Abstract:

Abstract:The vibration protection setting value of hydraulic generator unit is set by trial statistics generally and the accuracy of setting result strongly depends on the accuracy of experimental data and the experience of technician,therefore it has certain limitations and cannot meet the demands of flexibility and accuracy for setting the protection value.This paper proposed a new protection value setting method which based on the statistical analysis of historical data.Firstly,different setting value areas are divided according to the operation condition of hydraulic generator unit.After that,a numerical prediction model of the vibration monitoring points for different setting value areas is established through the extreme learning machine (ELM) algorithm.At last,the setting values of different areas are set based on the output values of the model. Experiments proved that the method can set the setting values of different areas flexibly and accurately according to the historical operating data.

Key words: Keywords:hydraulic generator unit, vibration protection setting value, extreme learning machine algorithm, trial statistics, method of value setting