论文部分内容阅读
通过研究合金热挤压过程中的突变现象与观测参数之间的关系,提出了一种观测值增量估值模型,在此基础上构建了一种基于增量估值模型的混杂系统,实现对挤压过程连续状态和离散状态的有效辨识;其中,针对连续状态的辨识,为了对参量进行有效检测,将参量设计为增量加速度向量,根据观测量瞬间变化时原有稳定系统观测模型与实际情况不匹配,从而导致新息序列统计特性发生变化的事实,构造了一种增量向量检测器;针对离散状态的辨识,采用隐马尔科夫模型,结合混杂模式转换概率矩阵,构建了一种离散状态辨识器。最后,基于新提出的混杂系统模型,设计并实现了一种有效的突变预测算法,仿真实验结果表明,新算法具有良好的突变预判性。
By studying the relationship between the phenomenon of abrupt change during hot extrusion and the observed parameters, an incremental estimation model of the observed value is proposed. Based on this, a hybrid system based on incremental estimation model is constructed and implemented For the continuous state identification, in order to effectively detect the parameters, the parameters are designed as incremental acceleration vector, according to the observation of the instantaneous changes in the original observation system and the stable system observation model The actual situation does not match, which led to the fact that the statistical characteristics of the new interest rate changes. An incremental vector detector was constructed. For the identification of discrete states, a hidden Markov model was used in combination with the mixed mode transition probability matrix to construct a Discrete state recognizer. Finally, based on the newly proposed hybrid system model, an effective mutation prediction algorithm is designed and implemented. The simulation results show that the new algorithm has good mutation predictability.