Research on Lean Analysis Algorithm for Equipment Centralized Monitoring in Big Data Era

Abstract

The equipment monitoring brought by the smart grid big data is difficult to effectively supervise the operation status of the whole network substation, and the typical defects (familial defects) are difficult to classify and locate. This paper proposes the substation operation state evaluation algorithm and typical defect classification algorithm. The operating state evaluation algorithm of the substation is based on different operation data generated by the substation. By normalizing the data of different dimensions, the substation is divided into different operating state levels. The typical defect classification algorithm establishes and maintains the historical experience database, and calculates the conditional probability of each information item to realize the correlation between the signal and the defect, and finally judge whether the signal is from a typical defect. These two algorithms are effective means for equipment monitoring professionals to realize intelligent supervision of substations and equipment in the era of big data.

A study of sound recognition algorithm for power plant equipment fusing MFCC and IMFCC features

“…The sound signal of the power plant equipment collected by the acoustic camera is pre-processed, and the processed sound signal is more uniform and smooth, which can provide high-quality parameters for feature extraction [1][2] . Feature generation is the process of characterizing the sound signal with fewer dimensions, so the high recognition rate of equipment sound depends on the accuracy of the generated features.…”

“…First, the obtained sound signal of the power plant equipment is cut into 0.5 seconds, that is, the sound sample, and the duration of the sound sample is unified. Then, find the spectrum in the low-frequency to high-frequency frequency bands with the same signal-to-noise ratio [5][6] , ensure the spectrum flattening, and pre-emphasize the sound signal of each sample by passing the sound signal through the high-pass filter of formula (1).…”