基于人工神经网络的造纸废水处理动态仿真

研究了人工神经网络对废纸造纸废水处理过程动态仿真的可行性,采用误差反向传播网络(BP网)建立了表征原水COD、加药量、进水流量、历史出水COD与预计出水COD之间复杂关系的动态模型,并对不同训练方法进行了比较,发现带有动态调整的方法具有较好的效果,其模型的计算输出值与过程的实际输出值具有较好的一致性.对造纸厂现场排放的废水的实验表明,该模型可用于废纸造纸废水处理的动态描述.

An ANN based dynamic simulation for paper making wastewater treatment

The feasibility of dynamic simulation based on artificial neural network (ANN) for paper making wastewater treatment was studied. With Error Back propagation(BP) network, a dynamic simulation model showing the complex relationship between the influent COD, the added medicament quantity, the influent quantity, the historical efflent water COD and the predictive effluent water COD was established. The comparision between the different training methods of ANN shows that the method of dynamic weight adjusting indicated the better effect and the output of the model is consistent with that of the experiment. The paper shows the dynamic simulation model can effectively describe the process of paper making wastewater treatment.