Volatile fatty acid (VFA)-rich leachate generated from acidogenesis of kitchen waste in a leach bed reactor (LBR) was utilized in an earthen microbial fuel cell (EMFC) to generate electricity. Effects of organic loading rate (OLR, 5–10 g VS/L·day) and pH (5–7) on LBR enumerated optimized parameters of OLR (10 g VS/L·day) and pH (5.74) to obtain total VFA (TVFA) of 7.7 ± 0.3 g/L in the leachate, with maximum contribution from acetic acid. Leachate obtained from the LBR was fed to the EMFC with varying OLR (2–7 kg COD/m³·day). The highest power density of 0.76 W/m³ (at OLR 7 kg COD/m³·day) was obtained with higher VFA content in the leachate. A neural network based on the Levenberg–Marquard function effectively predicted chemical oxygen demand and TVFA removal. This study establishes LBR as a techno-economic method to obtain useful substrate for EMFC. Furthermore, the response modelling of EMFC demonstrates the potential of utilizing machine learning in biological treatment.

中文翻译：

---

用于微生物燃料电池的渗滤液床反应器中挥发性脂肪酸生成的优化和建模

浸出床反应器 (LBR) 中厨房垃圾酸化产生的富含挥发性脂肪酸 (VFA) 的浸出液被用于土质微生物燃料电池 (EMFC) 中发电。有机负荷率（OLR，5–10 g VS/L·天）和 pH（5–7）对 LBR 的影响列举了获得总 VFA 的 OLR（10 g VS/L·天）和 pH（5.74）的优化参数渗滤液中的 TVFA 为 7.7 ± 0.3 g/L，其中乙酸的贡献最大。^{从 LBR 获得的渗滤液以不同的 OLR（2-7 kg COD/m 3} ·天）送入 EMFC 。最高功率密度0.76 W/m ³ (OLR 7 kg COD/m ³·天）的渗滤液中VFA含量较高。基于 Levenberg-Marquard 函数的神经网络有效预测化学需氧量和 TVFA 去除。本研究将 LBR 确立为一种获得 EMFC 有用基质的技术经济方法。此外，EMFC 的响应模型展示了在生物治疗中利用机器学习的潜力。