基于遗传神经网络的伺服机构健康状态预测

为了及时把握伺服机构的健康状态，为装备的管理维护与任务执行提供必要的决策支持，从装备的自然退化趋势出发，提出了一种基于遗传算法优化BP神经网络的预测模型。利用BP神经网络优秀的非线性映射能力构造预测模型，将神经网络初始权值阈值编码，利用改进的自适应遗传算法确定最优解。将该模型应用到伺服机构的健康状态预测上，并与标准BP神经网络及径向基神经网络做比较。结果表明该模型有更好的预测精度及收敛速度，从而验证了模型的有效性。

Health Condition Prediction of Servomechanism Based on GA-BP Network

In order to command health condition of servomechanism in time, and provide necessary decision support for equipment maintenance managing and assignment executing, a prediction model based on GA-BP neural network was proposed by equipment natural degradation trend. BP neural network with excellent nonlinear mapping capability was used to construct prediction model; improved adaptive genetic algorithm was used to search the optimal solution after coding the initial weights and thresholds of neural network. The model was applied to health status prediction of servomechanism, and was compared with normal BP neural network and particle swarm optimization BP neural network. The result shows the model has better prediction precision and convergence speed, which verifies the validity of the model.