The Fault Diagnosis Method Based On HMM Is Applicated In The Blower

Sintering blower is served as a key equipment used by ironmaking raw materialsworkshop of the metallurgical industry, which acts an important role involved in theproduction process of ironmaking anterograde and equipment safety. However, there wereless researches on the vibration test ways, fault diagnosis technology and diagnosisquipment for a long time in the domestic，which only stayed in the low lever stage of thecircuit testing, the analysis of the post-accident and dynamic balance analysis. Takeadvantage of vibration test signals with the modern signal processing theory, and it isnecessary that dig out the deeper level of sintering blower fault. Therefore, this paper carryout the following research work:(1) Choose vibration testing method according to the blower scene. Firstly,according to the test scheme, we choose portable ZonicBook/618e eight channels ofvibration testing system, and collect the blower vibration signal through accelerationsensor. At last, eZ-Analyst software of the IOtech company that match the data acquisitionsystem display and save data. Collect two different conditions of blower vibration signalRespectively, and lay the experiment foundation to use the vibration signal to analysisi theblower state based for the next step.(2) Adopt the method based on EMD-ICA that analyse three burn workshop SJ14500spindle blower rotor single channel information running state and extract the feature,which is effective to extract the multiple vibration source information in a single channelsensor. Therefore, the known characteristics informations are used in the HMM training,the unknown characteristics informations are taken into HMM to diagnosis the blowerrotor fault, and realize the blower fault diagnosis of the single senser, and solve theincorrect diagnosis caused by incomplete characteristic information because of using singlesensor.(3) Adopt the method based on theory of vector spectrum that analyse iron three burnworkshop SJ14500spindle blower rotor double channel information running state and extract the feature, which is effective to blend the two channel sensor characteristics andresponse the running state of the equipment. Therefore, the known characteristicsinformations are used in the HMM training, the unknown characteristics informations aretaken into HMM to diagnose the blower rotor fault.Therefore, this paper aims at a sintering blower of a sintering workshop, which iscoupled with the appropriate vibration testing to access to the steady state and transit dataof the blower, and realizes the blower fault diagnosis through some modern signal analysisand data processing. So provide new methods for blower fault diagnosis, which are signalchannel sensor and double channel sensors.