基于多传感器信息融合的LNG液位检测系统设计

发布时间：2018-03-22 01:05

本文选题：液化天然气　切入点：槽车运输　出处：《长安大学》2015年硕士论文　论文类型：学位论文

【摘要】：随着能源危机和环境污染的日益加剧,清洁能源在工业生产中所占的比重日益上升,液化天然气(LNG)由于清洁、安全、高效等特点,逐渐成为石化资源的主角。在LNG运输过程中,实时监测槽车位置和LNG的在途状态是实现安全生产的重要保障。本文以槽车运输途中LNG罐体液面为监测对象,以布局分散的超声波传感器为数据来源,通过单片机电路进行信号调理和信息预处理,实现在多传感器模式下的液位监测目标,最后探索了以卡尔曼滤波为核心的分布式多传感器信息融合技术。本文设计并搭建了LNG液位监测的实验室模拟软硬件平台。首先,利用超声波传感器测得模拟罐体的液位深度,以微小模拟电压信号通过信号调理电路后,以有线方式传给单片机;其次,单片机对数据进行采集、显示、报警以及相关模式设置,同时将数据提交给上位机;最后,利用上位机软件对数据进行曲线显示,再利用信息融合算法进行数据处理,得到较为准确的液位值。另外,本文基于经典卡尔曼滤波理论,将LNG罐体在运输途中的系统视为多传感器时滞模型,并针对可能存在的数据丢失和传输时延,采用按矩阵加权、引入状态增广矩阵等方式,修正卡尔曼滤波器的递推公式,设计分布式矩阵加权融合估计算法,并且给出了任意两个传感器之间的滤波误差互协方差阵的计算公式。通过实验,比较本设计测量结果与实际数据,该系统的精确度较高,实验数据的波动性较小,可以用来监测LNG在槽车运输途中的液位变化,具有一定的理论价值和现实意义。
[Abstract]:With the increasing of energy crisis and environmental pollution, the proportion of clean energy in industrial production is increasing day by day. Liquefied natural gas (LNG) has gradually become the protagonist of petrochemical resources because of its characteristics of cleanness, safety and efficiency. The real-time monitoring of the position of the tank car and the state of LNG in transit is an important guarantee for the safety of production. This paper takes the liquid level of the LNG tank during the transportation of the tank truck as the monitoring object, and takes the distributed ultrasonic sensor as the data source. The signal conditioning and information preprocessing are carried out by single-chip microcomputer circuit to realize the liquid level monitoring target in multi-sensor mode. Finally, the distributed multi-sensor information fusion technology with Kalman filter as the core is explored. In this paper, the software and hardware platform of LNG liquid level monitoring is designed and built. Firstly, the depth of simulated tank is measured by ultrasonic sensor. Small analog voltage signal through the signal conditioning circuit, the way to the single chip microcomputer; secondly, the single-chip microcomputer to collect data, display, alarm and related mode settings, at the same time the data to the host computer; finally, The upper computer software is used to display the curve of the data, and then the information fusion algorithm is used to process the data to get the more accurate liquid level. In addition, based on the classical Kalman filter theory, The system of LNG tank in transit is regarded as a multi-sensor time-delay model. In view of the possible data loss and transmission delay, the recursive formula of Kalman filter is modified by weighting by matrix and introducing state augmented matrix. The distributed matrix weighted fusion estimation algorithm is designed, and the calculation formula of the filtering error mutual covariance matrix between any two sensors is given. The experimental results show that the precision of the system is higher than that of the actual data. The fluctuation of the experimental data is small, so it can be used to monitor the change of liquid level of LNG in the course of tank truck transportation, which has certain theoretical value and practical significance.
【学位授予单位】：长安大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TE978;TP212

【相似文献】