检索项 检索词
  土壤学报  2021, Vol. 58 Issue (2): 391-400  DOI: 10.11766/trxb201910090434
0

引用本文  

任姮烨, 司炳成, 李敏, 等. 利用非线性模型提高加长型热脉冲双探针土壤热参数测定精度. 土壤学报, 2021, 58(2): 391-400.
REN Hengye, SI Bingcheng, LI Min, et al. Using Non-linear Model to Improve Measurement Accuracy of Soil Thermal Parameters with Extended Dual Probe Heat Pulse Sensor. Acta Pedologica Sinica, 2021, 58(2): 391-400.

基金项目

国家自然科学基金项目(41630860、41601222和41877017)资助

通讯作者Corresponding author

司炳成, E-mail:bing.si@usask.ca

作者简介

任烨, 女, 内蒙古包头人, 硕士研究生, 主要从事土壤物理研究。姮E-mail:hengye@nwafu.edu.cn
利用非线性模型提高加长型热脉冲双探针土壤热参数测定精度
任姮烨1, 司炳成1,2,3, 李敏1, 胡优1    
1. 西北农林科技大学旱区农业水土工程教育部重点实验室, 陕西杨凌 712100;
2. 萨斯喀彻温大学土壤科学系, 萨斯卡通 S7N5A8;
3. 鲁东大学资源与环境工程学院, 山东烟台 264025
摘要:热脉冲双探针技术被广泛应用于土壤热参数和含水率的测定。然而长度为2.8 cm的常规热脉冲双探针空间测试范围太小,加长探针可增加测定结果的代表性,但容易出现探针倾斜(探针间距改变),进而影响测量精度。设计制作了长度为10 cm的热脉冲双探针,通过室内土柱试验测定了四种倾斜方式(共面外倾、非共面外倾、共面内倾以及非共面内倾)下的土壤热参数和含水率,并利用线性和非线性修正模型对探针倾斜引起的测定误差进行了校正。结果表明:对于加长型双探针,非线性模型对探针间距的原位校正效果明显优于线性模型;非线性模型校正后的土壤容积热容和含水率准确度大幅提高。有效解决了加长型热脉冲双探针倾斜引起的测量误差,为探针的广泛应用提供了理论基础。
关键词非线性模型    长探针    热参数    含水率    间距校正    
Using Non-linear Model to Improve Measurement Accuracy of Soil Thermal Parameters with Extended Dual Probe Heat Pulse Sensor
REN Hengye1, SI Bingcheng1,2,3, LI Min1, HU You1    
1. Key Laboratory of Agricultural Soil and Water Engineering in Arid and Semiarid Areas, Ministry of Education, Northwest A & F University, Yangling, Shaanxi 712100, China;
2. Department of Soil Science, University of Saskatchewan, Saskatoon S7N5A8, Canada;
3. School of Resource and Environmental Engineering, University of Ludong, Yantai, Shandong 264000, China
Abstract: 【Objective】The technique of heat-pulse dual probe has been widely used for measuring soil thermal parameters and water content. However, being only 2.8 cm long, the conventional dual probes are very limited in monitoring range. Using lengthened probes may improve representativeness of the measurement, but long probes tend to get deflected, causing deviation of the distance between two probes, thus affecting measurement accuracy. Once the probe gets deflected by 1°the measurement of soil thermal diffusivity and volume heat capacity may deviate by 10%. By measuring temperatures at different points of the probe, spacing error caused by the deflection may be corrected with an in-situ spacing correction linear model or non-linear model. The linear model is preferred because it only requires the probe to have two thermistors rather than three thermistors as the non-linear model does, thus making it easier to assemble and lower in cost because it needs less data acquisition interfaces. However, as to which one, linear or non-linear, is more suitable for long probes, more work should be done.【Method】In this study, 10 cm long probes were designed and used, because 1) in measuring heat-pulses in large soil columns, the probe used by the single probe technique is 10 cm; 2) the thermo-TDR probes longer than 10 cm can improve accuracy of the measurement of soil water contents; and 3) the minimum resolution of the distributed optic fibre temperature measurement is about 10 cm. With development of the technique of thermo-TDR and thermo-distributed temperature sensing with heated fiber optics, the use of 10 cm long probes in this study may better embody the advantages of the above-described technique and cherish a bright future in applying the technique. In the indoor soil column experiment, soil thermal properties and water contents were measured with probes that might deflect off in four ways(coplanar extraversion, non-coplanar- extraversion, coplanar-introversion, non-coplanar-introversion), and errors of the measurements resulting from deflection-caused changes in needle spacing were corrected with the linear or the non-linear model, separately.【Result】In using the extended dual probes, the non-linear model performed better in correcting probe spacing in-situ than the linear model (relative error of the former was -8.30 and of the latter 43.90%). Correction with the non-liner model improved accuracy of the measurement of soil volumetric heat capacities and water contents by a large margin, controlling relative errors within 10%. It is worth noting that the main factor affecting the nonlinear model correcting probe spacing is that the corrected spacing acquired by the non-linear model is determined only determined by tm(the time when the temperature response curve reaches the highest value), while tm is determined jointly by spacing and soil moisture content. Therefore, it is expected to further improve the accuracy of probe spacing correction in using the non-linear model by modifying tm value in the nonlinear model to reflect changes in probe spacing and water content.【Conclusion】This study has effectively solved the problem of measurement errors caused by probe deflection of the extended dual thermal pulse probe, and hence provided a theoretical basis for the wide application of the probe. Compared to the linear model, the non-linear one can effectively minimize spacing errors. With the development of thermo-TDR and thermo-distributed temperature sensing with heated fiber optics, the use of 10 cm-long dual probes may embody advantages of the above-described measuring technique and has broad application prospects.
Key words: Non-linear model    Extended probe    Soil thermal parameters    Soil water content    Probe spacing correction    

热脉冲双探针对土壤破坏小,成本低廉,可连续获得原位数据,被广泛应用于土壤热参数、含水率[1-5]、土壤容重[6]等测定中。常规双探针长度仅为2.8 cm,测定土壤体积有限,不能满足水文、环境和农业监测的现实需求,增加探针长度可增加双探针测量体积,从而增大测量结果代表性,但同时增加探针长度会导致倾斜而产生误差[2, 7-8]

为减少探针倾斜的影响,Kamai等[9]和Satio等[10]通过将探针直径增至4 mm以上使探针更为坚固,而Mori等[11]将探针长度缩短至1.3 cm。但以上两种方法会偏离将有限性探针看作无限线性热源的假设。Liu和Wen等[12-14]通过改进感应探针(内含热感应元件)的设计先后提出了原位间距校正方法的线性模型和非线性模型,并分别采用一阶和二阶函数对探针间距进行校正。测试结果表明利用线性模型可有效减小探针共面倾斜下的土壤容积热容的误差,对于非共面、探针弯曲的倾斜无法准确校正[13, 15]。其优点是感应探针仅包含2个热敏电阻,制作简单,要求的数据采集仪接口数量少,可减小仪器成本,因此,优先选择线性模型校正探针间距变化。而非线性模型在此基础上还可校正非共面倾斜下和探针弯曲引起的测量误差,但要求感应探针中至少安装3个热敏电阻,增加了需要的数据采集仪接口数量和仪器成本。线性模型和非线性模型为增加探针长度从而增加测定体积提供了理论基础,但探针长度越长就越容易倾斜,导致的误差越大,当使用加长型探针时,优先考虑的线性模型是否仍然适用值得进一步探讨。

为了解决上述问题,本研究将双探针加长至10 cm,对于探针倾斜所产生的误差利用Liu等[13-14]提出的理论进行修正,通过对比不同模型校正后的土壤热参数值与理论值,评价了线性模型和非线性模型对于长探针间距误差校正的适用性。本研究将加长型探针长度定为10 cm的原因:(1)测定体积较大的热脉冲单探针长度为10 cm[16];(2)热脉冲-时域反射探针长度超过10 cm时,会提高测定含水率的精度[17]。(3)分布式光纤测温的最小分辨率也在10 cm左右[18]。随着热脉冲-光缆和热脉冲-时域反射技术的发展[13, 19-22],本研究选取的10 cm探针长度是可以更好地体现上述测试技术的优势,具有较广的应用前景。

1 材料与方法 1.1 热脉冲双探针原理

热脉冲双探针由一组间距为r的加热探针和感应探针构成(图 1)。根据热传导理论,在一个无限大的均匀等温介质中,无限线性热源发出t0时长的热脉冲,距热源径向距离r处温度随时间的变化可表达为[1]

$ \left\{ {\begin{array}{*{20}{l}} {\Delta T(r, t) = - \frac{q}{{4\pi \kappa \rho c}}Ei\left( {\frac{{ - {r^2}}}{{4\kappa t}}} \right)\;\;\;{\kern 1pt} t < {t_0}}\\ {\Delta T(r, t) = - \frac{q}{{4\pi \kappa \rho c}}\left( {Ei\left( {\frac{{ - {r^2}}}{{4\kappa \left( {t - {t_0}} \right)}}} \right) - Ei\left( {\frac{{ - {r^2}}}{{4\kappa t}}} \right)} \right)\;\;\;{\kern 1pt} t > {t_0}} \end{array}} \right. $ (1)
图 1 加长型双探针示意图 Fig. 1 Schematic of the apparatus with extended dual probes

式(1)中,ΔT为温度变化值(℃),t为时间(s),q’为单位长度加热丝在单位时间内释放的热量(W·m-1),κ为土壤热扩散系数(m2·s-1),ρc为土壤容积热容(MJ·m-3·K-1),Ei(-x)为指数积分。对式(1)求t的偏微分并使结果等于零,得到κρc为:

$ {\kappa = \frac{{{r^2}}}{4}\left\{ {\frac{{1/\left( {{t_m} - {t_0}} \right) - 1/{t_m}}}{{\ln \left[ {{t_m}/\left( {{t_m} - {t_0}} \right)} \right]}}} \right\}} $ (2)
$ {\rho c = \frac{q}{{4\pi \kappa \Delta {T_m}}}\left\{ {{E_i}\left( {\frac{{ - {r^2}}}{{4\kappa \left( {t - {t_0}} \right)}}} \right) - Ei\left( {\frac{{ - {r^2}}}{{4\kappa t}}} \right)} \right\}} $ (3)

式(2)和式(3)中,tm为最大升温值对应的时间(s)。

根据de Vries模型[4],土壤容积热容表达式为:

$ C = {\rho _b}{c_s} + {\rho _w}{c_w}{\theta _w} $ (4)

式(4)中,ρwcw为水的容积热容量(MJ·m-3·℃-1);θw为土壤体积含水率(m3·m-3);cs为土壤固体的比热(J·kg-1·℃-1),ρb为土壤容重(kg·m-3)。由式(4)可以反推土壤含水率θw的计算公式:

$ {\theta _w} = \frac{{C - {\rho _b}{c_s}}}{{{\rho _w}{c_w}}} $ (5)

纯水在20℃时容积热容量为4.18 MJ·m-3·K-1,土壤固体比热值在干土中使用长探针测定获得,其值为844.35 J·Kg-1·K-1

1.2 原位间距校正方法

双探针测定的容积热容对两支探针的间距很敏感。将热探针插入土壤时,易造成倾斜,从而引起探针间距的改变。考虑倾斜的原位间距校正方法(线性模型和非线性模型)可减小探针间距改变带来的测量误差。当探针发生共面倾斜时,可在感应探针中放置2个热敏电阻并用线性模型校正探针间距;当探针发生非共面倾斜时,则需要在感应探针内放置至少3个热敏电阻并使用非线性模型校正探针间距[12, 23]。其中共面倾斜指倾斜后的加热和感应探针共平面,非共面倾斜则指倾斜后的两支探针不共平面。Wen等[24]进一步将线性模型和非线性模型简化为一个模型并给出了求解方法:

假设感应探针分别距底座距离l1l2l3处放置3个热敏电阻(图 1,以下简称热阻1,热阻2,热阻3;其中,线性模型仅有热阻1和2),定义tmi为热敏电阻i的温度响应曲线达到最大值ΔTm时对应的时间(线性模型:i=1,2;非线性模型:i=1,2,3),并定义ηi为:

$ {\eta _i} = \frac{{\frac{1}{{\left( {{t_{mi}} - {t_0}} \right)}} - \frac{1}{{{t_{mi}}}}}}{{\ln \left[ {\frac{{{t_{mi}}}}{{\left( {{t_{mi}} - {t_0}} \right)}}} \right]}} $ (6)

对于均质土壤,各处热敏电阻热扩散率相同,根据式(2)可得:

$ \frac{{{r_1}}}{{{r_2}}} = \sqrt {\frac{{{\eta _2}}}{{{\eta _1}}}} = p;\;\;\;{\kern 1pt} \frac{{{r_2}}}{{{r_3}}} = \sqrt {\frac{{{\eta _3}}}{{{\eta _2}}}} = {p_2} $ (7)

式中,线性模型只用到p值,非线性模型则会用到pp2值。定义探针未发生倾斜时各热敏电阻与加热探针的间距为初始间距ri0,探针间距偏移量为Δri,则探针发生倾斜后的新间距(校正间距)ri为:

$ \begin{array}{*{20}{l}} {{r_i} = {r_{i0}} + \Delta {r_i}}\\ {\Delta {r_i} = a{l_i} + bl_i^2} \end{array} $ (8)

外倾Δri为正值,内倾反之;ab值为回归系数,对于线性模型,b=0。对于共面倾斜,可利用线性模型校正(b=0),非共面倾斜则需利用非线性模型校正(b≠0),ab值的求解过程参考Wen等[24]。求得的ab值代入式(8)则可获得倾斜后的校正间距(线性模型:r1r2;非线性模型:r1r2r3)。

1.3 试验装置

加长型热脉冲双探针的长度为10 cm,外径为1.3 mm,内径为0.9 mm,材料为304不锈钢。探针设计应符合一定规则,Blackwell等[25]指出探针长与半径比值大于25,Liu等[26]建议热敏电阻距离探针顶部与底部的距离应大于1 cm,Kluitenberg等[27]提出探针建议不宜过大(< 8 mm),否则需要较大的功率加热,易造成水分蒸发。本文采用探针,长度半径比为77,间距为6 mm,符合上述要求。加热探针内置镍铬绝缘电阻丝(阻值为86 Ω·m-1)。感应探针中有3个热敏电阻(10K3MCD1,Betatherm Corp.,Shrewbury MA),为满足线性热源模型,分别放置在距底座距离l1 = 70 mm,l2 = 50 mm,l3=30 mm的位置。并在两个探针内填充高导热性能的环氧树脂(Omega engineering,Stamford,CT)。两根探针通过环氧树脂胶固定在圆柱形PVC底座(壁厚5 mm,直径30 mm,高35 mm),组成热脉冲探头(图 1)。热脉冲探头固定在PVC土柱底部(高13.4 cm,外径5.5 cm,壁厚0.5 cm)。室内恒温条件下(20℃),利用浓度为5 g·L-1的琼脂溶液标定双探针初始间距(图 1),其中琼脂溶液的容积热容与去离子水相同,为4.18 MJ·m-3·K-1。双探针与数据采集仪(Model CR1000,Campbell Scientific,Logan,UT)采用四线半桥的方法连接。为保持探针加热电流稳定,数据采集仪和探针电阻丝分别采用独立的12V恒压蓄电池供电。热脉冲时长t0 = 15 s,加热功率q'≈W·m-1,采样频率1 Hz。

1.4 试验过程

研究表明双探针测定土壤水热参数和原位间距校正方法的规律不受土壤质地的影响[1, 12, 14, 24, 28],本研究选择砂土作为供试土样。土样取自陕西杨凌渭河沿岸,其黏粒(< 0.002 mm)、粉粒(0.002~0.05 mm)和砂粒(> 0.05 mm)的质量分数分别为0.03、0.09和98.8 kg·kg-1。供试土壤进行过筛(1 mm)和烘干处理后,称取所需质量的烘干砂土,加入一定质量的去离子水,配成质量含水率梯度为0、5%、10%、15%、20%的土壤,按目标容重1.60 g·cm-3装于PVC土柱并于恒温条件下使土壤水分充分平衡。测定过程控制室内环境温度温差小于±1℃。

为满足3个热敏电阻处热扩散率相等的前提假设,将热脉冲双探针和土柱平行于水平面横向布置,以减小重力作用带来土壤水分分布不均匀的影响。如图 2在土样中人为模拟探针在田间土壤里可能出现的四种倾斜方式(图 2),共面外倾(OD1)、非共面外倾(OD2)、共面内倾(ID1)和非共面内倾(ID2)。人为倾斜以聚氨酯橡胶材料控制,因其用量极小,其对测量结果的影响可忽略不计。倾斜后3个热敏电阻处探针校正间距见表 1。以上土柱试验对每种倾斜方式(OD1、OD2、ID1和ID2)设4个重复(独立填装4次),每个重复加热3次。

注:OD1为共面外倾、OD2为非共面外倾、ID1为共面内倾和ID2为非共面内倾Note:OD1,OD2,ID1 and ID2 stands for coplanar-extraversion,non-coplanar extraversion,coplanar-introversion and non-coplanar-introversion,respectively 图 2 不同倾斜方式示意图 Fig. 2 Schematic of different ways of deflection

表 1 利用非线性模型计算得到不同倾斜方式倾斜后的新间距 Table 1 New probe spacings of different types of deflection corrected by the non-linear model
1.5 数据处理和误差分析

为了分析线性模型和非线性模型对于加长型探针的适用性,利用线性模型和非线性模型分别校正上述干土土壤中四种倾斜方式(OD1、OD2、ID1、ID2)下的探针间距,将以新间距获得的土壤比热作为实测值并与理论值进行相对误差分析,理论值为探针未倾斜时测定的土壤比热。相对误差(ER)可定义为:

$ {\rm{ER}} = \frac{{{x_{实测值}} - {x_{\rm{理论值}}}}}{{{x_{\rm{理论值}}}}} \times 100\% $ (11)

式(11)中,ER代表相对误差,x实测值代表实测值,x理论值代表理论值。

为验证长探针测定土壤热参数的可行性,利用适用于长探针的校正模型校正不同含水率下的土壤在探针外倾和内倾时的探针间距,以校正的新间距拟合土壤容积热容和热扩散率,二者的乘积即为土壤热导率。感应探针共有三个热敏电阻,每个热敏电阻位置均可得到新的校正间距并拟合得到相应的容积热容和热导率,取这三个新间距拟合的容积热容和热导率的平均值作为土壤容积热容和热导率的实测值。土壤容积热容的理论值则可根据式(4)计算,其中土壤容重为填装容重,含水率由烘干法获得,土壤的固体比热为利用长探针测定的探针未倾斜时的土壤比热值。土壤热导率的测定不受探针间距和长度的影响[29-31],以初始间距拟合计算的热导率作为理论值。为验证长探针测定含水率的可行性,将土壤含水率实测值与烘干法测定的含水率进行比较,实测土壤含水率根据式(5)计算,其中土壤容积热容为新间距拟合得到的容积热容。本研究通过Matlab内置的solve函数和lsqcurvefit函数计算校正后的探针间距和非线性拟合热参数。

2 结果与讨论 2.1 长针原位间距校正模型的选择

图 3代表外倾(OD2)时热阻1、热阻2和热阻3在干土中的温度响应曲线。可见探针外倾时,随着热阻1、热阻2和热阻3的探针间距依次减小(图 2),温度变化达最大值的时间tm值依次减小,温度变化最大值ΔTm值依次增大。这是由于热量在相同的土壤介质中的传播速度相同,探针间距越小,温度上升就越快。探针间距分别使用线性模型和非线性模型校正后的新间距(表 2),结合相应的实测温度响应曲线,以最小二乘法拟合,获得相应的拟合曲线并得到容积热容和热扩散率。使用非线性模型计算的新间距(表 2)结合热阻1、热阻2和热阻3的实测温度响应曲线拟合热传导方程(公式(1))得到了图 3中的拟合曲线,残差平方和结果见表 2。线性模型计算的新间距结合实测温度响应曲线进行拟合也可获得相似的拟合曲线,残差平方和结果见表 2。对于相同的实测温度响应曲线,不同的新间距拟合热传导方程(公式(1))得到的拟合曲线相似,校正得到的新间距不同,则所得的容积热容和热扩散率不同(表 2)。以上以OD2为例,OD1、ID1和ID2的相应结果也很类似。

图 3 温度响应曲线和拟合曲线 Fig. 3 Measured curves and fitting curves of temperature

表 2 校正探针外倾(OD2)的间距和拟合的热参数 Table 2 Corrected probe spacings and fitted thermal parameters for probes deflected outwardly(OD2)

表 3可知,对于共面倾斜(OD1、ID1),未校正时干土比热相对误差为33.94%,经非线性和线性模型校正探针间距后,干土比热相对误差分别为-8.30%和43.90%。对于非共面倾斜(OD2、ID2),未校正时干土比热相对误差为32.65%,经非线性模型和线性模型校正探针间距后,干土比热相对误差分别为6.07%和31.94%。上述结果表明,无论是共面倾斜还是非共面倾斜,经非线性模型校正间距后,比热测定准确度大幅提高,而线性模型未能明显减小加长型探针的间距误差,说明加长型探针的间距校正宜用非线性模型,而不宜用线性模型。这是由于线性模型只能校正未发生弯曲的探针在共面倾斜情况下的间距误差。而对于加长型探针,加热探针和感应探针在被人为控制发生共面倾斜的过程中,不可避免地会发生轻微弯曲,从而影响线性模型的校正效果,而非线性模型则可有效地考虑轻度弯曲,有效减小探针弯曲时比热测定误差。综上,非线性模型校正可明显减少加长型双探针弯曲造成的间距误差。

表 3 线性模型校正、非线性模型校正间距误差和未校正间距的比热(cs)相对误差 Table 3 Relative errors in heat capacity with spacing error corrected with the linear model, non-linear model
2.2 非线性模型校正后土壤热参数与理论值比较

由2.1可知,相比线性模型,非线性模型更适于长探针的间距校正,故下文只考虑使用非线性模型校正长探针测定土壤热参数和含水率的间距误差。探针倾斜对于土壤热导率的测试结果影响很小[30]图 4是长探针实测的经非线性模型校正的土壤热导率和理论值的比较,可以看出,无论探针处于外倾还是内倾,长探针测定的土壤热导率与理论值基本位于1︰1线上(R2为0.996),说明利用非线性模型校正,可以获得准确的土壤热导率。

图 4 非线性模型校正后的热导率与理论值的比较 Fig. 4 Comparison of thermal conductivities acquired and corrected with the non-linear model with theoretical thermal conductivities

图 5可以看出,探针间距变化给土壤容积热容测定带来较大误差(未校正时外倾和内倾的RMSE分别为0.152 MJ·m-3·K-1和0.766 MJ· m-3· K-1)。非线性模型修正后的土壤容积热容准确度大幅提高(外倾和内倾的RMSE分别为0.136 MJ·m-3·K-1和0.173 MJ·m-3·K-1,图中仅列出非线性模型校正后的结果)。未校正间距误差时,探针外倾会高估容积热容,探针内倾结果相反,这与Wen等[14]对常规短探针的研究结果一致。内倾未校正结果偏离1︰1线程度较高,这是由于内倾倾斜程度更大。经过校正后,内倾和外倾校正在高含水率范围(>0.2 m3·m-3)会高估土壤容积热容(外倾平均相对误差+6.8%,内倾相应值为+8.7%);低含水率范围(<0.2 m3·m-3)校正的土壤容积热容与理论值相比较整体表现为低估(外倾平均相对误差为-5.8%,内倾相应值为-6.1%),低含水率范围的校正结果略优于高含水率范围。分析其原因在于:第一,根据非线性模型,校正后的新间距只由温度到达最大值的时间tm决定,而tm值却由间距和土壤含水率共同决定。含水率增加或探针间距减小,均导致tm值减小,但非线性模型中的tm值难以体现间距和含水率的双重变化,当含水率较高时非线性模型校正会低估探针间距,从而造成土壤容积热容的高估,含水率较低时则相反。第二,探针长度越长,初始间距校准的不确定性更大[32],长探针会引起更大的初始间距测定误差,从而影响容积热容的准确测定。第三,热扩散率相等是精确校正容积热容间距误差的前提。使用长探针时应尽量避免土壤水分重力作用,保证三个热敏电阻处含水率相同,以满足三个热敏电阻热扩散率相等。由于土壤含水率较低时受重力影响较小,故受低含水率范围的校正结果优于高含水率范围。

图 5 非线性模型校正间距和未校正间距容积热容计算结果 Fig. 5 Volumetric heat capacities acquired with probe spacing corrected or not with the non-linear model
2.3 非线性模型校正后的土壤含水率

土壤质地和容重一定时双探针测定的土壤含水率取决于土壤容积热容。图 6说明长探针测定的土壤含水率和容积热容的规律基本一致,探针间距变化导致较大土壤含水率测定误差(外倾和内倾的RMSE分别为0.035 m3·m-3和0.188 m3·m-3),但经非线性模型修正后,土壤含水率的测量精度提高(外倾和内倾的RMSE分别为0.034 m3·m-3和0.038 m3·m-3,图中仅列出非线性模型校正后的结果)。

图 6 非线性模型校正间距和未校正间距土壤含水率计算结果 Fig. 6 Soil water contents acquired with probe spacing corrected or not with the non-linear model
2.4 测试结果的不确定性

测试结果的不确定性来自四个方面:第一个方面是,热源本身造成的不确定性有热源尺寸、加热功率、探针倾斜、电阻随温度变化等。第二个方面是土壤的不均匀性、各向异性和导热系数变化的多解性,主要有土壤矿物成分、密度、颗粒大小、孔隙比、孔隙分布特征、有机质含量等。第三个方面是由于确定温度上升曲线到达最大值的时间tm时存在误差,tm值受到信噪比、热敏电阻的测定精度及响应速度的影响。第四个方面,原位间距校正模型基于不同热敏电阻处土壤的热扩散率相等的前提假设,因此土壤水分分布的均匀程度对测定结果也会产生一定的影响。

3 结论

本研究基于原位间距校正方法,将热脉冲双探针的长度加长至10 cm,人为进行探针倾斜,并对砂土进行长针水热参数的测定,以评价线性模型和非线性模型对于长探针的适用性。结果表明,无论是共面倾斜还是非共面倾斜,非线性模型能够明显提高长针测定土壤比热的准确性,相对误差为-8.30%;而线性模型校正效果不明显,相对误差达43.90%,因此,非线性模型适于长探针,而线性模型对于长探针间距校正存在一定局限性。探针间距变化对土壤热导率测定无影响,但给土壤容积热容和含水率测定带来较大误差;使用非线性模型修正后,长探针测定土壤容积热容和含水率的准确度大幅提高,相对误差在10%以内。当前影响非线性模型校正长探针间距误差的主要因素是:非线性模型得到的校正间距由温度响应曲线到达最大值的时间tm决定,而tm值却由间距和土壤含水率共同决定。因此,修正非线性模型中的tm值以体现间距和含水率的双重变化有望进一步提高非线性模型校准探针间距的精度。

参考文献
[1]
Bristow K L, Kluitenberg G J, Horton R. Measurement of soil thermal properties with a dual-probe heat-pulse technique[J]. Soil Science Society of America Journal, 1994, 58(5): 1288-1294. DOI:10.2136/sssaj1994.03615995005800050002x (0)
[2]
Campbell G S, Calissendorff C, Williams J H. Probe for measuring soil specific heat using a heat-pulse method[J]. Soil Science Society of America Journal, 1991, 55(1): 291-293. DOI:10.2136/sssaj1991.03615995005500010052x (0)
[3]
Li Y, Shao M A. Latest advance of thermo-pulse method for measuring soil thermal properties (In Chinese)[J]. Acta Pedologica Sinica, 2005, 42(1): 134-139. DOI:10.3321/j.issn:0564-3929.2005.01.021 [李毅, 邵明安. 热脉冲法测定土壤热性质的研究进展[J]. 土壤学报, 2005, 42(1): 134-139.] (0)
[4]
de Vries D A. Thermal properties of soils//van Wijk.Physics of plant environment . Amsterdam: North-Holland Publishing Company, 1963: 210-235. (0)
[5]
Liu Z P, Xu J N, She D L, et al. Effects of biochar addition on thermal properties of loamy soil (In Chinese)[J]. Acta Pedologica Sinica, 2018, 55(4): 933-944. [刘志鹏, 徐杰男, 佘冬立, 等. 添加生物质炭对壤土热性质影响机理研究[J]. 土壤学报, 2018, 55(4): 933-944.] (0)
[6]
Ren T S, Shao M A, Ju Z Q, et al. Measurement of soil physical properties with Thermo-Time Domain Reflectometry Ⅱ. Application (In Chinese)[J]. Acta Pedologica Sinica, 2004, 41(4): 523-529. DOI:10.3321/j.issn:0564-3929.2004.04.005 [任图生, 邵明安, 巨兆强, 等. 利用热脉冲时域反射技术测定土壤水热动态和物理参数Ⅱ.应用[J]. 土壤学报, 2004, 41(4): 523-529.] (0)
[7]
Liu G, Li B G, Ren T S, et al. Analytical solution of heat pulse method in a parallelepiped sample space with inclined needles[J]. Soil Science Society of America Journal, 2008, 72(5): 1208-1216. DOI:10.2136/sssaj2007.0260 (0)
[8]
Bristow K L, Campbell G S, Calissendorff K. Test of a heat-pulse probe for measuring changes in soil water content[J]. Soil Science Society of America Journal, 1993, 57(4): 930-934. DOI:10.2136/sssaj1993.03615995005700040008x (0)
[9]
Kamai T, Tuli A, Kluitenberg G J, et al. Soil water flux density measurements near 1 cm d–1 using an improved heat pulse probe design[J]. Water Resources Research, 2008, 44(4): 155-166. (0)
[10]
Saito H, Šimůnek J, Hopmans J W, et al. Numerical evaluation of alternative heat pulse probe designs and analyses[J]. Water Resources Research, 2007, 43(7): 88-101. (0)
[11]
Mori Y, Hopmans J W, Mortensen A P, et al. Multi-functional heat pulse probe for the simultaneous measurement of soil water content, solute concentration, and heat transport parameters[J]. Vadose Zone Journal, 2003, 2(4): 561-571. DOI:10.2136/vzj2003.5610 (0)
[12]
Liu G, Wen M, Chang X, et al. A self-calibrated dual probe heat pulse sensor for in situ calibrating the probe spacing[J]. Soil Science Society of America Journal, 2013, 77(2): 417-421. DOI:10.2136/sssaj2012.0434n (0)
[13]
Wen M M, Liu G, Horton R, et al. An in situ probe-spacing-correction thermo-TDR sensor to measure soil water content accurately[J]. European Journal of Soil Science, 2018, 69(6): 1030-1034. DOI:10.1111/ejss.12718 (0)
[14]
Wen M M, Liu G, Li B G, et al. Evaluation of a self-correcting dual probe heat pulse sensor[J]. Agricultural and Forest Meteorology, 2015, 200: 203-208. DOI:10.1016/j.agrformet.2014.09.022 (0)
[15]
Qiao Z Y, Yuan X H, Liu G. Application of spacing correction method of heat pulse probe in field measurements (In Chinese)[J]. Acta Pedologica Sinica, 2018, 55(1): 122-130. [乔照钰, 原晓辉, 刘刚. 野外试验中热脉冲探针间距校正方法的应用[J]. 土壤学报, 2018, 55(1): 122-130.] (0)
[16]
Li M, Si B C, Hu W, et al. Single-probe heat pulse method for soil water content determination:Comparison of methods[J]. Vadose Zone Journal, 2016, 15(7): 25-37. (0)
[17]
Dalton F N, van Genuchten M. The time-domain reflectometry method for measuring soil water content and salinity[J]. Geoderma, 1986, 38(1/2/3/4): 237-250. (0)
[18]
Sayde C, Buelga J B, Rodriguez-Sinobas L, et al. Mapping variability of soil water content and flux across 1-1000 m scales using the Actively Heated Fiber Optic method[J]. Water Resources Research, 2014, 50(9): 7302-7317. DOI:10.1002/2013WR014983 (0)
[19]
He H L, Dyck M F, Horton R, et al. Distributed temperature sensing for soil physical easurements and its similarity to heat pulse method . Amsterdam: Elsevier, 2018: 173-230. (0)
[20]
He H L, Dyck M F, Horton R, et al. Development and application of the heat pulse method for soil physical measurements[J]. Reviews of Geophysics, 2018, 56(4): 567-620. DOI:10.1029/2017RG000584 (0)
[21]
Hu Y, Li M, Ren H Y, et al. Measurement of soil water content using distributed temperature sensor with heated fiber optics (In Chinese)[J]. Transactions of the Chinese Society of Agricultural Engineering, 2019, 35(10): 42-49. [胡优, 李敏, 任姮烨, 等. 基于加热光纤分布式温度传感器的土壤含水率测定方法[J]. 农业工程学报, 2019, 35(10): 42-49.] (0)
[22]
Cao D F, Fang H Y, Wang F M, et al. A fiber bragg-grating-based miniature sensor for the fast detection of soil moisture profiles in highway slopes and subgrades[J]. Sensors, 2018, 18(12): 4431-4446. DOI:10.3390/s18124431 (0)
[23]
Liu G, Wen M M, Ren R Q, et al. A general in situ probe spacing correction method for dual probe heat pulse sensor[J]. Agricultural and Forest Meteorology, 2016, 226/277: 50-56. (0)
[24]
Wen M, Liu G, Horton R. Simplified computational approach for determining in situ probe spacing of a dual Probe heat pulse sensor[J]. Soil Science Society of America Journal, 2018, 82(5): 1113-1116. DOI:10.2136/sssaj2017.11.0402 (0)
[25]
Blackwell J H. The axial-flow error in the thermal- conductivity probe[J]. Canadian Journal of Physics, 1956, 34(4): 412-417. DOI:10.1139/p56-048 (0)
[26]
Liu G, Li B G, Ren T S, et al. Analytical solution of the heat pulse method in a parallelepiped sample space[J]. Soil Science Society of America Journal, 2007, 71(5): 1607-1619. DOI:10.2136/sssaj2006.0390 (0)
[27]
Kluitenberg G J, Ham J M, Bristow K L. Error analysis of the heat pulse method for measuring soil volumetric heat capacity[J]. Soil Science Society of America Journal, 1993, 57(6): 1444-1451. DOI:10.2136/sssaj1993.03615995005700060008x (0)
[28]
Ren T S, Ochsner T E, Horton R. Development of thermo-time domain reflectometry for vadose zone measurements[J]. Vadose Zone Journal, 2003, 2(4): 544-551. DOI:10.2136/vzj2003.5440 (0)
[29]
Kluitenberg G J, Das B S, Bristow K L. Error analysis of heat pulse method for measuring soil heat capacity, diffusivity, and conductivity[J]. Soil Science Society of America Journal, 1995, 59(3): 719-726. DOI:10.2136/sssaj1995.03615995005900030013x (0)
[30]
Kluitenberg G J, Kamai T, Vrugt J A, et al. Effect of probe deflection on dual-probe heat-pulse thermal conductivity measurements[J]. Soil Science Society of America Journal, 2010, 74(5): 1537-1540. DOI:10.2136/sssaj2010.0016N (0)
[31]
Noborio K, McInnes K J, Heilman J L. Measurements of soil water content, heat capacity, and thermal conductivity with a single TDR probe1[J]. Soil Science, 1996, 161(1): 22-28. DOI:10.1097/00010694-199601000-00004 (0)
[32]
Ham J M, Benson E J. On the construction and calibration of dual-probe heat capacity sensors[J]. Soil Science Society of America Journal, 2004, 68(4): 1185-1190. DOI:10.2136/sssaj2004.1185 (0)