In view of the frequent failure of power generation equipment under the background of frequent deep peak regulation

flexible operation

energy saving and consumption reduction of thermal power units

a coal mill fault warning method based on multiple state estimation-analytic hierarchy process is proposed. Firstly

based on the characteristic parameters of coal mill Spearman correlation analysis for dimension reduction

equidistant sampling method is used to extract some samples from a large number of coal mill history data memory matrix

after normalization

memory matrix is formed. Then

multi-state estimation algorithm is adopted to calculate the corresponding memory estimated vector according to the memory matrix and the observation vector. The characteristic parameters are given different weights by using analytic hierarchy process(AHP)

and the fusion similarity between the observed vector and the estimated vector is calculated

and the fault warning of coal mill is carried out based on the adaptive threshold method. Finally

the actual fault data of a roller medium speed coal mill is taken as an example to verify the effectiveness of the method. The results show that

this method has less misalarm rate and false alarm rate for coal mill fault warning

which can reduce the actual fault probability of coal mill to a certain extent.