基于光谱发射率函数基形式不变的辐射测温技术

近年来,随着国防、工业、科技等领域飞速发展,无论是对于军用动力发射系统还是对于民用钢铁冶炼以及高科技新兴产业,辐射温度测量都具有重要意义。尤其在温度极高且伴随着瞬态测温(小于1 μs)需求的场合,多光谱辐射测温法被广泛运用。多光谱辐射测温法是通过选取被测目标多个特征波长,测量特征波长的辐射信息,再假设发射率与波长相关的数学模型,最终求解得到辐射温度。目前,利用该方法实际测温时,光谱发射率都采用固定的假设数学模型,而针对目标在不同温度状态下,该固定模型则无法进行自适应变化。同样,在不同温度下,如何解算最终的发射率和辐射温度也没有普适性的方法。基于普朗克黑体辐射定律,提出一种被测目标在不同温度下光谱发射率函数基形式不变的思想,简称发射率函数基形式不变法。通过该方法,发射率模型可以根据物体在不同温度状态下,函数系数动态改变来进行自适应变化。同时对于如何解算最终的发射率和辐射温度也相应提出了普适性的方法。通过大量仿真验证以及实际测量光谱辐射照度标准灯和溴钨灯温度实验,证明本文提出的方法比现有的光谱发射率处理方法更加简单实用并且能够有效地提高光谱发射率的计算精度,从而提高辐射温度测量精度。同时具有实用性好、应用广泛等特点。     

热辐射体真实温度的测试研究

在辐射测温中，普遍存在一个问题，被测温度物体表面发射率影响很大，而物体的发射率很难确测量，这是因为发射率不仅与材料有关，而且还与波长，温度，表面状态（表面粗糙度，氧化程度等）有关，本文叙述一种利用多波长辐射法测量实际物体真实温度的方法，该方法利用最小二乘法原理拟合出实际热辐射体的光谱发射率曲线，从而使测量目标的真实温度成为可能。     

Determining the temperature of an opaque object by its thermal radiation spectrum: forms of initial data presentation and methods

Spectral distributions of intensities, relative emissivity, and inverse radiance temperatures of an opaque, free-radiating object in a condensed state are used as the initial data. The methods of determining the thermodynamic (true) temperature corresponding to these distributions, when object emissivity is previously unknown, are considered. The advantages and disadvantages of each method and corresponding form of an initial data presentation are discussed. It is shown that the spectral distribution of inverse radiance temperatures gives the greatest information about the true temperature and emissivity of the measured object. The estimates of the temperature range to which the true temperature belongs are given based on the known experimental data for tungsten. The methods for additional verification of reliability of the obtained results are presented.     

Determination of temperature and emissivity of opaque heated bodies via thermal radiation spectrum: simulation of measurements in spectral window

Computer simulation of emitted radiation intensity spectrum of tantalum object was carried out. Simulation measurements occurred in a narrow spectral window, which moved along the spectrum with the given step. In this way, spectral ranges at which the dependence of the emissivity (or its logarithm) on the wavelength is the simplest and fairly accurate, in particular linear, were sought for. In the case of successful search for the specified spectral range, the required temperature was determined in the spectral window with the least squares technique. If the emissivity (or its logarithm) is linearly dependent on the wavelength then the alternative estimation of the true temperature is possible. In this case, the required temperature is determined via the change of the convexity of the spectral emissivity dependence at selection of its numerical value from the values lower than the true temperature value to the values higher than the true temperature value. This approach is simple, does not require solution of the system of equations and can significantly narrow temperature range to which the true temperature belongs.     

The measuring of spectral emissivity of object using chaotic optimal algorithm

There exist a considerable variety of factors affecting the spectral emissivity of an object. The authors have designed an improved combined neural network emissivity model, which can identify the continuous spectral emissivity and true temperature of any object only based on the measured brightness temperature data. In order to improve the accuracy of approximate calculations, the local minimum problem in the algorithm must be solved.Therefore, the authors design an optimal algorithm, i.e. a hybrid chaotic optimal algorithm, in which the chaos is used to roughly seek for the parameters involved in the model, and then a second seek for them is performed using the steepest descent. The modelling of emissivity settles the problems in assumptive models in multi-spectral theory.     

Thermal emission of semiconductors under nonequilibrium conditions

The thermal emission of semiconductors under the fundamental absorption edge at excitation of nonequilibrium charge carriers in them has been investigated. A broad spectrum of various actions upon a semiconductor—photoexcitation, contact injection, magnetoconcentration effect—has been used for the first time for modulation of the emissivity of crystals under isothermal conditions. Spectral, field, temperature and coordinate dependencies of the thermal emission as well as its transient characteristics at a pulsed excitation have been measured. It is shown that a considerable modulation of the thermal emission in a semiconductor can be obtained without a temperature change, through changing its emissivity alone. A number of features of the thermal emission in the region under consideration has been revealed: a spectral maximum which position depends on the optical thickness of the crystal, a kinetics delay effect, etc. It has been established that the thermal emission of an excited semiconductor can be used as a means for determining various parameters of a material. Experiments were carried out on Ge, Si plates and Ge diode structures at 310 ⩽ T ⩽ 350 K.     

Multispectral infrared pyrometer for temperature measurement with automatic correction of the influence of emissivity

A new method for contactless temperature measurement has been developed.⁽¹⁾ It is applied advantageously to highly reflecting objects of moderate temperatures (emitting no visible radiation), where conventional methods fail. A set of data is collected by radiance measurement m a number of n narrow spectral bands. These data represent numerical values for a set of analytical equations of the radiation emerging from the object. Solving the set of equations leads to u ⩽ n unknowns, which are the temperature of the object, its emissivity and the temperature of its environment. Balancing calculation enhances the accuracy. Software simulations, including the balancing calculation, the radiation transfer and the hardware of the prospected multispectral radiometer optimized the whole system prior to hardware development. The finally assembled laboratory model, a ten channel filter spectrometer, proved to function as predicted. The temperature of objects about 500 K and an emissivity as low as 0.03 has been determined with less than 1% deviation from thermocouple measurement. Measurement at lower temperatures and/or higher emissivities or higher temperatures and/or lower emissivities is possible with comparable accuracy.     

纯钨红外光谱发射率特性研究

光谱发射率是一个重要的热物性参数,在辐射测温、热传输计算等领域有着广泛的应用。钨作为一种重要的金属,关于其光谱发射率的研究报道较少。利用黑体炉、傅里叶红外光谱仪、加热装置和光学系统搭建了一套能量对比法光谱发射率测量装置,该装置能够测量3~20μm的光谱发射率,测量装置的整体不确定度优于5%。利用该装置测量了纯钨在4个温度点(573, 673, 773和873 K)的法向光谱发射率,重点探讨了氧化、温度、波长和加热时间对纯钨光谱发射率的影响。研究结果表明:纯钨在表面未氧化的情况下,光谱发射率在几个温度点的变化规律基本一致,且数值相差较小,而当其表面发生氧化后光谱发射率迅速增加,在某些波长处出现了强烈的振荡。表面未氧化时纯钨的光谱发射率受温度的影响较小,随着温度的增加仅出现微小的增加,但是当表面发生氧化后,随温度的升高而迅速增大。纯钨的光谱发射率整体上随着波长的增加而减小,但是当表面发生氧化后,由于表面氧化膜与钨金属基底发生干涉效应,在4, 9, 12.5和16.5μm处均出现了峰值。在573和673 K,纯钨的光谱发射率随着加热时间的增加无明显变化。然而,随着温度的升高,在773和873 K时,光谱发射率随着加热时间增加而增大,在773 K时光谱发射率随加热时间的增加增幅较大,因为在该温度点,纯钨表面刚开始发生氧化,氧化速率较大,在873 K时光谱发射率随加热时间的增加增幅较为平缓,并且随着加热时间的增长呈现稳定的趋势。综上,纯钨的光谱发射率在温度较低和表面未氧化时较为稳定。随着温度的升高,当表面发生氧化后,光谱发射率迅速增大,并且在多个波长位置出现了强烈的振荡。由此可见,纯钨光谱发射率受温度、波长、加热时间的影响较大,在实际应用过程中,特别是在辐射测温过程中,如果把纯钨的光谱发射率看做常数将会带来较大的测量误差。该研究将进一步丰富钨的光谱发射率数据,并为其在科学研究和应用中提供数据支持。     

基于约束优化的多光谱辐射真温反演算法

多光谱辐射测温是通过测量待测物某点的多个光谱辐射强度信息,通过普朗克公式反演获得真实温度。但是,通过普朗克公式获得的多光谱辐射测温方程组,是欠定方程组,即N个方程,N+1个未知数(N个未知的光谱发射率ε_λi和1个待求真温T)。目前,多采用事先假设一组发射率模型(发射率-波长或发射率-温度模型),假设模型与实际情况如果相符,则反演结果能够满足要求,如果假设模型与实际情况不符,则反演结果误差很大。但是,发射率模型受温度、表面状态、波长等诸多因素影响,难以事先确定发射率模型。因此受未知光谱发射率的制约一直是多光谱辐射测温理论面临的主要障碍,能否在无需任何光谱发射率假设模型的情况下,实现真温和光谱发射率的直接反演一直是多光谱辐射测温理论研究的热点和难点。通过对参考温度模型的分析表明,多光谱辐射测温反演过程的实质是寻找一组光谱发射率,使得每个通道方程解得的真温都相同,如不相同则继续寻找合适的光谱发射率,直到每个通道解得的真温都相等。为此,提出将多光谱辐射测温参考温度模型的求解过程转换为约束优化问题,即在光谱发射率0≤ε_λi≤1的约束条件下,通过梯度投影算法不断寻找光谱发射率,带入多光谱辐射测温参考温度模型方程组后,计算温度反演值的方差,直到每个光谱通道方程获得的温度值应该近似相等,此时各个光谱通道的温度反演值方差最小,这样就把多光谱辐射真温和发射率的反演问题转换为约束优化问题。约束优化算法是解决这一类问题的主要方法,但为了满足Ax≥b的约束条件,将0≤ε_λi≤1分解为ε_λi≥0和-ε_λi≥-1的两个约束条件,从而满足了约束优化问题Ax≥b的约束条件。这样就可以通过约束优化算法在无需任何光谱发射率假设模型的条件下,直接求解真温和光谱发射率。实验采用六种不同光谱发射率分布模式(随波长递增、递减、凸波动、凹波动、“M”型波动、“W”型波动)的材料为研究对象,以验证新算法对不同材料光谱发射率分布反演的适应性,利用Matlab的minRosen函数,选择光谱发射率的初始值均为0.5(取中间值,提高计算效率)。针对六种不同光谱发射率模型的仿真结果表明,新算法无需任何有关发射率的先验知识,对不同发射率模型反演结果均表现较好,在真温1 800 K的情况下,绝对误差均小于20 K,相对误差均小于1.2%,新算法具有无需考虑任何光谱发射率先验知识、反演精度较高及适合于各种发射率模型等优点,进一步完善了多光谱辐射测温理论,在高温测量领域具有良好的应用前景。     

基于发射率温差模型的多光谱辐射测温理论研究

对于辐射测温数据的处理过程实质就是要解决光谱发射率与真温之间的关系。如果假设的光谱发射率模型与实际光谱发射率模型不符,则会造成较大的测温误差。因此,如何减少光谱发射率和真温对测量模型的依赖程度进而使算法具有一定的通用性,是该领域亟待解决的主要问题之一。提出的算法无需预先假定的发射率与波长之间的模型关系即可找出求解出光谱发射率和真温。通过仿真和实验验证结果表明,使用该算法可以求解出一个相对合理的光谱发射率和满足一定精度要求的真温。算法简单、可靠、具有一定的通用性,适合光谱发射率和真温的测量。     

多目标普朗克极小值优化法的多光谱真温反演研究

光谱发射率是辐射体辐射能力的重要参数,通过光谱发射率可以建立辐射体与黑体的之间的桥梁,从而黑体辐射的相关理论就可以应用于辐射体。采用普朗克公式,光谱高温计的每一个光谱通道可以构成一个方程,这个方程中包含有真温、亮度温度和光谱发射率。对于N个光谱通道可以构成N个方程,这N个方程中也包含一个真温、N个亮度温度和N个光谱发射率,其中亮度温度是已知量,真温和光谱发射率是未知量。由于方程组是欠定的,理论上存在着大量的解。为了求解这个方程组常需要假设光谱发射率与波长和温度之间的数学模型,使方程组未知数的个数降为N个,实现真温的求解。当光谱发射率与波长或温度之间的规律被正确获得后,多光谱辐射测温法才能反演出正确的真温。通过对上述较为常用两种光谱发射率模型的分析可知,这两种方法的基本思想都是试图找到光谱发射率与波长或温度之间的函数关系,确立光谱发射率与波长或温度之间数学模型。用含有波长或温度的表达式代替光谱发射率,实现方程的求解。由于光谱发射率具有一定的不确定性,假设的光谱发射率模型与实际光谱发射率的变化之间存在一定的差异,有可能导致真温反演产生较大的误差。光谱发射率与波长或温度之间的数学模型是需要通过大量的实验和经验才能获得的,而且这种数学模型通用性较差,尤其是当待测辐射体发生改变时,这种数学模型也就失去了意义。为了解决多光谱高温计在实际测量中存在的问题,找到一种无需假定光谱发射率与波长或温度之间数学模型而且又具有一定通用性的多光谱真温反演方法成为一种迫切的需要。为此,将优化的思想引入到了多光谱求解过程中,将多光谱真温的求解问题转化为多目标普朗克极小值优化(MMP)问题,从而不再需要建立光谱发射率与波长或温度之间的数学模型,降低了系统的复杂性与难度。该方法以普朗克公式和光谱发射率之间的等式约束条件为基础,构造了六个目标函数,实现了真温的求解。新方法在反演精度上得到了较大幅度的提高,仿真数据的误差都小于1%。借助于以往的真实测量数据,利用多目标普朗克极小值优化法实现了真温的反演。     

温度和加热时间对大气环境中纯铁光谱发射率的影响

纯铁的光谱发射率受温度的影响很大,尤其是在大气环境中,由于温度升高加剧了表面的氧化,导致其光谱发射率发生了“无规律”变化。基于基尔霍夫定理,利用研制的反射法光谱发射率测量装置对纯铁1.55μm波长的光谱发射率进行了系统的研究,探讨了温度、加热时间等因素对纯铁光谱发射率的影响。研究结果表明：纯铁的光谱发射率随着温度的升高而增大,并且在一定的温度下出现了峰值和谷值,通过分析有氧化层时金属的发射率模型,解释了这种现象的发生。恒温长时间测量结果表明,在不同的温度下,加热时间对光谱发射率的影响不同。研究结果将进一步丰富纯铁的光谱发射率数据,并为其光谱发射率在大气环境中的应用提供了实验依据。     

基于有限立体角测量的多光谱测温

辐射测温以Planck定律为基础通过测量物体表面的发射辐射来反演温度。推导了有限立体角辐射测量条件下的单色测温方程,发现多光谱辐射测温能够实现温度和光谱发射率同时求解通常需满足特定的辐射测量条件：进行微元立体角辐射测量或仅针对漫发射体的有限立体角辐射测量。引入多项式发射率模型,经过数学转化,可以摆脱以上测量限制,得到具有测量普适性的单色测温方程,但却不一定能同时测量光谱发射率。对测温方程组的多解问题进行了初步研究,提出使测量通道数大于待求变量数及采用非线性最小二乘来解决此问题。     

Temperature recovery of opaque bodies by thermal radiation spectrum: the use of relative emissivity to select the optimal spectral range

The approach based on relative emissivity was tested and developed using the experimental data. It was assumed that the medium separating an opaque body and measuring device was diathermic or nonradiating (it is characterized by its transmittance); radiation source emissivity and medium transmittance were unknown. Data on comparison of spectral radiances (spectral intensities), obtained within 220–2500 nm for the temperature lamps in the metrological laboratories of Europe, Russia, and USA were used as the initial experimental data. It is shown that the use of relative emissivity allows graphical interpretation for the solution to the initial nonlinear system of equations. In this case, the problem of determining the true temperature of the body by the thermal radiation spectrum in a graphical interpretation is reduced to the choice depending on relative emissivity at the desired temperature. It is shown that to narrow the interval, which includes the true temperature, the criterion was based on a change in convexity of spectral dependence of the relative emissivity in the process of desired temperature selection. The use of relative emissivity in a spectral range, where the Rayleigh—Jeans approximation is satisfied, allows unambiguous determination for the shape of emissivity dependence on the wavelength. The relationship for determination of the peak wavelength within the registered thermal radiation spectrum on the basis of data about the true temperature of the body and its spectral emissivity is presented.     

Quantitative study of the temperature dependence of normal LWIR apparent emissivity

In a previous work, a method of measurement of apparent emissivity in situ was implemented. This approach has the decisive advantage of being suitable for any commercial infrared systems. It was tested successfully to characterize the normal LWIR apparent emissivity of an aluminium nitride plate in the temperature range [40–550 °C]. Apparent emissivity exhibits a tight temperature dependence. By using the classical model of apparent emissivity and taking into account the spectral emissivity of aluminium nitride ceramic and the spectral response of the IR sensor, we modelled our apparent emissivity measurement with 5% of accuracy and with dispersion better than 1% within the overall temperature range. The effect on the apparent emissivity of both the detection window and the temperature dependence of the spectral emissivity are highlighted.     

白铜热氧化过程中光谱发射率特性研究

随着科技的发展,工业领域对白铜产品质量的要求日益提升;利用辐射测温技术对白铜在冶炼和加工时的温度进行精确测量,是决定产品质量的重要手段,因此研究白铜的光谱发射率特性就显得尤为重要。基于傅里叶红外光谱仪搭建的光谱发射率测量装置,测量了白铜在四个温度点（673,773,873和973 K）,波长范围2～22 μm内的光谱发射率,分别研究了波长、温度、加热时间和氧化对白铜光谱发射率的影响。研究发现,在氮气环境下白铜的光谱发射率随温度的升高而增加,随波长的增加而减少。当白铜暴露在空气环境中,随着温度的升高,其光谱发射率迅速增加。673 K时,白铜表面生成一层细微的氧化物颗粒,阻止白铜进一步氧化,这些氧化物颗粒的光谱发射率大于白铜基底,所以此温度下短波处的光谱发射率略微增加。773 K时,白铜表面氧化物的主要成分是Cu₂O,在实验过程中也观察到白铜表面逐渐变红的现象,这也是白铜在773 K温度下其光谱发射率迅速增加的原因。873 K时,白铜表面氧化物的种类和含量明显增多,氧化膜的厚度满足干涉效应条件,在白铜的光谱发射率曲线中可以明显地观察到干涉极值的演变,随着加热时间的增加,干涉极值逐渐向长波移动。随着温度的升高,白铜的抗氧化能力下降。973 K时,白铜表面的氧化程度最深,在XRD图中氧化物的峰值也最大,因此氧化1 h后由于干涉效应产生的干涉极值数最多。综上所述,波长、温度和氧化对白铜的光谱发射率有重要的影响,在运用辐射测温技术测量白铜温度时应充分考虑上述因素的影响。该研究丰富了白铜的光谱发射率数据,为辐射测温提供了真实可靠的数据支撑。     

Spectral emissivity modeling of steel 201 during the growth of oxidation film over the temperature range from 800 to 1100 K in air

This study explores the spectral emissivity modeling of steel 201 during the growth of oxidation film over the temperature range from 800 to 1100 K at 1.5 μm. The radiance coming from the specimen is received by an InGaAs photodiode detector. The specimen temperature is obtained by averaging the two platinum–rhodium thermocouples, which are tightly welded in the front surface of specimen near the measuring area viewed by the detector. The variation of spectral emissivity with the temperature is studied at a given heating time. The variation of spectral emissivity with the heating time is evaluated at a definite temperature. The strong oscillations of spectral emissivity are observed and discussed in detail, which originate from the interference effect between the radiation stemming from the oxidization film on the specimen surface and the radiation coming from the specimen surface. The measurement uncertainties of spectral emissivity contributed only by the surface oxidization are about 3.2–14.1%. At a given heating time, the variation of spectral emissivity with the temperature abides well by a simple analytic functional form. And at a definite temperature, the variation of spectral emissivity with the heating time can also be well reproduced by fitting except for the periodical oscillations.     

Determination of true temperature of opaque materials via spectral distribution of thermal radiation intensity: application of Wien’s displacement law

The equation for the derivative connecting surface spectral emissivity, wavelength, and thermodynamic (true) temperature of an opaque heated body at the point of spectral maximum of thermal radiation was obtained. It is suggested to solve the problem of determining the true temperature of an opaque surface in two stages. At the first stage, the spectral range, most comfortable for approximation of body emissivity, is distinguished using a special function (relative emissivity), and the true temperature is determined. At the second stage, the true temperature is determined again using the resulting equation for the derivative. The dimensionless parameter that connects the radiative properties of material with the peak wavelength and characterizes deviation from Wien’s displacement law was found. If the absolute value of this parameter is low, the value of true temperature obtained at the first step can be specified at the second stage. This approach is illustrated by experimental data obtained at comparison of spectral radiance of the temperature lamps.     

Experimental study of the relationships between the spectral emissivity of brass and the temperature in the oxidizing environment

Effect of surface oxidization on the spectral emissivity of brass is studied over the temperature range from 800 to 1070 K at the wavelength of 1.5 μm. The temperature of brass surface is measured by averaging the two R-type platinum–rhodium thermocouples. The radiant energy emitted by the brass surface is received by an InGaAs photodiode detector. Two kinds of relationships between the spectral emissivity and the temperature are investigated in the oxidizing environment at the elevated temperature. One is the variation of spectral emissivity with the heating-duration time at the given temperature. The other is the variation of spectral emissivity with the temperature at the given heating-duration time. The interference effect of radiation coming from the brass surface and coming from the oxidization film is discussed when the oxidation film on the surface is grown. The resonant structures of spectral emissivity are observed during the whole heating period, in particular at the early stage of heating duration. The analytic formula of spectral emissivity versus the temperature is derived at the heating-duration time of 30, 60, 90, 120, 150, 180, 210, 240, 270 and 300 min, respectively. The conclusion is obtained that coefficients of analytic expressions between the spectral emissivity and the temperature are different from each other for the experimental results obtained at the different heating-duration time, though the polynomial functional form is suitable to fit all the measurements obtained in the present work.     

Optimization of the width of the working spectral region of a video pyrometer

We have obtained expressions allowing us to estimate the uncertainty in measurement of the temperature by a broadband video pyrometer in the case when the known value of the spectral emissivity of the object applies to the central wavelength of the working spectral region. We show that for monitoring industrial production processes, the working spectral region of the video pyrometer in a number of cases can be broadened up to 100–400 nm, which enables a many-fold increase in its sensitivity and a faster response time.