基于人工神经网络的水处理厂混凝投药系统智能建模

摘要：凝固-絮凝是水处理过程中非常重要的一个部分。由于涉及多个物理和化学过程，因此极易产生人为因素导致的错误。为了减少出错率，达到最佳的处理效率，提出以人工神经网络为基础建立智能混凝投药模型。混凝投药模型的设计思路是以辣木作为促凝剂，通过ANN的辅助来实现水质预测和软测量。通过人工智能工具的优化，能够实现对凝固过程最优化处理。基于ANN的混凝投药系统将是解决水处理过程中人为失误的一个有效方法。模拟和试验结果表明，这个新发展的系统能够为一个小规模的农村社区精确地预测在水处理过程中所需要的混凝投药剂量。实际与ANN模拟得到的混凝投药剂量的相关度达到1：0.97，表明ANN模型是完美的。

Abstract: Coagulation –flocculation process remains a very essential part in the water treatment chain. It involves both physical and chemical phenomena and hence susceptible to high percentage of errors due to human factor.  In order to reduce this percentage error and obtain optimal treatment efficiency, an intelligent coagulant dosing based on Artificial Neural Network (ANN) was proposed. Design of the Coagulant dosing using processed Moringa oleifera seed as coagulant was achieved through ANN that helps in water quality forecast and soft measure. Effort was made to suggest the optimization tips in the form of Artificial Intelligent tools that can be used for optimization of coagulation process. Such coagulant dosing based ANN will be a useful method to address most errors common in water treatment cause by human factors.  Experimental results with simulated and real data show that the newly developed system is able to accurately predict coagulant dosage needed in water treatment for a small size rural community. The correlation between actual and ANN estimation of coagulant dosing model is 0.97 of 1.00. This high Correlation of coefficient indicates that the ANN model is a perfect match.