当前位置: X-MOL 学术Rev. Sci. Instrum. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Decreasing the uncertainty of classical laser flash analysis using numerical algorithms robust to noise and systematic errors
Review of Scientific Instruments ( IF 1.3 ) Pub Date : 2020-06-01 , DOI: 10.1063/1.5132786
Artem Lunev 1 , Robert Heymer 1
Affiliation  

The laser flash method is highly regarded due to its applicability to a wide temperature range, from cryogenic temperatures to the melting point of refractory metals, and to extreme environments involving radioactive or hazardous materials. Although instruments implementing this method are mostly produced on a commercial basis by major manufacturers, there is always room for improvement both in terms of experimental methods and data treatment procedures. The measurement noise, either due to the detector performance or electromagnetic interferences, presents a significant problem when accurate determination of thermal properties is desired. Noise resilience of the laser flash method is rarely mentioned in the published literature; there are currently no data treatment procedures that could guarantee adequate performance under any operating conditions. In this paper, a computational framework combining finite-difference solutions of the heat conduction problem with nonlinear optimization techniques based on the use of quasi-Newton direction search and stochastic linear search with the Wolfe conditions is presented. The application of this framework to data with varying level of noise is considered. Finally, cross-verification and validation using an external standard, a commercial, and an in-house built laser flash instrument are presented. The open-source software implementing the described computational method is benchmarked against its industrial counterpart.

中文翻译:

使用对噪声和系统误差具有鲁棒性的数值算法降低经典激光闪光分析的不确定性

激光闪光法因其适用于广泛的温度范围而备受推崇,从低温到难熔金属的熔点,以及涉及放射性或危险材料的极端环境。虽然实施这种方法的仪器大多由主要制造商在商业基础上生产,但在实验方法和数据处理程序方面总是有改进的余地。当需要准确确定热特性时,由于检测器性能或电磁干扰引起的测量噪声会带来重大问题。激光闪光法的抗噪性在已发表的文献中很少提及;目前没有任何数据处理程序可以保证在任何操作条件下都有足够的性能。在本文中,基于使用拟牛顿方向搜索和具有沃尔夫条件的随机线性搜索,提出了一种将热传导问题的有限差分解与非线性优化技术相结合的计算框架。考虑将该框架应用于具有不同噪声水平的数据。最后,介绍了使用外部标准、商用和内部构建的激光闪光仪器进行的交叉验证和验证。实现所描述的计算方法的开源软件与其工业对应物进行了基准测试。提出了一种计算框架,它将热传导问题的有限差分解与基于使用拟牛顿方向搜索和沃尔夫条件的随机线性搜索的非线性优化技术相结合。考虑将该框架应用于具有不同噪声水平的数据。最后,介绍了使用外部标准、商用和内部构建的激光闪光仪器进行的交叉验证和验证。实现所描述的计算方法的开源软件与其工业对应物进行了基准测试。提出了一种计算框架,它将热传导问题的有限差分解与基于使用拟牛顿方向搜索和沃尔夫条件的随机线性搜索的非线性优化技术相结合。考虑将该框架应用于具有不同噪声水平的数据。最后,介绍了使用外部标准、商用和内部构建的激光闪光仪器进行的交叉验证和验证。实现所描述的计算方法的开源软件与其工业对应物进行了基准测试。介绍了使用外部标准、商业和内部构建的激光闪光仪器进行的交叉验证和验证。实现所描述的计算方法的开源软件与其工业对应物进行了基准测试。介绍了使用外部标准、商业和内部构建的激光闪光仪器进行的交叉验证和验证。实现所描述的计算方法的开源软件与其工业对应物进行了基准测试。
更新日期:2020-06-01
down
wechat
bug