摘要
南水北调中线各节制闸、分水口门处均设置了流量计。分水口门处的流量测量是分水水量计量的依据,节制闸的流量数据是输水调度的重要参考,流量计的率定是至关重要的工作。传统率定流量计的方法是通过水文精测法对流量计测量数据进行修正,一般根据比测数据的大致分布,人工选择数学函数拟合流量计与流速仪所测流量数据之间关系。这种方法在函数形式的选择上存在较大的人为随意性,并且往往需要多次试验,反复比较才能找到较为满意的线型,工作量大。针对上述问题,将人工神经网络应用于南水北调中线京石段滹沱河节制闸流量计率定中,对32组实测样本进行数据拟合,绝对值误差和为7.715,而传统利用二次函数进行拟合的误差为8.019,人工神经网络拟合精度较高,且该技术能够根据实测数据自动生成回归模型,适应性强,有较好的推广应用价值。
Many flowmeters are set up in South- to- North Water Diversion,and the accuracy of flow rate is very important to water delivery. Traditional method to calibration need choose confirmed function in advance,and expense a lot of time. A new model of information diffusion is established to regress the flux relation between flowmeter and velocimeter. The model that is more intelligent and flexible can find the optimal structure automatically by neural networks. Take the new method into calibration of flowmeter in south- to- north water diversion middle route project,and compare neural networks model with traditional methods. Results show that the new model has high accuracy and is effective. The paper provides a new approach to calibrate flowmeter in south- to- north water diversion project.
出处
《人民黄河》
CAS
北大核心
2015年第3期129-131,共3页
Yellow River
关键词
流量计率定
人工神经网络
南水北调中线
滹沱河节制闸
Calibration of flowmeter
Artificial neural networks
South-to-North Water Diversion Middle Route
Hutuohe gate
