Fault Diagnosis of Wind Turbine Pitch System Based on Multiblock KPCA Algorithm

When the wind turbine pitch system is in operation, due to strong coupling of internal structure system, it difficult accurately locate fault only relying on prior knowledge. And when using data-driven contribution graph method for location, influence variable, value non-fault variable will become larger, which cause a tailing effect and misdiagnosis. In this paper, multiblock kernel principal component algorithm (MBKPCA) proposed. algorithm, variables operation process are divided into several blocks based historical data perform faults each block diagnosis. Taking an area North China as actual calculation example, proposed analysis compared with traditional algorithm. The results show that MBKPCA can effectively reduce “tailing effect” used identify source, source achieves higher detection accuracy.