基于多项式回归算法的电量水量转换模型构建——以宁夏西干渠管理处马场滩灌溉站为例

Construction of Electricity-Water Conversion Model Based on Polynomial Regression Algorithm—A Case Study of the Machangtan Irrigation Station of Ningxia West Main Canal Management Office

摘要: 我国农业水价综合改革背景下，严格灌区农业灌溉用水计量，针对各分散灌溉站存在设备陈旧、测量水量成本高、点多面广等问题，急需大规模、低成本、全覆盖的灌溉站水量测量方案。本文采用电流间接测量灌溉站实时灌溉用水量，以宁夏西干渠管理处马场滩灌溉站为例，借助超声波流量计来获取灌溉站电流和水量等历史数据，采用多项式回归算法，构建灌溉站电量水量转换模型，通过验证模型预测的绝对值误差为0.0092，表明构建的模型具有较好的鲁棒性和稳定性，能够实现较为准确地计算灌溉站的用水量，该模型集成到马场滩灌溉水量检测平台中，具有在线动态计算水量、确准度较高、成本低、易于实施等特征，有效满足农业灌溉用水计量的要求，具有较好推广应用价值。

Abstract: Under the background of the comprehensive reform of agricultural water price in my country, the measurement of agricultural irrigation water in irrigation areas is strictly required. In view of the problems of outdated equipment, high water measurement cost, and wide range of points in each decentralized irrigation station, a large-scale, low-cost, and full-coverage irrigation station water measurement solution is urgently needed. In this paper, the current indirectly measures the re-al-time irrigation water consumption of the irrigation station. Taking the Machangtan irrigation station of Ningxia West Main Canal Management Office as an example, the ultrasonic flowmeter is used to obtain historical data such as the current and water volume of the irrigation station, and the polynomial regression algorithm is used to construct the electricity and water volume of the irriga-tion station. Converting the model, by verifying that the absolute value error of the model predic-tion is 0.0092, it shows that the constructed model has good robustness and stability, and can cal-culate the water consumption of the irrigation station more accurately. The platform has the char-acteristics of online dynamic calculation of water volume, high accuracy, low cost, and easy imple-mentation, which can effectively meet the requirements of agricultural irrigation water metering, and has good application value.