摘要
针对水处理过程非线性、时变和大滞后的特点,本文采用RBF和BP神经网络分别建立了水处理过程模型,利用水厂实际运行数据对两个模型分别进行了训练和检验。与BP神经网络模型相比,RBF神经网络模型具有逼近能力强、收敛速度快等优点。该模型可以实现对水处理过程的在线辨识,并可进一步用于该过程的神经网络预测控制。
Considering the nonlinear, time-varying and time-delaying property of water treatment process, this paper develops two models of water treatment separately based on RBF and BP neural network. The models have been trained and checked separately by practical data of water plant. Compared with the model based on BP neural network, the model based on RBF neural network has better features such as excellent approximation and fast converge speed. Online identification of water treatment process can be implemented by this model, which also can be applied in neural network predictive control.
出处
《微计算机信息》
北大核心
2007年第34期294-296,共3页
Control & Automation
基金
江苏省高校自然科学研究计划资助项目(06KJB510040)
关键词
水处理
RBF
BP
神经网络
建模
water treatment, RBF, BP, neural network, modeling
