Research On Equipment Fault Intelligent Alarming Based On Data Mining

When detecting bearing,to judge whether the bearing is working normally,in most cases,the dynamic signal which receives from single vibrating sensor is analyzed.However,the information which reflected by single sensor is always not good enough.The Vector spectrum technology can effectively integrate the homologous information,get more comprehensive vibrating feature,thus it can extract reliable omen.Because of the higher automation and continuous extent during the process of production,the higher requirement is asked about the reliability of equipment's running and intelligence of fault's judgement.So it plays an important role in the aspects of supporting equipment and productive safety.In the research about predicted method,more and more artificial intelligence arithmetic is discussed and used.Association rules is an effective data mining algorithm,in this paper,the vector spectrum information fusion technology and association rule are combined to do a fault early warning research which its research object is the important parts of rolling bearing.The main work is as follows:(1)The data disposition of the early warning model,the Vector spectrum Hilbert mediation method is used to extract the fault character data under the character frequency of antifriction bearing,The simulation research and experimental result show that the method can effectively extract the fault character data under the character frequency of antifriction bearing.First,receive the two groups of homologous vibration signal in antifriction bearing through sensor,then do Hilbert transformation for the two groups of signal,construct analytic signals,and take the envelope of them,in the end,do the Vector spectrum analyze for the two perpendicular homologous signals,receive Vector spectrum structure,which is used as character data of early warning model.(2)The Vector-Apriori fault early warning model is established.The experiment shows that this model can effectively predict the running state of equipment,then predict the fault of equipment.However,because the Apriori arithmetic require to circular scan to mined databases,produce the candidate item sets that can't satisfiedthe requirement,which results in fewer strong association rules and insufficient real-time warning.Vector Hilbert is used to integrate the information for the original sample date in the normal running state,then do the pretreatment,such as discretization,an d establish the database that are needed in late work.The apriori algorithm is used to find the unknown relationship among the amplitude of eight characteristic frequency,then establish a rule base by using the association rule the has been dug.Match the rule base and running base that has been disposed,to reduce the economic loss,whether current operating machinery is in fault formation stage is needed to be detected.(3)Establish the Vector-FP-Growth fault early early warning model and improve the deficiency of the Vector-Apriori fault early warning model.The experiment showed that the vector-FP-Growth fault early warning model can advance 23 min effectively comparing with the vector-Apriori early warning model,warn the possible fault timely and remind the equipment management personnel to take action.