摘要
水电大数据算法平台能够通过其内嵌的数据计算功能组件,对水电站实时监控数据进行计算,进而求出其特征值、特征差及缓变率。通过故障诊断模型对其值的逻辑判断来预测故障,因而故障表征量的科学选取就显得尤为重要。本文通过数学统计分析的方法,首先利用MATLAB进行T检验,检验监测量是否是正态分布,讨论用哪种相关分析较为科学,再用SPSS进行双变量相关分析,分析在水电大数据平台中故障模型所选择表征量相互是否为强相关量,强相关就代表两者有相同变化趋势,如此才能判断出一个故障,达到对故障模型的合理性验证目的。经过本论文的故障模型验证,所得结论支撑了水电大数据算法平台故障模型的科学性,相关量选择合理,通过所选相关量建立的相关模型能够判断并预测出故障是否会发生。
The hydropower big data algorithm platform can calculate the real-time monitoring data of the hydropower station through its embedded data computing function component,and then find its characteristic value,characteristic difference and ramp rate. The fault diagnosis model predicts the fault by logical judgment of its value,so the scientific selection of fault characterization is particularly important. In this paper,through the method of mathematical statistical analysis,first use MATLAB to carry out T test to check whether the monitoring quantity is normal distribution,discuss which correlation analysis is more scientific,and then use SPSS for bivariate correlation analysis to analyze the fault in the hydropower big data platform. Whether the selected quantities of the model are strongly correlated with each other,the strong correlation means that the two have the same trend,so that a fault can be judged and the rationality of the fault model can be verified. After the failure model verification of this paper,the conclusions support the scientificity of the fault model of the hydropower big data algorithm platform,and the correlation quantity selection is reasonable. The correlation model established by the selected correlation quantity can judge and predict whether the fault will occur.
作者
杨洋
YANG Yang(Sichuan Huaneng Jialingjiang Hydropower Co.,Ltd.,Nanchong 637000 Sichuan,China)
出处
《电力大数据》
2018年第10期74-81,共8页
Power Systems and Big Data
关键词
故障诊断模型
预测故障
检验
双变量相关分析
强相关量
fault diagnosis model
predictive failure
test
bivariate correlation analysis
strong correlation
