基于SA-ELM的声学层析成像温度分布重建算法

发布时间：2018-03-20 03:07

本文选题：声学层析成像　切入点：温度分布测量　出处：《化工学报》2017年06期 　论文类型：期刊论文

【摘要】：在化工领域中,获取准确的温度分布信息具有极其重要的意义。声学层析成像(AT)具有非侵入传感、廉价等优点,因而被认为是一种具有广阔发展前景的可视化温度分布测量方法。将AT用于温度分布测量,提出一种SA-ELM算法改进重建质量。首先,利用稳健估计建立了基于L1范数的目标泛函,采用模拟退火算法(SA)对目标泛函进行求解,得到粗网格下的温度分布;最后,采用极限学习机(ELM)来预测经细化网格后的温度分布。数值仿真和实验研究途径评估该方法的可行性与有效性,结果表明SA-ELM算法能够有效提高温度分布重建质量和鲁棒性,从而为AT反问题的求解提供了一种新的有效方法。
[Abstract]:In the field of chemical industry, it is very important to obtain accurate temperature distribution information. Acoustic Tomography (ATT) has the advantages of non-invasive sensing, low cost, and so on. As a result, it is considered to be a visual temperature distribution measurement method with broad development prospects. Using AT to measure temperature distribution, a SA-ELM algorithm is proposed to improve the reconstruction quality. Firstly, a target functional based on L1 norm is established by robust estimation. The simulated annealing algorithm (SAA) is used to solve the target functional, and the temperature distribution under coarse mesh is obtained. An extreme learning machine (ELM) is used to predict the temperature distribution after mesh refinement. The feasibility and effectiveness of the method are evaluated by numerical simulation and experimental research. The results show that SA-ELM algorithm can effectively improve the quality and robustness of temperature distribution reconstruction. This provides a new and effective method for solving AT inverse problem.
【作者单位】：华北电力大学控制与计算机工程学院;华北电力大学能源动力与机械工程学院;
【基金】：高等学校学科创新引智计划项目(B13009) 中央高校基本科研业务费专项资金(2016XS42)~~
【分类号】：TQ02

【相似文献】