Detecting method for power anomaly of turbulence-considering wind turbine generator set

Abstract

The invention discloses a detecting method for power anomaly of a turbulence-considering wind turbine generator set. The method comprises the steps that first, the model of the set to be detected is determined, and then, the quantitative relationship between the turbulent intensity of wind, the wind speed average value and the set power is determined according to monitoring data of a same model set; then, according to the monitoring data of the set to be detected, the power of normal working of the set is estimated, whether the degree and probability of the drift normal value of the actual powder of the set to be detected in a continuous time exceed a specified range or not are calculated and counted, and whether the power of the set to be detected is abnormal or not is detected. Accordingto the detecting method, the influence of turbulent intensity and the average wind speed on the output power of the set are carefully considered, anomaly criterion is provided based on a statistic method, and the detecting method has the characteristics of being high in accuracy and low in misjudgment rate; and data sample establishment of a normal power model of the set only needs model corresponding, the engineering usability is good, the method can be implemented only through pure software, additional deployment of hardware facilities is not needed, the cost is low, and the project difficulty is low.

Description

Technical field [0001] The invention relates to the field of power detection of wind turbines, in particular to a method for detecting abnormal power of a wind turbine considering turbulence. Background technique [0002] There are many factors affecting the output power of the unit, including wind speed stability, anemometer failure, blade contamination or icing, drive shaft failure, unit resonance, equipment corrosion, and pitch system failure. A high-precision mastery of the actual output efficiency of wind turbines is conducive to improving the accuracy of wind power forecasting, and helps to understand the operating status and health status of the wind turbines in time. [0003] Many studies believe that when the unit is healthy, turbulence is an important reason for the deviation of the output power of the wind turbine from the theoretical value. Even if the average wind speed is the same and the turbulence intensity is different, the output power of the unit will be differe...

AI Technical Summary

Problems solved by technology

However, the power curve provided by the manufacturer often does not accurately give the corresponding relationship between the average wind speed, turbulence intensity and output power, and its accuracy is limited.

Therefore, there is also a problem of accuracy in detecting the abnormal power of the unit based on this power curve.

[0005] On the other hand, in actual engineering, the judgment of the abnormal power of the existing unit focuses on the degree of deviation between the actual measured power of the unit and the power curve of the unit manufacturer, and does not consider much about the deviation probability, which is not rigorous enough

For example, the yaw of the unit and the short-term fluctuation of the grid voltage can cause the output of the wind turbine to deviate greatly from the power curve given by the manufacturer for a short time. However, the deviation probability is not high in a long-term perspective

Therefore, the existing method has the defect of not being rigorous enough, which easily leads to misjudgment of whether the output power of the unit can maintain normal

[0006] It can be seen that the current abnormal detection method of wind turbine output power has the problems of insufficient accuracy and easy misjudgment, which needs to be improved urgently

To create a method for identifying abnormal output power of wind turbines with high precision and low false positive rate for wind turbines operating at unstable wind speeds is a problem that needs to be solved at present.

Images