热压过程多传感器数据融合方法  

Multisensor Data Fusion Methods in Hot-Pressing Process

作　　者：于海英[1] 花军[1] 

机构地区：[1]东北林业大学,哈尔滨150040

出　　处：《东北林业大学学报》2013年第6期168-171,共4页Journal of Northeast Forestry University

基　　金：哈尔滨市科技创新人才研究专项资金项目(2010RFXXG005)

摘　　要：热压过程中人造板板坯内部参数的检测受到工艺状况、环境干扰及传感器自身性能的影响,往往存在较大的误差。采用多传感器技术检测板坯内部温度及压力参数,应用加权最小二乘法准则,对各传感器观测值进行检测级融合,使估值精度高于每个传感器的局部估值;将控制策略设定为目标识别框架,再应用D-S证据理论对温度和压力的概率赋值进行关联融合,从而确定控制策略,提高系统不确定性处理能力。试验结果证明了算法的有效性和可靠性。During hot pressing of the wood-based panels, there are parameter measurement errors because of technological conditions, environment noises and performance of the sensors. Multi-sensor technology was used for mat temperature and pressure measuring. The weighted least square algorithm was applied to acquire high accuracy of overall estimation value by de- termining the information weight coefficient with the corresponding measurement variance of each sensor. The precision of overall estimation value is preferable than the partial estimation of each sensor. The control strategies were organized into a set of target identification frame. According to the basic probability assignment of temperature and pressure, D-S evidence combination principles and reasoning theory were introduced to identify the control strategy of the control system. The resuits show the effectiveness and reliability of the algorithm.

关 键 词：人造板 热压 多传感器数据融合 加权最小二乘法 D—S证据理论 

分 类 号：X791[环境科学与工程—环境工程]