当前位置: X-MOL 学术Process Saf. Environ. Prot. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
The use of artificial intelligence models in the prediction of optimum operational conditions for the treatment of dye wastewaters with similar structural characteristics
Process Safety and Environmental Protection ( IF 6.9 ) Pub Date : 2020-11-01 , DOI: 10.1016/j.psep.2020.06.020
Alain R. Picos-Benítez , Blanca L. Martínez-Vargas , Sergio M. Duron-Torres , Enric Brillas , Juan M. Peralta-Hernández

Abstract This work assesses the effectiveness of an artificial intelligence (AI) model based on an artificial neural networks (ANN) – genetic algorithm (GA) in the prediction of the behavior and optimization of the treatment of sulfate wastewaters with Bromophenol blue dye using an electro-oxidation (EO) process. Trials were made with a filter press-type reactor with a boron-doped diamond (BDD) anode. The ANN model was trained with 51 electrolytic experiments by using the electrolysis time, flow, current density, pH and dye concentration as input variables and the discoloration efficiency as the output one. The performance of ANN was measured with RMSE and MAPE values of 10.73 % and 8.81 %, respectively, calculated from real and predicted values. Optimum conditions determined by GA were reached for the inputs of 10 min, 11.9 L min−1, 31.25 mA cm−2, 2.8 and 41.25 mg L−1, giving a discoloration efficiency of 88.8 ± 0.3 %, close to 95.5 % predicted by the model. To validate the AI model, the same experimental conditions were applied to treat wastewaters with Bromothymol blue and Thymol blue, with analogous structures to Bromophenol blue, and a mixture of the three dyes by EO. In all cases, the loss of color decayed following a pseudo-first-order kinetics, with similar apparent rate constants. For the dye mixture, 69 % COD was reduced at 60 min, with 13 % average current efficiency and 0.26 kW h (g COD)-1 energy consumption. The AI model is a strong tool to design, control and operate the EO process with a BDD anode to treat wastewaters with similar dyes.

中文翻译:

人工智能模型在预测处理具有相似结构特征的染料废水的最佳操作条件中的应用

摘要 这项工作评估了基于人工神经网络 (ANN) – 遗传算法 (GA) 的人工智能 (AI) 模型在预测行为和优化使用溴酚蓝染料处理硫酸盐废水时的有效性。 -氧化 (EO) 过程。使用带有掺硼金刚石 (BDD) 阳极的压滤式反应器进行了试验。通过使用电解时间、流量、电流密度、pH 值和染料浓度作为输入变量,以变色效率作为输出变量,对 ANN 模型进行了 51 次电解实验训练。根据实际值和预测值计算出的 RMSE 和 MAPE 值分别为 10.73 % 和 8.81 %,测量了 ANN 的性能。对于 10 分钟、11.9 L min-1、31.25 mA cm-2 的输入,达到了由 GA 确定的最佳条件,2.8 和 41.25 mg L-1,变色效率为 88.8 ± 0.3 %,接近模型预测的 95.5 %。为了验证 AI 模型,使用相同的实验条件处理含有溴百里酚蓝和百里酚蓝的废水,其结构与溴酚蓝类似,并且通过 EO 处理三种染料的混合物。在所有情况下,颜色的损失都遵循伪一级动力学衰减,具有相似的表观速率常数。对于染料混合物,60 分钟时 COD 降低了 69%,平均电流效率为 13%,能耗为 0.26 kW h (g COD)-1。AI 模型是一种强大的工具,可用于设计、控制和操作带有 BDD 阳极的 EO 过程,以处理具有类似染料的废水。应用相同的实验条件,用溴百里酚蓝和百里酚蓝处理废水,其结构与溴酚蓝类似,EO 作用下三种染料的混合物。在所有情况下,颜色的损失都遵循伪一级动力学衰减,具有相似的表观速率常数。对于染料混合物,60 分钟时 COD 降低了 69%,平均电流效率为 13%,能耗为 0.26 kW h (g COD)-1。AI 模型是一种强大的工具,可用于设计、控制和操作带有 BDD 阳极的 EO 过程,以处理具有类似染料的废水。应用相同的实验条件,用溴百里酚蓝和百里酚蓝处理废水,其结构与溴酚蓝类似,EO 作用下三种染料的混合物。在所有情况下,颜色的损失都遵循伪一级动力学衰减,具有相似的表观速率常数。对于染料混合物,60 分钟时 COD 降低了 69%,平均电流效率为 13%,能耗为 0.26 kW h (g COD)-1。AI 模型是一种强大的工具,可用于设计、控制和操作带有 BDD 阳极的 EO 过程,以处理具有类似染料的废水。具有相似的表观速率常数。对于染料混合物,60 分钟时 COD 降低了 69%,平均电流效率为 13%,能耗为 0.26 kW h (g COD)-1。AI 模型是一种强大的工具,可用于设计、控制和操作带有 BDD 阳极的 EO 过程,以处理具有类似染料的废水。具有相似的表观速率常数。对于染料混合物,60 分钟时 COD 降低了 69%,平均电流效率为 13%,能耗为 0.26 kW h (g COD)-1。AI 模型是一种强大的工具,可用于设计、控制和操作带有 BDD 阳极的 EO 过程,以处理具有类似染料的废水。
更新日期:2020-11-01
down
wechat
bug