Benchmarking surrogate models for activated sludge processes using BSM1/ASM-type simulations: performance and residual diagnostics for effluent ammonium, nitrate, and dissolved oxygen
Description
The uploaded dataset contains 100 sets of operational and effluent data generated from the BSM1 /ASM-type activated sludge process simulations. Each set records five key operational inputs (flow rate Q, readily biodegradable substrate XS, influent ammonium SNH_in, and aeration coefficients KLa3 and KLa4) and the corresponding three effluent indicators (ammonium SNH_out, nitrate SNO_out, and dissolved oxygen SO_out). These data are intended for constructing and validating machine-learning surrogate models, which can rapidly predict effluent performance of activated sludge systems under varying operational conditions, serving as a computationally efficient alternative to mechanistic simulations. The dataset spans typical operating ranges, capturing system responses under different loadings and aeration regimes, providing a foundation for rapid screening and sensitivity analysis.
Files
Institutions
- Heilongjiang UniversityHeilongjiang, Harbin