Research On 3D Printing Process Monitoring Method Based On Fused Deposition Molding

In recent years,3D printing technology is applied more and more widely,and with the quality problem is becoming more and more prominent,so many experts and scholars carried out extensive research both at home and abroad,in the increase of material manufacturing fused deposition molding 3D printers,most widely used,but the fused deposition molding 3D printing made the stability of product quality is bad,it is difficult to implement in the field of industrial application,at the same time,as a result of the fused deposition molding 3D printing processing time is longer,so the print will encounter in the process of abnormal working condition of print,now on the market of printer are not equipped with process monitoring technique,which is one of the important problems of 3D printing.These problems have seriously restricted the development of 3D printing technology,so it is of great significance to study the process monitoring technology of melt deposition molding 3D printing technology.Firstly,a process monitoring platform based on melt deposition molding 3D printer was built,typical printing conditions of melt deposition molding 3D printer were analyzed,and fault conditions often encountered in the printing process were discussed,as well as the most fundamental causes of these fault conditions.The optimal position of the vibration acceleration sensor is selected as the wire feeding mechanism through the analysis of working conditions,so that the extrusion state of the nozzle material can be well monitored.Secondly,the EEMD method was used to decompose the pre-processed signals in different print conditions,and the IMF component was obtained through the decomposition of this method.The correlation coefficient between each IMF component and the original signal was calculated and the standard deviation was calculated as the data to judge the effectiveness of the IMF component,and the first two IMF components were all effective components.The energy calculation will be carried out for the IMF component judged as effective and the eigenvectors will be formed.All the eigenvectors will be identified through KNN under three working conditions.Again,the EEMD method in different flow rate ratio of fused deposition molding for feature extraction of 3 d printing conditions used by KNN classification method to extract the characteristics of classification processing,classification effect is poorer,and analyzes the cause of this result is due to the different flow ratio condition,the working process of the printer,the nozzle motor motion has no obvious change,then aimed at this problem,the proposed feature extraction method based on frequency,by comparing the frequency amplitude is different,can extract characterization under different flow rate ratio of amplitude frequency characteristic value,the KNN method is used to implement different classification of working condition of wire feed ratio.Finally,the KNN classification method exist some disadvantages,such as unbalanced sex differences of sample classification of fault signals,large amount of data in computation time is very long,and fault tolerance of training data such as faults,is proposed based on Dirichlet process mixed model in fused deposition molding 3D printing different wire feeding conditions of process monitoring,and in the execution Dirichlet process in the process of the hybrid model,need not specify the size of the clusters,the algorithm itself will be determined according to the intrinsic nature of the input data of each sample point,for a new input sample under test,Dirichlet process in the hybrid model will assign it to the group with the highest probability.It has good fault-tolerant performance for samples,and is less sensitive to samples with unbalanced hierarchy,which solves the shortcomings in KNN algorithm and can better realize the identification of different wire feeding conditions.