某电站大坝安全监测数据处理方法探究

发布时间：2018-10-10 10:32

【摘要】：大坝安全内部和外部变形监测是检测大坝安全运行状况，及时发现工程是否隐患的重要日常工作，是大坝运行管理的重中之重。为保证大坝安全运行，按照《混凝土坝安全监测技术规范》（DL/T5178-2003）、《水电站大坝运行安全管理规定》（国家电力监管委员会第3号令）、《水电站大坝安全监测工作管理办法》（电监安全[2009]4号）等要求，需定期对大坝进行水工监测。在现场检查结合定期监测资料进行对比分析等工作的基础上，，对整个大坝的运行工况做出初步评价，对监测设施的保护、更新改造等提出意见和建议。为了掌握航电工程大坝、电厂、船闸运行期的工作状态，验证设计的各项参数，及时发现异常情况采取措施，以保证大坝、电厂、船闸能够安全得正常运行，因此，要对其进行变形监测。变形监测的工作包括每月一次的变形点测量和每年一次的控制网校核。由于变形点监测是在控制网提供精确坐标的基础上进行，因此控制网校核工作非常关键。大坝安全监测系统的重要观测项目之一是水平位移观测。结合对某电站大坝坝顶上11个观测点以及船闸边墙上18个观测点的观测过程，介绍了在视觉观察法、前方交会法行水平位移的原理，步骤和误差分析，结合历史数据对水平位移方面的观测资料进行统计整理，建立模型和多元线性回归分析，并讨论了温度、水位、时效对水平位移的影响深度，坝顶的水平位移，发现船闸和水库水位温度的影响是显而易见的，老龄化的水平位移的影响不大;采用第二代回归方法偏最小二乘回归重新建立统计模型，与多元线性回归分析做简要对比，发现偏最小二乘回归的拟合效果比多元线性回归更好。本文主要介绍了用回归分析和偏最小二乘法两种方法对某电站大坝及船闸上点的测量数据进行分析，并对分析结果进行比较，可以得到偏最小二乘法方法得到的拟合结果更好，建议在以后数据处理方面多偏向于用最小二乘法进行分析处理，得到的拟合结果综合看误差更小，结果比其他方法更精确。
[Abstract]:The internal and external deformation monitoring of dam safety is an important daily work to detect the safe operation condition of the dam and to discover the hidden trouble of the project in time. It is the most important task of the dam operation management. In order to ensure the safe operation of the dam, According to the Technical Specification for Safety Monitoring of concrete dams (DL/T5178-2003), the regulations on the Safety Management of Hydropower Dam Operation (order No. 3 of the State electricity Regulatory Commission), and the measures for the Management of Dam Safety Monitoring of Hydropower stations (Power Supervision Safety [2009] No. 4), etc. Hydraulic monitoring of the dam is required on a regular basis. On the basis of comparison and analysis of site inspection and periodic monitoring data, this paper makes a preliminary evaluation of the operation conditions of the whole dam, and puts forward suggestions and suggestions on the protection of monitoring facilities and the renewal and renovation of the monitoring facilities. In order to master the working state of the dams, power plants and locks in avionics engineering, verify the design parameters, and find out the abnormal situation in time, measures should be taken to ensure the safe operation of the dams, power plants and locks, so that the dam, power plant and lock can operate normally and safely. It should be monitored for deformation. Deformation monitoring includes monthly deformation point measurement and annual control network checking. Because the deformation point monitoring is based on the precise coordinates provided by the control network, it is very important to control the network checking. One of the important observation items of dam safety monitoring system is horizontal displacement observation. Combined with the observation process of 11 observation points on the dam top of a hydropower station and 18 observation points on the side wall of a ship lock, the principle, steps and error analysis of horizontal displacement in visual observation method and forward intersection method are introduced. Based on the historical data, the observation data of horizontal displacement are analyzed, the model and multivariate linear regression analysis are established, and the influence of temperature, water level and aging on horizontal displacement is discussed, and the horizontal displacement of dam top is also discussed. It is found that the influence of shiplock and reservoir water level temperature is obvious, and that the influence of aging horizontal displacement is not obvious, the second generation regression method partial least square regression is used to re-establish the statistical model, and a brief comparison is made with the multivariate linear regression analysis. It is found that the fitting effect of partial least square regression is better than that of multivariate linear regression. This paper mainly introduces two methods of regression analysis and partial least square method to analyze the measured data of the upper point of dam and lock of a certain power station. By comparing the results of the analysis, the fitting results obtained by the partial least square method are better. It is suggested that the least square method should be used in the analysis of the data processing in the future, and the result of fitting is more accurate than other methods.
【学位授予单位】：西北农林科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TV698.1

【参考文献】