基于二维云推理的球磨机料位软测量  

Soft Sensor for Ball Mill Fill Level Based on Two-dimensional Cloud Reasoning

作　　者：寄珊珊 龚杏雄[1] 阎高伟[1] 

机构地区：[1]太原理工大学信息工程学院,太原030024

出　　处：《科学技术与工程》2015年第25期61-65,共5页Science Technology and Engineering

基　　金：国家自然科学基金(61450011);山西省自然科学基金(2015011052)资助

摘　　要：针对单一信息反映料位的局限性,采用基于二维云模型不确定性推理方法实现球磨机料位软测量。首先,对振动和振声信号进行预处理,通过Welch方法获得功率谱密度并提取特征值;其次,利用二维云模型对所得特征值进行融合,获得推理系统的前件云、后件云以及相应的规则库;最后,将测试样本作为推理系统的输入,进行二维云模型不确定性推理,实现球磨机料位软测量。实验结果表明,本方法的测量精度能够满足现场测量应用的需求,且具有较高的稳定性。Considering the limitations of single information for fill level measuring, a novel soft sensor for ball mill fill level based on uncertainty reasoning of two-dimensional cloud model is employed. At first, Welch＇ s method is used to extract features from power spectral density of vibration and acoustic signals. Then two-dimensional cloud model is introduced for information fusion, and the antecedent clouds and consequent clouds as well as rule base of reasoning system can be acquired. At last, according to the test samples, the uncertainty reasoning of two-dimensional cloud model is realized and the measurement of ball mill fill level is accomplished. The experiment results show that the proposed method can meet the requirement of field measurement application and is powerful with respect to the robustness.

关 键 词：球磨机料位 二维云模型 不确定性推理 软测量 

分 类 号：TP212.6[自动化与计算机技术—检测技术与自动化装置]