基于水位与符号回归的灌区无干扰量水技术研究

Study on non-interference flow gauging method for irrigation ditch based on water level and symbolic regression

为克服传统量水方法水头损失大、技术复杂、经济成本高等问题,提出了一种基于水位和符号回归的灌区无干扰量水方法,并依据高邮灌区典型斗渠2012~2013年实测水位流量数据,对方法应用效果进行了评估。结果表明:符号回归优选出来的5组模型均能较好地估算渠道流量,模拟值与实测值均方根误差RMSE和平均绝对误差MAE分别为71.81~73.43,52.15~56.38m^3/h,调整决定系数超过0.8,说明模型能够较好地反映渠道水位流量关系,模拟精度较高。该方法精度可靠、应用简便、经济可行,在灌区量水方面具有较广的应用前景。

In order to overcome the problems of large head loss,complex technology and high cost in traditional flow gauging,a non-interference flow gauging method based on water level and symbolic regression was proposed.The application effect of this method was evaluated by using the measured data of typical lateral canal at Gaoyou irrigation district from 2012 to 2013.The results show that all the five groups of models selected by symbolic regression can estimate the flow well.RMSE and MAE of the flow rates simulated by five groups of models and by measurement were ranged from 71.81m^3/h to 73.43m^3/h and from 52.15m^3/h to 56.38m^3/h,respectively.The adjusted determination coefficient was more than 0.8,indicating that the model can reflect the relation of water level and flow rate with high simulation accuracy.The proposed method is reliable,simple and economical,and has a wide application prospect in irrigation district.