Biofuel is one of the promising alternatives for petroleum-based fuels and the bioethanol produced from agricultural residues/industrial wastes are a viable alternative, and current trend in research. In this context, this work reports the utilisation of the waste potatoes (Solanum tuberosum L.) from food industries as precursors for bioethanol. Waste potato mass (WPM) was pre-treated using ultrasonication and hydrolysed using either hydrochloric acid (US–HCl) or α-amylase (US-Enzyme). The process conditions such as initial S. cerevisiae concentration (10–20 g/L), enzyme concentration (10–30 U/mL), HCl concentration (1–3% v/v) and ultra-sonication time (5–15 min) were modelled using Box-Behnken RSM design and artificial neural networks (ANN). ANN modelled the experimental data better than RSM which was evident from different errors. Optimum parameters were evaluated using genetic algorithm. The optimal parameters predicted for US-HCl hydrolysis was 65.8 mg/L at HCl concentration 2.1%, ultrasonication time 10.7 min, and S. cerevisiae concentration 19.2 g/L with R² 0.979, whereas, for US-Enzyme hydrolysis was 54.1 g/L at α-amylase concentration 25.3 U/mL, ultrasonication time 10.2 min, and S. cerevisiae concentration 19.2 g/L with R² 0.959. Hence, ultrasonic pretreatment increases the bioethanol yield from waste potatoes and US-HCl process was efficient than the US-Enzyme process.

生物燃料是石油基燃料的有前途的替代品之一，由农业残留物/工业废物生产的生物乙醇是可行的替代品，并且是当前的研究趋势。在这种情况下，这项工作报告了利用来自食品工业的废马铃薯（马铃薯）作为生物乙醇的前体。马铃薯废料（WPM）使用超声波进行预处理，然后使用盐酸（US-HCl）或α-淀粉酶（US-Enzyme）水解。工艺条件，例如初始酿酒酵母浓度（10–20 g / L），酶浓度（10–30 U / mL），HCl浓度（1-3％v / v）和超声处理时间（5–15分钟）使用Box-Behnken RSM建模设计和人工神经网络（ANN）。从不同的误差可以明显看出，ANN比RSM更好地模拟了实验数据。使用遗传算法评估最佳参数。在HCl浓度为2.1％，超声处理时间为10.7分钟，酿酒酵母浓度为19.2 g / L，R ^2为0.979的条件下，预测US-HCl的最佳水解参数为65.8 mg / L ，而US-Enzyme水解的最佳参数为54.1 g / L。L在α-淀粉酶浓度为25.3 U / mL，超声处理时间为10.2分钟，酿酒酵母浓度为19.2 g / L且具有R ^{2的情况下}0.959。因此，超声预处理增加了废土豆中生物乙醇的产量，而US-HCl工艺比US-Enzyme工艺更有效。