Modeling of a heat-integrated biomass downdraft gasifier: Estimating key model parameters using experimental data

• Published in: Energy conversion and management 2025-02, Vol.325, p.119372, Article 119372
• Main Authors: Haidar, Houda M., Butler, James W., Lotfi, Samira, Vo, Anh-Duong Dieu, Gogolek, Peter, McAuley, Kimberley
• Format: Article
• Language: English
• Subjects: Biomass gasification; Downdraft gasifier; Heat integration; Mathematical model; Parameter estimation
• ISSN: 0196-8904
• DOI: 10.1016/j.enconman.2024.119372

[Display omitted] •Subset selection is used to aid parameter estimation in a downdraft gasifier model.•Model parameters were ranked and 27 out of 40 were tuned for reliable predictions.•Parameter estimates result in 50% reduction in objective function from 1565 to 785.•Diagnosis of 13 unestimated parameters reveals correlation and low sensitivities.•Model is validated and used to study effects of bed height and producer-gas demand. Kinetic and transport parameters in a model of a heat-integrated biomass downdraft gasifier are poorly known and require estimation. The large number of parameters (40) arises from pyrolysis, combustion, and gasification reactions, as well as heat-transfer phenomena inside the gasifier and associated heat-integration system. Due to complexity of the model and the limited available data, only a subset of the parameters can be reliably estimated. A sensitivity-based approach is used to determine the appropriate number of parameters to estimate while preventing overfitting. It is hypothesized that estimating these important parameters will result in better model predictions. The 40 parameters are ranked from most-estimable to least-estimable based on sensitivity information and initial parameter uncertainties. A mean-squared-error criterion is then used to determine that 27 parameters should be estimated using data from 15 experimental runs, with the remaining 13 parameters fixed at their initial values. A diagnosis of the 13 low-ranked parameters reveals that 8 parameters are not estimated due to correlation with high-ranked parameters and that the remaining 5 parameters have little influence on model predictions. The model is validated using two runs not used for parameter tuning. The updated model is used to predict that a taller gasifier would not improve the quality of the producer gas. Simulations show that increasing the producer-gas demand by 50% results in a 15.2% decrease in H2/CO ratio, a 52.6% increase in tar content in the producer gas, and a 44% increase in electrical energy output.