论文部分内容阅读
为了对电动舵机进行致命故障检测,根据模糊聚类方法建立了舵机标准状态样本,通过计算待测状态样本与标准状态样本之间的距离,将其归为最近的一类状态。首先,对舵机原始状态样本数据进行了归一化处理。接着,选用夹角余弦法建立了各样本间的模糊相似矩阵。然后,对初始聚类中心矩阵和隶属度矩阵进行迭代运算,设定最大迭代误差并结束迭代过程,得到了舵机标准状态样本。最后,搭建了舵机状态检测的试验平台,在拷机试验中舵机控制器实时运行状态检测程序,实时计算待测样本与标准状态样本的距离。实验结果表明:状态检测程序运行时间只需0.23 us,且检测结果全部正确。此状态检测方法满足了电动舵机故障检测准确性和实时性的要求。
In order to detect the fatal fault of the electric steering gear, a standard sample of the steering gear was established based on the fuzzy clustering method. The distance between the test state sample and the standard state sample was calculated and classified as the latest type of state. First of all, the original state of the steering gear sample data were normalized. Then, using the included cosine method to establish the fuzzy similarity matrix between samples. Then, the initial clustering center matrix and membership matrix are iteratively computed, the maximum iterative error is set and the iterative process ends. The standard sample of steering gear is obtained. Finally, a test platform for steering condition detection is set up. During the test, the servo controller real-time running status test program is used to calculate the distance between the test sample and the standard status sample in real time. The experimental results show that the running time of the state detection program is only 0.23 us, and the test results are all correct. This state detection method to meet the electric servo fault detection accuracy and real-time requirements.