removal and N-DP levels: At V-D/V-T = 0.4, N-DP associated with TKN = 85 mg N L-1 was approximately 60 mg N L-1, while it remained basically around 40 mg N L-1 with TKN = 57 mg N L-1. Simulation also indicated that the hydrolysis kinetics significantly affected N-DP generation. CONCLUSION Simulation results revealed that stoichiometric estimations always underestimated N-DP, recommending unnecessarily high anoxic volumes, where the oxidized nitrogen was consumed within the 20-25% portion of the provided volume. The remaining volume fraction was then totally wasted since it was forced to operate under anaerobic conditions, with practically no useful microbial activity and negligible substrate utilization. (c) 2019 Society of Chemical Industry