Date of Graduation


Document Type


Degree Name

Doctor of Philosophy in Engineering (PhD)

Degree Level



Biological and Agricultural Engineering


Jun Zhu

Committee Member

Wen Zhang,

Second Committee Member

Thomas A. Costello

Third Committee Member

Yi Liang


Anaerobic sequencing batch reactor;Artificial neural network;Biochar addition;Methane yield;Micro-aeration;Response surface methodology


Anaerobic co-digestion (Co-AD) of poultry litter (PL) and wheat straw (WS) for sustainable waste management and simultaneously renewable bioenergy generation was improved by advanced technologies including micro-aeration (MA), biochar addition (BC), nanoparticles supplementation (NP) and their integrations. Response surface methodology (RSM) and artificial neural network (ANN) were used for the process modeling and optimization. First, a method by gradually increasing OLR was developed for the start-up of Co-AD of PL and WS in an anaerobic sequencing batch reactor (ASBR), with a continuous daily biogas production of (13.06 ± 0.21) L and methane content (MC) of (54.38 ± 0.53) % observed. The overall microbial community became more uniform, and the dominant aceticlastic methanogens of Methanosaeta were enriched. Based on the first experiment, Co-AD of PL and WS was continuously performed in the ASBR under different operating parameters including carbon to nitrogen (C/N) ratio, total solids level (TS, %), and hydraulic retention time (HRT, day) by central composite design. The RSM models (R2 = 0.9554, RMSE = 0.813; R2 = 0.9618, RMSE = 4.940) were more accurate than the ANN models (R2 = 0.9163, RMSE = 1.114; R2 = 0.9037, RMSE = 7.847) in predicting MC (%) and daily methane yield (DMY, mL CH4/g VS added), respectively. The optimal conditions for maximum DMY obtained by RSM were C/N ratio of 22.73, TS of 2.27 %, and HRT of 11.45 days, under which the validation trials showed MC of (58.37 ± 0.25) % and DMY of (184.36 ± 0.51) mL CH4/g VS added. The results could provide support for the operation and prediction of the continuous Co-AD of PL with WS for real bioenergy production applications. The third experiment investigates the micro-aeration (MA) strategy for Co-AD of PL with WS. Batch Co-AD of WS and PL with daily MA obtained an improved (15.1%) CMY of 225.44 mL CH4/g VS added. Daily MA shortened the lag phase of Co-AD by 3.4%. The sequencing batch reactor for the Co-AD of WS and PL showed an increased (21.5%) daily methane yield when 0.5-h/d MA was employed. The results provided support for the application of MA in the AD of agricultural wastes. The fourth study employed RSM and ANN coupled with genetic algorithm (GA) for the process modeling and optimization of batch Co-AD of PL and WS with biochar addition. Biochar addition increased the observed CMY (mL CH4/g VS added) by 10.07 % and increased k (day-1) by 12.5 %. RSM was used to model the response of CMY, with R2 =0.9981 and RMSE=0.91. The optimal conditions by RSM were C/N ratio = 28.54, TS = 8.08 %, and Biochar = 9.93 % TS, under which the maximum CMY was 284.61 mL CH4/g VS added. The trained ANN (3-3-1) for predicting the output of CMY was slightly less accurate (R2 = 0.9926, RMSE = 1.80). But the ANN-GA generated a 2.03 % higher maximum CMY of 290.39 mL CH4/g VS added under the optimal conditions of C/N ratio of 24.46, TS level of 5.03 %, and Biochar of 8.73 % TS, compared to RSM. The results could provide support for the process simulation of Co-AD of PL and WS with biochar addition in applications. Lastly, the fifth part investigated different strategies (MA, BC, Fe3O4 NP (FNP)) and their integrations for improving the Co-AD of PL and WS. The SMY was increased most in BC + FNP by 18.4 %, with a value of (324.8 ± 3.8) mL CH4/g VS added. BC + FNP also showed the most improved removal of substrate TS and TVS by 38.6 % and 24.2 %, respectively. Based on the comparison results, BC + FNP was recommended as the best strategy to improve the methane yield efficiency of Co-AD of PL and WS.

Available for download on Saturday, August 30, 2025