Method for judging temperature anomaly of water turbine by using big data

Abstract

The invention discloses a method for judging temperature anomaly of a water turbine by using big data, and belongs to the field of water turbine anomaly monitoring, the method for judging the temperature anomaly of the water turbine by using the big data performs judgment by using a difference value of a plurality of monitoring points on a single component in a water turbine set as a feature, and can avoid the situation that the temperature anomaly of the water turbine is abnormal due to the difference of installation processes. The temperature difference between different parts in the same hydraulic turbine set and the statistical value are used for judging whether the equipment runs normally or not, and the situation that the equipment runs normally or not due to changes of working conditions and environmental factors can be avoided. According to the method, the temperature of each component in the hydraulic turbine set is affected, a historical data model is constructed through historical data calculation, anomaly analysis can be carried out by means of a historical trend, real-time data anomaly detection and multiple analysis and judgment can also be carried out based on a sliding window, and the accuracy of temperature anomaly judgment is further improved.

Description

technical field [0001] The invention relates to the field of abnormality monitoring of hydraulic turbines, and more particularly, to a method for judging abnormal temperature of hydraulic turbines by using big data. Background technique [0002] In the prior art, in order to ensure the safety and reliability of the operation of the hydraulic turbine unit, a large number of temperature sensors are added to each part of the number theory unit to measure its working temperature in real time. Multiple sensors are installed at the location, that is, multiple temperature measurement points are set on the same component to be monitored, and the data is collected into the monitoring system through the automation system. [0003] In the prior art, by configuring the maximum threshold alarm, statistical historical statistical thresholds or statistical interval thresholds according to working conditions, these do not consider the influence of the environment. Even under the same workin...

AI Technical Summary

Problems solved by technology

[0003] The existing technology configures the maximum threshold alarm, statistical historical statistical threshold or statistical interval threshold method according to working conditions, none of which considers the influence of the environment. In addition, due to installation process problems, the measurement values ​​of different measurement points of the same monitoring component also vary greatly, and the use of historical statistics or working condition intervals often lead to differences in statistical values ​​due to differences in the external environment, even for the same in different years. There are also certain differences in the day and night data values ​​of a month, so the statistical method can only be used to view the overall trend changes, and is not suitable as an alarm judgment threshold. When using standard or empirical alarm thresholds, due to the differences in the values ​​​​of each measurement point, If a single point exceeds the limit alarm, some points exceed the limit but other point values ​​do not exceed the limit, so it is impossible to judge whether the value is abnormal, which affects the detection accuracy

Images