摘要
应用数学方法和BP神经网络技术对水处理工艺建模进行研究,研究对象为哈尔滨市绍和水厂的原水水质、投药量与沉后水浊度之间的模型关系。在神经网络模型中,为了确保较好的建模效果,挑选代表不同水质特征的数据作为样本集,并采用提前停止法提高网络的推广能力。仿真试验证明与传统的数学模型相比,神经网络模型的建模效果更佳。
Research was conducted for modeling of water treatment process on the basis of mathematical method and BP neural network science. The chemical dosage and turbidities of inlet and outlet of sedimentation tank at Shaohe Waterworks in Harbin were studied by the established BP model. For the best result, samples of different characteristics and ahead stop algorithm were used to improve the promotion sphere of the neural network. Emulated experiments showed that the result of the neural network model was better than that of the traditional mathematical model.
出处
《给水排水》
CSCD
北大核心
2007年第10期110-114,共5页
Water & Wastewater Engineering
基金
国家自然科学基金资助项目(50678047)
