基于方差补偿AKF-BP的电解液式测斜仪精度提升研究
2024,32(10):319-325
摘要：电解液式测斜仪广泛适用于测量土石坝、面板坝、岩土边坡、路堤、基坑等结构物的水平位移监测。为消除环境温度变化和测量噪声引起的电解液式测斜仪测量误差，进行了传感器温度实验，并对硬件本身、工作环境、数据采集三种测量误差影响因素和温度变化影响因素进行详细分析，提出了一种基于方差补偿自适应卡尔曼滤波（AKF）和BP神经网络（BPNN）的精度补偿模型。结果表明，补偿后的电解液式测斜仪水平位移的均方误差相较于补偿前平均降低了80%以上，从而有效减少计算水平位移的误差，极大提升了电解液式测斜仪的测量精度，可应用于电解液式测斜仪实际工程。
关键词：电解液式倾角传感器；固定式测斜仪；精度提升；方差补偿自适应卡尔曼滤波；BP神经网络
Research on Accuracy Improvement of Electrolyte Inclinometer Based on Variance Compensation AKF-BP
Abstract：Electrolyte inclinometer is widely used to measure the horizontal displacement of earth-rock dam, face dam, rock slope, embankment, foundation pit and other structures. In order to eliminate the measurement error caused by ambient temperature change and measurement noise, the sensor temperature experiment is carried out, and the hardware itself, working environment, data acquisition three factors affecting the measurement error and temperature change are analyzed in detail. A precision compensation model based on variance compensation adaptive Kalman filter (AKF) and BP neural network (BPNN) is proposed. The results show that the mean square error of the horizontal displacement of the electrolytic inclinometer after compensation is reduced by more than 80% compared with that before compensation, which effectively reduces the error of calculating the horizontal displacement and greatly improves the measuring accuracy of the electrolytic inclinometer.
Key words：electrolyte tilt sensor; fixed inclinometer; accuracy improvement; adaptive Kalman filtering with variance compensation; BP neural network
收稿日期：2023-09-08
基金项目：上海市2021年度“科技创新行动计划”社会发展科技攻关项目
