BibTeX21C

@Article{Casallas2021,

author = {Casallas, Miguel and Soto-Paz, Jonathan and Alfonso-Morales, Wilfredo and Oviedo-Oca\~{n}a, Edgar R. and Komilis, Dimitrios},

title = {Optimization of Operational Parameters during Anaerobic Co-digestion of Food and Garden Waste},

journal = {Environmental Processes},

year = {2021},

month = mar,

issn = {2198-7505},

pages = {769--791}

abstract = {Anaerobic co-digestion (AcoD) is a solution to recover renewable energy, such as methane, from food waste (FW) and garden waste (GW). Methane yield efficiency can be improved through the optimization of the operating conditions, such as mixing ratio of substrates (MRS) and Mixing Ratio of Inocula (MRI). The objective was to optimize the AcoD of FW with GW by combining Artificial Neural Networks (ANN) and Particle Swarm Algorithm (PSO). The AcoD of FW and GW were initially evaluated experimentally at a laboratory scale through a central design composed of two factors, at three levels each: FW:GW MRS at 80:20, 70:30 and 60:40 (w/w) and MRI with sludge mixture Granular (GSL) and Flocculant sludge (FSL) equal to 10:90, 30:70 and 50:50 (GSL:FSL, v/v). The response variables were the biochemical methane potential (BMP), bicarbonate alkalinity (BA), volatile fatty acids (VFA), hydrolysis (kh) and process stability index (If). The optimization identified the best operational conditions, which were later validated with a second experiment. Results of ANN and PSO showed that the maximized methane yield (270 mL CH4/g VS) can occur at the ratios of MRS 64:36 (w/w) and at MRI 44:56 (v/v) that increase the methane yield by 26% compared to mono-digestion of FW (70 mL CH4/g VS), while BA of 1645 mg L−1, 1955 mg L−1 of VFA, kh of 0.32d−1 and a stability (If) of −59 are achieved. The second experiment showed the robustness and applicability of the optimization tools, which resulted in a production of 265 mL CH4/g VS as a maximum yield.Highlights• Simultaneous optimization of substrate and inoculum ratios increase methane yields• A mixing ratio of granular:flocculent sludge (44:56%) increased methane yields by 26%• The optimal neural network topology was 2-4-1 for predicting response parameters},

doi = {10.1007/s40710-021-00506-2},

url = {https://doi.org/10.1007/s40710-021-00506-2},

}