Uncertainty analysis of heat flux measurements estimated using a one-dimensional, inverse heat-conduction program.
Uncertainty analysis of heat flux measurements estimated using a one-dimensional, inverse heat-conduction program.
复制标题
使用一维逆热传导程序估计的热通量测量的不确定性分析。
DOI：
--
复制
发表时间：
2005
期刊：
影响因子：
--
通讯作者：
J. E. Murphy
中科院分区：
文献类型：
--
作者：
J. Nakos;V. Figueroa;J. E. Murphy
烃类燃料火灾中热通量的测量（例如，柴油或JP-8）由于高温和潮湿的环境而很困难。未冷却的市售热通量计不能在长时间火灾中存活，并且冷却的热通量计经常被烟灰覆盖，从而改变热通量计校准。另一种耐用且相对便宜的方法是基于逆热传导方法。逆热传导方法使用温度/时间历史、边界条件、材料性质以及通常一维（1-D）热流假设来估计特定材料界面处的吸收热通量。这种方法通常用于桑迪亚的防火测试设施。在本报告中，针对一个具体示例进行了不确定性分析，以量化使用逆热传导法时输入参数变化对估算热通量的影响。所采用的方法是将使用修改后的输入的若干案例的结果与基本案例进行比较。304不锈钢圆柱体[约30.5厘米（12英寸）]的响应直径0.32厘米厚（1/8英寸）]填充2.5厘米厚（1英寸）检查陶瓷纤维绝缘。逆热传导程序的输入参数变化是钢壁厚度，导热系数，和体积热容;绝缘厚度，导热系数，和更多»体积热容，温度不确定性，边界条件，温度采样周期;和数值输入。在所有情况下都假设一维热传递。分析结果表明，在最大热流密度下，最重要的参数是温度不确定性、钢厚度和钢体积热容。使用恒定的热特性而不是温度相关值也会导致所得热通量的显著差异;因此，应使用温度相关值。作为一个例子，改变几个参数来估计热通量的不确定性。在最高通量下，结果为15-19%的不确定性和95%的置信度，忽略多维效应。«少
The measurement of heat flux in hydrocarbon fuel fires (e.g., diesel or JP-8) is difficult due to high temperatures and the sooty environment. Un-cooled commercially available heat flux gages do not survive in long duration fires, and cooled gages often become covered with soot, thus changing the gage calibration. An alternate method that is rugged and relatively inexpensive is based on inverse heat conduction methods. Inverse heat-conduction methods estimate absorbed heat flux at specific material interfaces using temperature/time histories, boundary conditions, material properties, and usually an assumption of one-dimensional (1-D) heat flow. This method is commonly used at Sandia.s fire test facilities. In this report, an uncertainty analysis was performed for a specific example to quantify the effect of input parameter variations on the estimated heat flux when using the inverse heat conduction method. The approach used was to compare results from a number of cases using modified inputs to a base-case. The response of a 304 stainless-steel cylinder [about 30.5 cm (12-in.) in diameter and 0.32-cm-thick (1/8-in.)] filled with 2.5-cm-thick (1-in.) ceramic fiber insulation was examined. Input parameters of an inverse heat conduction program varied were steel-wall thickness, thermal conductivity, and volumetric heat capacity; insulation thickness, thermal conductivity, andmore » volumetric heat capacity, temperature uncertainty, boundary conditions, temperature sampling period; and numerical inputs. One-dimensional heat transfer was assumed in all cases. Results of the analysis show that, at the maximum heat flux, the most important parameters were temperature uncertainty, steel thickness and steel volumetric heat capacity. The use of a constant thermal properties rather than temperature dependent values also made a significant difference in the resultant heat flux; therefore, temperature-dependent values should be used. As an example, several parameters were varied to estimate the uncertainty in heat flux. The result was 15-19% uncertainty to 95% confidence at the highest flux, neglecting multidimensional effects.« less