基于人工免疫算法的离散隐马尔科夫故障诊断模型优化

目的解决离散隐马尔科夫模型在行星齿轮箱故障诊断中的自适应性和泛化性问题。方法建立人工免疫优化模型,将包含易被误判样本的多样本集作为抗原,以其正确识别率为适应度函数,不断对初始观测矩阵进行增殖、变异和选择,获得识别率最高时的初始观测矩阵,利用人工免疫算法对隐马尔科夫故障诊断模型的初始观测矩阵进行优化。结果将该方法应用于行星齿轮箱的故障诊断中,通过不同工况下的对比试验、单样本和多样本优化对比试验,验证了优化后的隐马尔科夫故障诊断模型的诊断效果。结论优化后的隐马尔科夫故障诊断模型具有更好的适应性,诊断精度显著提高。

Optimization of Discrete Hidden Markov Fault Diagnosis Model Based on Artificial Immune Algorithm

Objective To solve the adaptivity and generalization of discrete hidden Markov fault diagnosis model in planetary gearbox. Methods An artificial immune optimization model was established for the initial observation matrix of hidden Markov fault diagnosis model. To obtain the highest recognition rate, the multi-sample set containing the samples that are easily to be misjudged was used as the antigen. And the correct recognition rate was used as the fitness function. The initial observation matrix was continuously propagated, mutated and selected to obtain the highest recognition rate. The initial observation matrix of hidden Markov fault diagnosis model was optimized by the artificial immune algorithm. Results The method established was applied to the fault diagnosis of planetary gearbox. The diagnostic results of the optimized hidden Markov fault diagnosis model were verified by comparison test under different working conditions and single and multi-sample optimization comparison test. Conclusion The optimized hidden Markov fault diagnosis model has better adaptability and significant diagnostic accuracy.