Vibration Analysis of Shaft Misalignment Using Machine Learning Approach under Variable Load Conditions

Table 6

Results obtained using classification and prediction of the shaft misalignment method.

Rotational speed (rpm)	Type of misalignment	Misalignment (mm)	SVM results		ANN result
			SVM: distance to H	Classed by SVM	Expected outcome (Eo)	Actual outcome (Ao)	% error
500	Parallel	0	0.002	0	0.00	0.0030	0.30
		0.02	1.239	1	0.02	0.0204	2.00
		0.04	2.920	1	0.04	0.0414	3.50
		0.06	3.572	1	0.06	0.0622	2.75
	Angular	0	0.022	0	0.00	0.0019	0.19
		0.01	−1.925	−1	0.01	0.0102	2.00
		0.03	−1.110	−1	0.03	0.0312	4.02
		0.04	−1.592	−1	0.04	0.0410	2.69
1200	Parallel	0	0.014	0	0.00	0.0025	0.25
		0.02	1.239	1	0.02	0.0205	2.79
		0.04	2.920	1	0.04	0.0406	1.57
		0.06	3.572	1	0.06	0.0617	2.89
	Angular	0	0.022	0	0.00	0.0040	0.40
		0.01	−1.925	−1	0.01	0.0103	3.89
		0.03	−1.110	−1	0.03	0.0308	2.79
		0.04	−1.592	−1	0.04	0.0418	4.68
2100	Parallel	0	0.012	0	0.00	0.0010	0.10
		0.02	1.239	1	0.02	0.0251	2.57
		0.04	2.920	1	0.04	0.0408	2.15
		0.06	3.572	1	0.06	0.0616	2.67
	Angular	0	0.022	0	0.00	0.0030	0.30
		0.01	−1.925	−1	0.01	0.0104	4.21
		0.03	−1.110	−1	0.03	0.0311	3.92
		0.04	–1.592	–1	0.04	0.0408	2.21
Average % error:							2.28