Anaerobic digestion (AD) is a commonly adopted technology to treat liquid dairy manure. However, an increase in ammonia nitrogen (NH3-N) concentrations in anaerobically digested liquid dairy manure (DLDM) has been observed due to the conversion of organic nitrogen to NH3-N during AD. The removal and recovery of NH3-N from DLDM has therefore gained increased attention among researchers to ensure environmental sustainability along with making better uses of manure N fertilizer values. In this study, 150 mL of DLDM was treated under different operating conditions using a modified domestic microwave and the key operational parameters viz. pH, irradiation time (min), and microwave power output (W) were optimized and modeled via response surface methodology (RSM) and RSM-artificial neural network (ANN). The role of operational parameters on the performance of the treatment process was first evaluated using RSM. Afterward, the RSM was combined with ANN to predict the NH3-N removal from DLDM. Although all the operational parameters at a linear level along with the interactive effect of irradiation time, microwave power output, and pH at a quadratic level significantly influenced NH3-N removal from the waste stream (p < 0.001). The results also revealed that pH acted as the primary factor affecting NH3-N removal during the microwave irradiated treatment process. Under the optimal conditions (pH: 10.8, irradiation time: 4.1 min, and microwave power output: 625 W), the NH3-N removal efficiency of 86.71 ± 1.85% was observed and found to be in line with the computational values from both the RSM and RSM-ANN models, while the recovery efficiency was found to be 77.98 ± 3.71%. Furthermore, a statistical comparison between the models demonstrated the RSM-ANN model has a greater prediction capability than the RSM model. The high NH3-N removal efficiency obtained under the optimal conditions indicates the effectiveness of the process and will further encourage the application of microwave irradiation to treat DLDM in pilot and full-scale processes.
Read full abstract