Feature Extraction And Abnormal Condition Identification Based On Image And Acoustic Signal For Fused Magnesia Furnace

In the smelting process of the fused magnesia furnace,on the one hand,it is difficult to establish a mathematical model due to the complicated mechanism.The three-phase electrodes have a strong coupling characteristic.The AC arc has a strong nonlinear characteristic.On the other hand,the industrial environment is very poor.Many workers rely on their experience to operate the machines.Labor intensity is strong.Therefore,it is easy to lead to the occurrence of abnormal conditions and high energy consumption in the production process.At present,for the abnormal conditions identification of the fused magnesia furnace,the operators make decisions by observing the fluctuation law of three-phase electrode currents and the conditions of smelting equipment.However,the operators can’t keep a high degree of attention for a long time at night and identify the abnormal conditions in time,which will lead to the occurrence of production accidents.To solve this problem,the image features and sound features are extracted based on image information and sound information in the smelting process of the fused magnesia furnace.Then the expert system based on rule reasoning is used to identify the abnormal conditions.Image feature extraction and sound feature extraction are beneficial to the realization of automatic identification of the abnormal conditions for fused magnesia furnace.This way can reduce the labor intensity of workers,increase production efficiency and provide the basis for the self-healing control decision.The main contents of this thesis are as follows:(1)Considering high-frequency noise interference of image and the large computation of the color image processing algorithm,the image is preprocessed using the butterworth filtering algorithm and the gray transform algorithm.(2)The threshold segmentation algorithm is used to realize the segmentation of bright spot.The threshold segmentation based on Otsu algorithm is used to realize the segmentation of the smelting zone.(3)By analyzing the image characteristics of the semi-melting abnormal condition in the smelting process,the bright spot area,the average gray of smelting zone and the gray variance of smelting zone are extracted.By analyzing the image characteristics of the overheated abnormal condition in the smelting process,the smelting rate,the change rate of the smelting zone and the change rate of the overall average gray are extracted.(4)By analyzing the sound characteristics of the exhaust abnormal condition in the smelting process,the frequency characteristic and the amplitude characteristic under the frequency characteristic are extracted.(5)According to the extracted image features and sound features,the rules base is constructed.The expert system based on rule reasoning is used to identify the abnorma conditions of the smelting process for the fused magnesia furnace.