论文部分内容阅读
在饮用水的处理中,采用颗粒式过滤介质过滤是个很重要的过程。过程用以确保充分去除携带病原体(如:贾第虫和隐孢子虫,Giardia and Cryptosporidium)颗粒。通常,通过检测过滤后的水浊度来反映过滤性能的优劣。不过,颗粒数对于过滤操作中细微的变化极其敏感,所以颗粒数可以进一步反映过滤效率。埃尔金区水处理厂(WTP)应用人工神经网络(ANN)通过考察过滤后的水中颗粒数,对过滤进行优化。成功地开发了过程模型预测过滤后颗粒数。开发了两套逆过程模型,用来预测颗粒数达到要求时,凝聚剂的最佳用量。对模型进行检验显示,实测值和预测值之间具有较高的相关性。然后将这些ANNs集成到一个优化应用中,通过这个优化应用,模型和在线监控和数据采集系统(SCADA)之间可以进行实时数据传输。
In the treatment of drinking water, the use of particulate filter media filtration is a very important process. The procedure is to ensure that particles carrying pathogens (eg Giardia and Cryptosporidium) are sufficiently removed. In general, the quality of the filtration is reflected by the turbidity of the filtered water. However, the number of particles is extremely sensitive to subtle changes in the filtration operation, so the number of particles can further reflect the filtration efficiency. Elgin District Water Treatment Plant (WTP) Applications Artificial Neural Networks (ANNs) Optimize filtration by examining the number of particles in the filtered water. The process model was successfully developed to predict the number of particles after filtration. Two sets of inverse process models were developed to predict the optimal amount of coagulant when the number of particles is required. Testing the model shows a high correlation between measured and predicted values. These ANNs are then integrated into an optimized application that allows real-time data transfer between the model and the SCADA.