|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011年12期
| 標題 |
| 粒子群優(yōu)化的神經(jīng)網(wǎng)絡在齒輪故障診斷中的應用(20卷) |
| 英文標題 |
| Application to Neural Network of particle swarm optimization in gear fault diagnosis |
| 摘要 |
| 針對BP神經(jīng)網(wǎng)絡容易陷入局部極小及收斂速度慢的問題，本文利用粒子群優(yōu)化算法代替BP算法中的梯度下降法訓練神經(jīng)網(wǎng)絡的權(quán)重和閾值，有效地改善了BP網(wǎng)絡診斷性能；利用訓練后的神經(jīng)網(wǎng)絡對齒輪進行了故障診斷，并比較了基于粒子群優(yōu)化算法與BP 算法的診斷結(jié)果，通過仿真實驗表明：無論是在診斷速度上還是在診斷精度上， PSO- BP神經(jīng)網(wǎng)絡診斷性能都比單獨的運用神經(jīng)網(wǎng)絡有很大提高。 |
| 作者 |
| 新聞作者:張勝召，齊金平，陶海龍 |
| 關(guān)鍵字 |
