The safety and stability of thermal power units directly affect the reliability of the power supply.Axial blowers are key auxiliary equipment of thermal power units.The research on its operation monitoring has significant application value.This paper proposes an operation monitoring method for axial flow blowers based on multilayer perceptron and delay timer.Firstly

the data preprocessing is performed by integrating the mechanism model of the axial blowers to enhance the interpretability of the neural network model.Secondly

the multilayer perceptron is utilized to model the differential pressure of the axial blowers.The residual between the model output and the actual output is calculated as the monitoring indicator.The alarm threshold is designed through probability density function estimation.Finally

the delay timer is designed based on the analysis of the alarm duration of alarm events.The alarm performance

including the false alarm rate and miss alarm rate

is improved with the use of the delay timer.The mechanism knowledge driven data preprocessing improves the modeling performance of the multilayer perceptron. Compared with the univariate monitoring method used in the industrial site

the proposed method considers the relationship between process variables

can detect the abnormality of axial blowers earlier.The effectiveness of the proposed method is illustrated by the stall case of an axial blowers in a thermal power unit.