Intelligent Fault Diagnosis of Delta 3D Printers Using Local Support Vector Machine by a Cheap Attitude Multi-sensor

Identifying the most effective features and improving the accuracy of health status identification have an important impact on the printing reliability of 3D printers. In this paper, local support vector machine (LSVM) is used as an attitude monitoring method for condition recognition of delta 3D printers. To reduce the cost of the experiment, a cheap nine-channel attitude sensor was installed on the printer's mobile platform to monitor the printer's working status. The attitude data of 13 fault types and 1 normal condition under different working conditions were collected to analyze its diagnosis of health conditions. To improve the fault classification performance, locality preserving projections (LPP) was adopted for dimensionality reduction and to obtain a feature space by using all the channels dataset from attitude sensor. To this end, the support vector machine was adopted for fault-diagnosis modeling. For comparison, the attitude monitoring data of each channel is adopted for model training and testing. Furthermore, peer methods such as KLPP and NPE were used for fault diagnosis with the same data. The results indicated that multi-information attitude monitoring with LSVM was effective for condition recognition of delta 3D printers.