Outlier discrimination method for power curve data of a wind turbine generator

Abstract

The invention discloses an outlier discrimination method for power curve data of a wind turbine generator. The method comprises the following steps: dividing data according to a certain wind speed andpower interval after a series of preprocessing steps based on real-time operation data of a wind turbine data acquisition and monitoring control (SCADA) system including wind speed, active power andthe like; further detecting suspected outliers based on three outlier detection algorithms of mean distance discrimination (AVDC), local abnormal factor (LOF) and density-based clustering (DBSCAN) considering noise; and finally, identifying the real outlier from the suspected outliers based on a real outlier discrimination criterion. The method is based on data driving, has no special requirementsfor other information of the wind turbine generator, and has high universality. Compared with the prior art, the method has the advantages that the characteristics of the power curve data set are considered while the advantages of the mainstream outlier detection method are combined, the data quality is guaranteed, and the method has high theoretical property and applicability.

Description

technical field [0001] The invention relates to a method for discriminating data outliers, in particular to a method for discriminating outliers facing wind turbine power curve data. Background technique [0002] In a modern society where traditional fossil energy reserves are increasingly scarce and environmental degradation is intensifying, wind energy, as one of the alternatives to new energy sources, has attracted widespread attention from the public due to its environmental friendliness, abundant reserves and excellent renewable characteristics. The total installed capacity and the growth rate of installed capacity have also gradually ranked among the top three in the world. In my country, with the gradual popularization of big data technology in the industrial field, in the past ten years, the wind power industry has gone from site selection, layout, construction, to wind turbine optimization, scheduling, control, to subsequent evaluation, operation and maintenance, ma...

AI Technical Summary

Problems solved by technology

However, the current existing power curve data mainly has the following deficiencies in the methods of detecting outliers and outliers: (1) Most of the research on abnormal points and outliers uses data under different working conditions Abnormalities are detected as a whole, ignoring the impact of differences in different working conditions; (2) Traditional power curve data outliers and abnormal point discrimination methods often only use a single algorithm for detection, and the effect is greatly limited by the subjective nature of a small number of parameters. Select constraints

[0004] Therefore, the existing power curve data cannot process the wind turbine power curve data information with high accuracy from the data set in terms of abnormal point and outlier point detection, and it is necessary to improve the related algorithm of comprehensive power curve data characteristics

Images