Diagnosis of Elevator Faults with LS-SVM Based on Optimization by K-CV

Table 1

Results for identifying elevator faults with a LS-SVM optimized by cross validation.

Faults of tested samples	Input of SVM									Output
	Energy characteristics					Time Domain characteristics			Noise extremes (dB)
							Peak-to-peak values in direction (m/s^2)	Peak-to-peak values in direction (m/s^2)
Normal state (1)	0.089	0.167	0.124	0.118	0.068	0.889546	0.009	0.01	47.9	1
Normal state (2)	0.093	0.173	0.136	0.121	0.079	0.796675	0.009	0.012	47.1	1
Deviation of guide rail (1)	0.044	0.242	0.276	0.171	0.026	4.139485	0.191	0.024	58.1	2
Deviation of guide rail (2)	0.041	0.241	0.285	0.166	0.024	3.247673	0.178	0.025	59.5	2
Deviation of shape of guide shoe (1)	0.56	0.294	0.028	0.031	0.012	3.173433	0.021	0.02	55.6	3
Deviation of shape of guide shoe (2)	0.534	0.295	0.027	0.034	0.018	4.272636	0.025	0.019	55.2	3
Abnormal running of tractor (1)	0.545	0.294	0.028	0.038	0.022	2.380825	0.021	0.02	56.4	4
Abnormal running of tractor (2)	0.582	0.273	0.026	0.036	0.02	3.567978	0.025	0.018	55.9	4
Error of rope groove of traction sheave (1)	0.814	0.031	0.022	0.022	0.024	3.677453	0.024	0.022	57.1	5
Error of rope groove of traction sheave (2)	0.79	0.026	0.021	0.024	0.022	2.824331	0.023	0.022	56.4	5
Deviation of guide wheel (1)	0.02	0.212	0.11	0.066	0.105	3.187661	0.124	0.1	53.8	6
Deviation of guide wheel (2)	0.022	0.207	0.108	0.056	0.105	2.095107	0.115	0.1	54.2	6
Uniformity of tension on wire rope (1)	0.12	0.525	0.028	0.118	0.022	2.109227	0.02	0.019	55.6	7
Uniformity of tension on wire rope (2)	0.124	0.566	0.028	0.109	0.021	3.099955	0.023	0.02	55.9	7