基于灰色-神经网络的城市湖库水华预测研究

City Lake Algal Bloom Forecast Research Based on Grey-neural Network

准确的预测水华是及时治理水污染及保护水环境的重点,而对水华形成的机理模型的深入研究是准确预测水华的关键所在.在对城市湖库藻类水华形成机理研究基础上,提出了一种基于改进NGM(1,1,k)及BP神经网络的藻类水华预测模型,结合了灰色模型建模所需信息量少及神经网络非线性预测优势,克服了灰色模型预测精度低和BP神经网络所需训练数据多的缺点,可以解决在监测信息有限条件下的藻类水华预测问题.经过实际验证表明,该模型相对神经网络预测精度高,适合应用于城市湖库藻类水华的预测.

The key point of water pollution control and water environment protection is how to forecast the algal bloom accurately, further more intensive study on the algae bloom formation mechanism is the main point of algal bloom forecast. In study of the city lake algae bloom formation mechanism, an improved me- tabolism BP neural network model of NGM（1,1, k ） is proposed by combining with the benefit that the grey model modeling needs less information and the prediction advantage of neural network. The method over- comes the low prediction accuracy of grey model and the weakness that the BP neural network requires too many training data. The algae bloom forecast problem under the limited monitoring information is solved. Ex- periments show that the prediction accuracy of the grey neural network is much higher than that of the BP neural network, and the method is suitable for city lake algae bloom forecast.