白光LED灯光通量测量不确定度

发布时间：2019-02-18 10:33

【摘要】：随着LED灯的光度测试的测量不确定度在实验室比对的重要性越来越突出,如何正确地评估不确定度成为一大难题。其中不确定度来源与人们对不确定度知识、测试原理、样品、设备和测试方法的理解程度和操作规范性直接相关,有些潜在的较大的不确定度因素难以发现；测量模型不完善,非线性模型以及不确定度因素的强相关加大了不确定度合成的难度；B类不确定度的难度在于确立被测量的误差分布范围和概率分布。因此本论文以白光lpf灯为研究对象,从不确定度来源、数学模型和评估方法三方面探究合理地评估白光LED灯光通量测量不确定度的方法。首先,通过文献调研和测试实践,尽可能地搜集影响因素。本论文主要考虑了能够评估的影响光通量测试的18个不确定度来源。结果发现,主要的不确定度因素只有8个,其他因素的影响是可以忽略的。其次,建立合理的光通量数学模型。本论文结合测试原理,建立的数学模型只包含了标准灯和LED灯的光谱响应。对于未包含在数学模型中的其它因素,要单独考虑其影响,最后把所有的不确定度分量合成到一起。结果发现,此模型把光谱和其它变量分离开来,非常方便不确定度的评估。然后,在评估过程中,关于数学模型中的五个光谱,在每个波长处存在光谱能量误差和波长误差以及相邻波长之间的强相关性,评估难度极大。本论文通过使用三次样条插值法获得整数波长点的光谱,再分别使用ISO《测量不确定度表示指南》(以下简称GUM)和蒙特卡罗法(以下简称MCM)来评估光谱能量误差引入的不确定度分量。同时,LED灯的光谱响应要考虑波长误差的影响。GUM和MCM评估过程复杂,需要使用Matlab编写所有的计算程序。结果发现,对于非线性的光通量数学模型,MCM比GUM的不确定度要小很多,方法更简单和精确。对于数学模型之外的其他因素,仍然使用GUM来评估其不确定度。关于A类不确定度分量,本论文设计了一系列独立重复实验。关于B类不确定度分量,通过查阅证书、文献、借鉴以前的实验数据或者经验,得到误差分布区间,并估计其概率分布。最后,通过制定不确定度预算表,发现标准灯的校准光谱、LED灯的响应、配光修正系数、LED灯响应的波长误差、标准灯的环境温度和测试电流、光谱仪的非线性度和测量误差是主要的不确定度来源。其中光谱仪的测量误差影响最大,其次是LED灯光谱响应的波长误差。在实际测试中,通过合理控制以上因素,最大程度地减小测量不确定度。
[Abstract]:With the increasing importance of the uncertainty in the photometric measurement of LED lamp in laboratory comparison, how to correctly evaluate the uncertainty has become a difficult problem. The source of uncertainty is directly related to people's understanding of uncertainty, test principle, sample, equipment and test method, and some potentially large uncertainty factors are difficult to find. The imperfection of measurement model, the strong correlation of nonlinear model and uncertainty factors increase the difficulty of uncertainty synthesis, and the difficulty of class B uncertainty lies in establishing the range of error distribution and probability distribution. Therefore, this paper takes the white lpf lamp as the research object, and explores the method of reasonably evaluating the uncertainty of the white light lpf light flux measurement from three aspects: the source of uncertainty, the mathematical model and the evaluation method. First of all, through literature research and testing practice, as far as possible to collect the factors. This paper focuses on 18 sources of uncertainty that can be evaluated for the impact of luminous flux testing. The results show that there are only 8 major factors of uncertainty, while the other factors can be neglected. Secondly, a reasonable mathematical model of luminous flux is established. Based on the test principle, the mathematical model only includes the spectral response of standard lamp and LED lamp. For the other factors which are not included in the mathematical model, the influence of the factors should be considered separately, and finally all the uncertainty components should be combined together. It is found that this model separates spectrum from other variables and facilitates the evaluation of uncertainty. Then, in the evaluation process, for the five spectra in the mathematical model, there is a strong correlation between the spectral energy error and wavelength error at each wavelength, as well as the strong correlation between adjacent wavelengths, so it is very difficult to evaluate. In this paper, the spectrum of integer wavelength points is obtained by cubic spline interpolation. Then the uncertainty components introduced by spectral energy error are evaluated by ISO (GUM) and Monte Carlo method (MCM) respectively. At the same time, the spectral response of LED lamp should take into account the influence of wavelength error. The evaluation process of GUM and MCM is complicated, and all calculation programs need to be written with Matlab. The results show that the uncertainty of MCM is much smaller than that of GUM, and the method is simpler and more accurate for the nonlinear luminous flux mathematical model. For factors other than mathematical models, GUM is still used to assess its uncertainty. For class A uncertainty components, a series of independent repeated experiments are designed in this paper. For class B uncertainty components, the error distribution interval is obtained and its probability distribution is estimated by consulting certificate, literature, and previous experimental data or experience. Finally, the calibration spectrum of the standard lamp, the response of the LED lamp, the correction coefficient of the light distribution, the wavelength error of the response of the LED lamp, the ambient temperature of the standard lamp and the test current are found by making the uncertainty budget table. The nonlinearity and measurement error of spectrometer are the main sources of uncertainty. The measurement error of the spectrometer is the most important, followed by the wavelength error of the LED light spectrum response. In the actual test, the uncertainty of measurement is minimized by reasonably controlling the above factors.
【学位授予单位】：北京交通大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TM923.07

【参考文献】