Short-Term Power Prediction of Wind Power Generation System Based on Logistic Chaos Atom Search Optimization BP Neural Network

Table 1

Results of power prediction values for the four prediction models.

Sampling points	5	10	15	20	25	30	35	40	45	50
Measured values	21.7242	19.7778	23.2253	35.6639	34.5001	29.7842	30.1208	30.6493	29.1846	41.0771
BP predicted values	22.4122	20.7560	24.2829	36.3452	35.3211	30.2889	30.6706	31.1948	29.6624	41.4405
GA-BP predicted values	21.5707	19.7018	22.9188	34.9702	33.7510	29.3096	29.6088	30.1063	28.6340	40.3058
PSO-BP predicted values	21.6433	19.7858	23.0024	35.0307	33.8162	29.3884	29.6889	30.1845	28.7146	40.3577
LCASO-BP predicted values	21.8132	19.8748	23.3081	35.6955	34.5366	29.8400	30.1753	30.7016	29.2429	41.0865
AE values of BP	0.6880	0.9782	1.0576	0.6813	0.8209	0.5048	0.5498	0.5454	0.4778	0.3634
AE values of GA-BP	0.1535	0.0760	0.3065	0.6936	0.7490	0.4746	0.5120	0.5430	0.5505	0.7712
AE values of PSO-BP	0.0809	0.0080	0.2228	0.6331	0.6838	0.3958	0.4319	0.4648	0.4700	0.7194
AE values of LCASO-BP	0.0889	0.0969	0.0827	0.0316	0.0364	0.0558	0.0545	0.0522	0.0583	0.0094
RE values of BP	0.0316	0.0494	0.0455	0.0191	0.0238	0.0169	0.0183	0.0178	0.0163	0.0088
RE values of GA-BP	0.0070	0.0038	0.0132	0.0194	0.0217	0.0159	0.0170	0.0177	0.0188	0.0187
RE values of PSO-BP	0.0037	0.0004	0.0096	0.0177	0.0198	0.0133	0.0143	0.0151	0.0161	0.0175
RE values of LCASO-BP	0.0040	0.0049	0.0035	0.0008	0.0010	0.0019	0.0018	0.0017	0.0020	0.0002