Parameters optimization using an artificial neural networks and release characteristics of humic acids during lignite bioconversion

The bioconversion of coal at ambient conditions is a promising technology for coal processing. However, there are few examples of the optimization of processes for industrial-scale use. In this work, the optimization of process parameters affecting lignite bioconversion by an isolated fungus WF8 using an artificial neural network (ANN) combined with a genetic algorithm (GA) was carried out for modeling of humic acids (HAs) yield and parameters. Kinetic models were used to understand the release characteristics of HAs from the bioconversion of lignite. The results of the present work indicate that the optimal process parameters (OPP) are 29 degrees C, initial pH of 7, 180 rpm, 0.6 mmol center dot L-1 of CuSO4, 0.4 mmol L-1 of MnSO4, and 6.4 mu mol center dot L-1 of veratryl alcohol (VA). The predicted experimental data obtained by ANN is similar to the actual and the significant correlation coefficient value (R-2) of 0.99 indicates that ANN has good predictability. The actual yield of HAs are 5.17 mg center dot mL(-1). During bioconversion, the fungus WF8 could loosen and attack the structure of lignite. The release of HAs produced by bioconversion of lignite under the OPP via diffusion and swelling is fit to zero-order model independent on concentration. This provides support for the industrial bioconversion of lignite.

Automated summary

This study optimized the process parameters of lignite bioconversion using ANN and GA, and the results showed that the optimal process parameters are 29 degrees C, initial pH of 7, 180 rpm, CuSO4 of 0.6 mmol·L-1, MnSO4 of 0.4 mmol·L-1, and VA of 6.4 μmol·L-1, with good predictability and support for industrial bioconversion of lignite.