基于GA-ANN模型的A2/O工艺运行参数优化

摘要

影响A2/O工艺运行的参数有许多,这些因素相互联系、相互作用,影响工艺效率.为了弥补控制单一变量法或者设计正交试验法的不足,综合考察多种运行参数对工艺运行效果的影响,建立了基于遗传算法进行全局寻优的神经网络模型(GA-ANN模型),并应用于某城市污水处理厂A2/O工艺的运行优化.获得该厂调试运行期间154组有效监测数据后,随机选取2/3的数据用于GA-ANN模型的求解,1/3的数据用于模型的检验,对工艺运行参数进行优化,得到最佳运行参数组合.结果显示,建立基于遗传算法的神经网络模型用于A2/O工艺运行参数的优化是可行的,可以为污水处理厂运行参数的设置提供理论参考,对调试工作、提高工艺运行效率具有一定的实际指导意义.%A2/O process is one of the major processes in municipal waste water treatment, but many parameters affect the operation effect of A2/O process. And these parameters interact with each other, affecting the efficiency of the process. In order to make up the insufficience of single variable control method or orthogonal designing method, it establishes the neural network model ( GA-ANN model) based on genetic algorithm. The model has been applied to an urban waste water treatment plant by A2/O process optimization. During the commissioning operation of the plant, it has obtained 154 effective monitoring data, and 2/3 of the data has been randomly selected for the GA-ANN model, and 1/3 of the data has been used for the model test. The process parameters have been optimized and get the best combination of operating parameters. The results show that it is feasible to establish the neural network model based on the genetic algorithm for the optimization of the A2/O process operation parameters. It can provide the theoretical reference for setting operation parameter of the waste water treatment plant. And it is also helpful to the practical production and application for adjustment and improvement of the operation efficiency.