论文部分内容阅读
针对高可靠性、长寿命复杂产品的可靠性评估过程,在加速寿命退化试验数据的基础上,提出了一种基于试验数据驱动的自适应智能方法,并对某型LED灯管的寿命与可靠性进行预测分析。首先,通过指数模型拟合性能退化曲线,推算出各组应力条件下的伪失效寿命值;再将蚁群算法结合BP神经网络等智能算法应用于寿命预测模型的建立,根据试验证明寿命服从对数正态分布,且检验寿命必须满足置信度区间范围内;最后,预测出正常应力条件下LED灯管的工作寿命。结果表明,基于蚁群神经网络预测LED灯管寿命的方法,预测误差较小,收敛速度快,能够满足工程要求。
Aimed at the reliability evaluation of high reliability and long life complex products, based on the accelerated life degradation test data, an adaptive intelligence method driven by experimental data is proposed. The life and reliability of a certain type of LED lamp Sex prediction analysis. First, the exponential model was used to fit the performance degradation curve to calculate the pseudo-failure life value of each group under stress conditions. Then the ant colony algorithm combined with BP neural network and other intelligent algorithms were applied to the life prediction model. According to the test, The number of normal distribution, and the test life must meet the confidence interval within the range; Finally, predict the working life of LED lamp under normal stress conditions. The results show that the method based on ant colony neural network to predict the life of LED lamp has small prediction error and fast convergence, which can meet the engineering requirements.