- Email: [email protected]
Comparative performance assessments of surface junction probes for stagnation heat flux estimation in a hypersonic shock tunnel
International Journal of Heat and Mass Transfer 114 (2017) 748–757
Contents lists available at ScienceDirect
International Journal of Heat and Mass Transfer journal homepage: www.elsevier.com/locate/ijhmt
Comparative performance assessments of surface junction probes for stagnation heat flux estimation in a hypersonic shock tunnel Sumit Agarwal a, Niranjan Sahoo a,⇑, K.J. Irimpan b, Viren Menezes b, Siddesh Desai a a b
Department of Mechanical Engineering, Indian Institute of Technology Guwahati, Guwahati 781 039, India Department of Aerospace Engineering, Indian Institute of Technology Bombay, Powai, Mumbai 400 076, India
a r t i c l e
i n f o
Article history: Received 4 March 2017 Received in revised form 20 May 2017 Accepted 23 June 2017
Keywords: Heat probe Coaxial surface junction thermocouples Hypersonic flow Shock tunnel Surface heat flux
a b s t r a c t The aerodynamic environment in the test section of a hypersonic shock tunnel is very hostile as the high enthalpy flows prevail over the test model for a very short test time (1 ms). It calls for the requirement of the high-speed thermal sensors to survive in this harsh environment. Coaxial surface junction thermocouples (CSJTs) have the characteristics features of very high response times that can be used to capture highly transient temperature data during short duration hypersonic flows in the shock tunnel. The present study aims at the comparative performance assessment studies of three different surface junction probes (E, J and T-type) in a hypersonic shock tunnel at a flow Mach number of 8.2. All of them are mounted simultaneously in a rake along with a pitot-probe, in the test section of the tunnel where they experience a step heat load through a slug of test gas prevailing for 1 ms flow duration. Subsequently, the transient temperature histories are recorded at stagnation point of the junction probe and surface heat fluxes are predicted through one-dimensional heat conduction modeling. Side by side, stagnation heat flux is calculated independently through numerical simulations and analytical methods under same experimental flow conditions. The surface heat flux is recovered within a reasonable accuracy for E and T-type probes when experimental results are compared with numerical simulation and analytical solutions. In contrast, substantial deviations in surface heat flux value are observed for a J-type probe, which is mainly due to inconsistency in ‘‘thermal product (TP)” value of CSJT. However, they are quite sensitive in acquiring transient data during short time scale of measurements prevailing in the shock tunnel. The performance indicators such as ‘‘sensitivity and response time” for E-type CSJTs are found to be 58.96 lV/°C and 21 ls, respectively as compared to 43.82 lV/°C and 29 ls for J-type and 28.47 lV/°C and 24 ls for T-type, respectively. Ó 2017 Elsevier Ltd. All rights reserved.
1. Introduction The ability to capture transient temperatures in short duration time scale is very challenging which warrants toward the need for a highly sensitive/fast response sensor [1]. Due to several constraints, the test flow durations are limited to few hundreds of microseconds or less, in the ground based impulse facilities (e.g. shock tubes/shock tunnels/expansion tubes). Again, the stagnation enthalpies of test gases in these facilities are very high, for which the thermal survival of measuring devices becomes difficult [2–4]. At the same time, measuring surface temperature and the heat transfer rate are the dominant issues for the design of high-speed flight vehicles. In many other applications, the experimental data in ground-based impulse facilities need to be supported by the ⇑ Corresponding author. E-mail address: [email protected] (N. Sahoo). http://dx.doi.org/10.1016/j.ijheatmasstransfer.2017.06.109 0017-9310/Ó 2017 Elsevier Ltd. All rights reserved.
design of heat flux sensors. The classical approach of measuring surface temperatures on wind tunnel models in hypersonic flows is based on surface coating with encapsulated thermochromic liquid crystals [5,6]. These crystals reflect light depending on the detected surface temperatures and the reflections are visible as a color response based on different temperatures. The unsteady temperatures are also measured on the aerodynamic surfaces in many transient facilities by employing temperature sensitive paints (TSP) on the model surface [7–10]. The fundamental operating principle of TSPs is the oxygen quenching of luminescence from the paint. The light intensity emitted by the paint, is measured by a photo detector and subsequently correlated with the transient variations of temperatures. In recent years, a typical choice for the researchers to capture short duration transient surface temperatures is the use of platinum/nickel thin film gauges (TFGs) for which the response times fall in the range of 1 ms with a thickness of film the range of few micrometers. These sensors are made out of base
S. Agarwal et al. / International Journal of Heat and Mass Transfer 114 (2017) 748–757
749
Nomenclature c cP k Pr pe p1 q q_ s Rn S T
specific heat of the thermocouple material (J/kg K) specific heat capacity of the test gas at constant pressure (J/kg K) thermal conductivity (W/m K) Prandtl number static pressure at the edge of the boundary layer (Pa) free stream static pressure (Pa) stagnation point heat flux (W/cm2) surface heat flux (W/cm2) nose radius of the blunt body (m) sensitivity (mV/°C) free stream total temperature (K)
metals in the form of ink/paste supported on an insulating substrate [10–14]. The resistance of the metal changes with a temperature that give rise to transient variations in the temperature when the sensor is energized by the suitable power source. Being residues of paints, all these measuring techniques become more prone toward wear and tear on the sensing surface due to the impact of high-speed test gases. In many cases during experiments in shock tunnels, the thin film sensors lose their adequate resistances due to unexpected heat loads. So, frequent replacement is essential during experiments, which becomes cumbersome as it affects the accuracy and repeatability of measurements. An alternative approach to overcome the limitations of TFGs is to employ a ‘‘surface junction probe” instead of metal-insulating substrate combination. Such types of probes can be prepared from metallic elements known as ‘‘coaxial surface junction thermocouple (CSJT)” for which response times are in the similar range as that of TFGs [14–19]. In a CSJT, the sensing element is a surface junction of few micron thickness while it is a point junction in the case of a conventional thermocouple. The CSJT is fabricated by using two metallic elements through a surface junction formed between them obtained by grinding process. The small scale plastic deformation allows the fast response characteristics and makes it suitable for short-duration transient measurements [16,18]. Moreover, the variation in the temperature change in the highspeed flow leads to a voltage change across the metallic elements that can be captured by data acquisition system. The robust design, small size, ease in fabrication, fast response characteristics, ability to be flush-mounted on model surfaces and cost-effectiveness, are some of the distinct advantages of a CSJT as compared to its counterparts. Emphasizing on the need for inevitable short duration time scale measurements and harsh test flow conditions in impulse facilities, the surface junction probes are preferred candidates for transient temperature measurements as compared to other techniques. The design features and various methods of fabrication of different types of thermocouples have been discussed in open literature [15–18]. With a view point of for measuring surface temperature history on aerodynamic bodies in shock tunnel facilities, an E-type (chromel-constantan) and K-type CSJTs have been employed separately [15,17,18]. However, the effective comparisons of different types of CSJTs with simultaneous exposure to high temperature slug of test gases in shock tunnel flows can provide a better insight about the surface junction probe’s characterization (such as response time, prediction of surface heat flux). In fact, knowledge of these parameters is very vital for an accurate estimate of surface heat flux and they are rarely attempted in the reported literatures. In the present work, the authors explore different types of surface junction probes and address their comparative performances in shock tunnel experiments. All the probes experience step heat loads simultaneously and the corresponding
Tw T s ðtÞ t DT s =Dt
q qe le
b
model wall temperature (K) surface temperature (K) time (s) temperature gradient (K/s) density of the thermocouple material (kg/m3) static density at the edge of the boundary layer (kg/m3) dynamic viscosity at the edge of the boundary layer (N s/m2) thermal product (J m2 s0.5 K1)
transient temperature responses are acquired to address their comparative performances. As discussed above, attention in the present study is focused toward three different types of surface junction probes (E-type, T-type and J-type), in terms of their characterization (sensitivity, thermal product and response times) and predicting surface heat fluxes. The experiments are performed in a hypersonic shock tunnel having effective test time of 1 ms at flow Mach number of 8.2. For this purpose, a pitot probe and a hemispherical model housing the CSJTs are mounted simultaneously in the test section of the tunnel, where all the probes experience step rise of heat loads within the test flow duration. Prior to the shock tunnel testing, the sensitivity of each thermocouple (i.e. the variation of voltage with respect to temperature) is determined through the calibration experiments. The transient temperatures from shock tunnel experiments are recorded from each probe and subsequently surface heat fluxes are obtained through one-dimensional heat conduction analysis. Simultaneously, the experimental surface heat fluxes from all the probes are compared with analytical expressions as well as by numerical simulation for stagnation point heat flux. The fabrication details of CSJTs, calibration, shock tunnel experiments and surface heat flux computations are described in the subsequent section.
2. Fabrication and calibration methods The duration of test gas flow in impulsive experimental facilities prevails in the test section only for a few milliseconds or less. In such cases, the technique used to capture the transient temperature through surface junction probes must have a fast response time and suitable enough for rapidly changing flow conditions. Thus, the CSJTs should withstand the adverse test flow environment of very high stagnation enthalpies and be able to trace the transient temperatures within the short test window [1]. Being surface temperature detector, the CSJTs qualify for capturing the highly transient temperature signals. They are usually formed from two elements; the inner wire of the negative element and the outer annulus wire with the positive element. Both the elements are electrically insulated along their lengths. The CSJT has a unique feature of having a surface junction which is formed in its construction process through plastic deformation caused by eroding both the elements. The micro scratches are artificially created by abrading one thermocouple material over the other material by using scalpel blade that makes it suitable to respond in short duration time scale [16]. The important factors such as high electromotive force, fast response time, stability over the specified temperature range and resistant to oxidation, are considered while selecting the materials.
750
S. Agarwal et al. / International Journal of Heat and Mass Transfer 114 (2017) 748–757
2.1. Fabrication of surface junction probes In the present investigation, an E-type (chromel-constantan), Ttype (copper-constantan) and J-type (iron-constantan) surface junction probes have been considered for shock tunnel experiments. For all the types, the inner element is the constantan wire of 0.8128 mm diameter while the outer element has a diameter of 3.25 mm for chromel/copper and 2.5 mm for iron. The wires for the outer elements (i.e. chromel, copper and iron) are of 10 mm long and inner element (i.e. constantan wire) is kept slightly longer than the outer element. A hole (1 mm diameter) is drilled in the outer element with a drill bit and the inner element is then inserted with a layer of epoxy resin into the machined outer element. An electrical insulation (epoxy-resin) is maintained so as to prevent any electrical contacts between the two thermocouple wires (Fig. 1a). The sensing junction is created by slightly abrading one thermocouple material over the other, so that a plastic deformation of one material over the other material is formed, thereby, bridging the gap between the two thermocouple materials. This plastic deformation forms a microscopic junction on the surface (thickness 20–40 mm) with a very low thermal mass of inertia [16]. After the formation of the surface junction, adequate resistance is ascertained with connecting wires. The coaxial thermal sensors fabricated in-house in the above process are shown in Fig. 1(b). More details on different fabrication techniques have been reported in the literature [15–18]. 2.2. Oil-bath calibration for surface junction probes An oil-bath based experimental unit is being used for calibration in which the gradual step variation in the temperature is achieved in a controlled manner [19]. It is mainly performed to check the linearity between the change in voltage signals with the corresponding changes in temperature across the sensing the material during heating and cooling process. For a temperature difference ðT T Þ, if there is a change in voltage ðV V Þ across the junction probe, then the sensitivity ðSÞ of the thermocouple can be given by the following expression;
S¼
V V T T
ð1Þ
The value of S is one of the ‘‘performance indicator” that emphasize the magnitude of the voltage change with respect to temperature in a particular application. Its value will change with different thermocouple material and is obtained experimentally. In the present case, all the surface junction probes are calibrated by following the standard procedure as reported in the literature [14,19]. The setup consists of a heater, oil, beaker, a scientific thermometer, an ice bath to compensate for reference temperature and a digital multimeter (Fig. 2a). One of the junctions of the thermo-
couple is dipped into a beaker and the other ends are kept inside an ice-bath. The oil is heated by a heating element so that the temperature of air in the beaker is gradually regulated in the range of 30 –75 °C. The voltage and temperature data are recorded in step of 5 °C, mainly for two cases; one with a gradual increase of temperature of air from 30 °C to 75 °C (heating process) followed by a decrease in temperature from 75 °C to 30 °C (cooling process). The entire procedure is repeated for three times in order to check the repeatability of the recorded data. Using the data acquisition system (DAS), the variations in voltage are noted against the temperature change of the junction for the heating and cooling process. The calibration curves plotted in Fig. 2(b–d), show the linearity in voltage change corresponding to the changes in temperature during heating and cooling processes, respectively. The slope of the ‘voltage-temperature’ curve gives the ‘sensitivity’ of the surface junction probes and the average values are found to be about 58.96 mV/°C, 28.47 mV/°C and 43.82 mV/°C for E-type, Ttype and J-Type CSJTs, respectively with an uncertainty of ± 3%. These values are mentioned in Table 1 and compare well as reported in literature [20]. 3. Shock tunnel experiments The intended experiments with surface junction probes are carried out in the hypersonic shock tunnel facility of IIT Bombay (Fig. 3). With a flow Mach number of 8.2 at high stagnation enthalpy, these temperature probes experience ‘‘step heat loads” from the test gas. The facility comprises of two major parts i.e. ‘‘shock tube section” separated by a paper diaphragm with a converging-diverging nozzle followed by rectangular test domain (300 mm 300 mm cross-section and 450 mm long) and a dump tank (1 m diameter) attached to a high efficiency multistage vacuum pump (up to 106 mbar). The shock tube portion of the tunnel has a driver section (high pressure) and driven section (lowpressure region) separated by a metallic diaphragm (typically, aluminum alloy of 1.2 mm thickness). During the operation of the tunnel, the driver section is pumped so that the diaphragm ruptures suddenly. Thus, it generates a shock wave that propagates into driven section of the tube, thereby compressing the test gas inside it. As a result, the test gas gets heated due to the sudden rise in temperature across the shock wave. After reaching the end of the driven section, the primary shock undergoes reflection and propagates back into a medium where the test gas is already at elevated temperature and pressure. This leads to further enhancement in temperature and pressures of the test gas due to the reflected shock. Since, the gas motion behind the reflected shock is almost zero, the slug of the test gas in the driven section experiences a momentary reservoir of high pressure and temperature. Subsequently, the paper diaphragm ruptures and the complete slug of test gas expands in the nozzle while establishing a free
Fig. 1. (a) Schematic representation of surface junction probes fabricated in the laboratory; (b) photograph of the fabricated sensors.
751
S. Agarwal et al. / International Journal of Heat and Mass Transfer 114 (2017) 748–757
3.5
Sensitivity=58.96 µV/oC
3
Voltage (mV)
2.5 2 1.5 1 Avg H
0.5
Avg C
0 30
40
50
60
Temperature
(a)
80
(b)
2.5
1.6
Sensitivity = 43.82 µV/oC
2
Sensitivity= 28.47 µV/o C 1.2
Voltage(mV)
Volatage (mV)
70
(oC)
1.5
1
0.8
AVG H
0.4
0.5
AVG H
AVG C
AVG C 0
0 30
40
50
60
70
30
80
40
50
60
70
80
Temperature(oC)
Temperature (oC)
(c)
(d)
Fig. 2. Experimental calibration and voltage-temperature variation of surface junction probes: (a) schematic of oil-bath experimental set-up; (b) calibration curve (E-type); (c) calibration curve (J-type); (d) calibration curve (T-type). Notations: Average value during heating process (AVG H); Average value during cooling process (AVG C).
Table 1 Characteristics and performance indicators for surface junction probes. Tunnel characteristics (based on Pitot signal, Fig. 5a): Steady time = 1 ms; test time = 1 ms Types of CSJT
Materials and size
Thermal product [31] (J m2 K1 s0.5)
Sensitivity (Fig. 2) (mV/°C)
Response time (Fig. 5b) (ms)
Experimental surface heat flux (Eq. (5)) (W/cm2)
E
Chromel (outer element; 3.25 mm diameter) Constantan (inner element; 0.813 mm diameter) Iron (outer element; 2.5 mm diameter) Constantan (inner element; 0.813 mm diameter) Copper (outer element; 3.25 mm diameter) Constantan (inner element; 0.813 mm diameter)
8452
59
21
53 ± 2%
9585
44
29
89 ± 2%
7046
28.5
24
57 ± 2%
J
T
stream flow (Mach 8.2) in the test section for a typical time duration of 1 ms. The test flow conditions generated through this process with ‘helium’ as driver and ‘air’ as test gas are given in Table 2. 3.1. Test model and instrumentation The measurement diagnostics for the shock tunnel involves the determination of shock speed, nozzle-supply pressure and effective time duration of the slug of the gas passing over the test model. For
this purpose, two pressure transducers (Model: 102A08, Make: PCB Piezotronics, USA) are mounted toward the end of the driven section (505 mm apart) to measure the speed of the primary shock and reservoir/nozzle-supply pressure. Another pressure transducer (Pitot probe) placed in the test section, measures the pressure history of the test gas flow. The test model for surface junction probe is a hemispherical model (made out of brass) of radius 12.2 mm and length 35 mm (Fig. 4a). Considering the experimental flow conditions and geometrical shape of the hemispherical model, a
752
S. Agarwal et al. / International Journal of Heat and Mass Transfer 114 (2017) 748–757
Fig. 3. Hypersonic shock tunnel experimental facility.
Table 2 Free stream flow conditions in the test section of the shock tunnel. Mach No. ðM1 Þ
Static pressure P 1 ðPaÞ
Static temperature T 1 ðKÞ
Unit Reynolds number, Re1 ðm1 Þ
Stagnation enthalpy ðMJ=kgÞ
8.2 ± 2.5%
155 ± 5%
90 ± 3%
(1.48 106) ± 1%
1.41 ± 3%
Fig. 4. Hemispherical model housing surface junction probes and pitot pressure transducer for shock tunnel experiments: (a) geometrical details; (b) photograph of the rake.
S. Agarwal et al. / International Journal of Heat and Mass Transfer 114 (2017) 748–757
bow shock is expected for a Mach number of 8.2. As a result, there is a sudden rise in temperature leading to significant increase in surface heat flux in the vicinity of surface junction probes [21,22]. Hence, the hemispherical model housing the CSJTs may also be called as ‘‘stagnation probe”. For the shock tunnel experiments, a rake assembly consisting of hemispherical models for CSJTs and a Pitot probe are mounted in the test section of the tunnel with their leading edges in the same vertical plane at the exit of the nozzle (Fig. 4b). Further, it ensures the fact that all the CSJTs and the Pitot probe encounters same test flow conditions at the same time during the experiments. Moreover, the dimensions of the rake are chosen such that all the measurements are performed in the core flow regions of the test gas. An isothermal ambience has been maintained during the test time by wrapping the connecting leads of the sensor and the cold junction of the thermal sensor in the hemispheres. An op-amp instrumentation amplifier (INA 128) is used to amplify the output obtained from all the thermal sensors viz., E, T and J-types, respectively. The amplifier has a gain factor of 500 with an operating bandwidth frequency in the range of 1– 40 kHz, which is found to be sufficient for capturing the voltage signals in the shock tunnel flow environment (typically 1 kHz frequency). After suitable amplifications, all the signals from the sensors are acquired by using a PC-based data acquisition system armed with NI-PCI-6115 S-series cards, having a sampling rate of about 1 MS/s. The acquired signals are post-processed further through a 4th order IIR low-pass filter having a cut-off frequency of about 10 kHz. 3.2. Signal processing and data interpretation The quality of the flow at free stream Mach number of 8.2 in the test section can be ascertained by measuring transient responses simultaneously from the Pitot probe and surface junction thermal probe. This Pitot pressure transducer has a response time of 1 ms, as specified by the manufacturer. At the free stream test flow conditions (Table 2), it is expected to have bow shock wave formation over the hemispherical model [21]. Along the axis of each probe, a normal shock behavior can be assumed so that the Pitot probe experiences stagnation pressure drop across the normal shock. The effective test duration of the flow in the tunnel is the time during which the Pitot signal is steady (1 ms) as shown in Fig. 5(a). In the same plot, the typical time histories of surface temperature change ½T s ðtÞ during test flow duration (1 ms) recorded from each surface junction probe are presented so that the experimental data can be interpreted during test flow duration. All these signals follow a parabolic trend depicting the characteristics feature of thermal sensors in shock tunnel flow environment. By comparing the time domain for thermal probes and Pitot probe, it is found that the ‘‘response time” of all the surface junction probe is in the range of few microseconds (Table 1). Since all the sensors are mounted in single plane (Fig. 4) and keeping the Pitot probe as a benchmark reference sensor, the response time of each CSJTs are calculated (E-type: 21 ms; J-type: 29 ms; T-type 24 ms) using the time lag as shown in Fig. 5(b). These numbers seem to be adequate for ultrashort duration experiments [15,23–25]. Further, the parabolic trend of voltage history from surface junction probes is observed with almost similar time duration from the Pitot history. This fact justifies the usability and applicability of CSJTs for recovering transient temperature change during microsecond time scale duration in hypersonic flow environment. With known values of ‘‘sensitivity” of each probe (Table 1), the temperature rise experienced by the thermocouples during the test time are plotted in Fig. 5(c). From this plot, the temperature gradient ðDT s =DtÞ is obtained as, 1779 K/s (E-type) and 1951 K/s (T-type) and 1862 K/s (J-type). It justifies the fact that microscopic thickness of surface junction (30 mm) probes have adequate thermal inertia on shock tunnel
753
flow time scales. Moreover, the calorimetric principle of heat flux gauges are utilized for surface heat flux prediction, as the thermal penetration of heat into the probe is negligible during the shock tunnel experimental time scale [1,26]. However, the actual temperature rise is only in the range of 3–5.5 K during the test time, despite the test gas temperature at the stagnation point being about 1780 K. 4. Computations of surface heat flux 4.1. Prediction based on correlation The stagnation enthalpy of test gases in the shock tunnel is very high while the flow encounters ‘‘step rise” in surface heat flux during experimental test window of 1 ms at the surface junction probe. The heat flux at the stagnation point of the model can be estimated by theoretical approach by ‘‘Fay and Riddell” correlation [22,27]. Moreover, there exists modified version of this correlation at high and enthalpy test conditions in the literature [23,24]. In this case, the authors use Eq. (2) to calculate the stagnation heat flux corresponding to test flow conditions in the shock tunnel, which are also adopted in the literatures [27].
1=4 1 2ðpe p1 Þ qo ¼ 0:763Pr0:6 ðqe le Þ0:5 pffiffiffiffiffi cp ðT o T w Þ qe Rn
ð2Þ
Here Pr is Prandtl number of the test gas (= 0.71 for air), pe; qe and le are the static pressure, static density and dynamic viscosity, respectively at the edge of the boundary layer. Further, p1 is the free stream static pressure, cp is the specific heat capacity of the test gas at constant pressure, T o and T w are the free stream total temperature, and model wall temperature, respectively and Rn is the nose radius of the blunt body. For the present study, the average value of stagnation heat flux is estimated as, 37 W/cm2 with repeatability margin of ±3 W/cm2 (Table 3). 4.2. Numerical computations for surface heat flux prediction In order to compliment the experimental studies, a numerical non-equilibrium compressible flow solver has been used to perform the numerical simulation with a second order accuracy [28]. Essentially, it is an upgraded version of ‘‘unstructured solver for hypersonic aerothermodynamics simulations (USHAS)” developed in-house. The governing equations (mass, momentum, energy and species conservation) are solved for two-dimensional axisymmetric laminar flow in finite volume formulation where eleven prominent chemical reactions of dissociation of air and associated five species (N2, O2, NO, N, O) are accounted. This solver employs AUSM scheme for inviscid flux computations and second order accuracy has been achieved by using Venkakrishnan limiter [29]. More details of the calculations for high-temperature reaction kinetics, transport properties and thermodynamic properties may be found in Refs. [28,29]. Simulations are performed to compliment the present experimental studies with the freestream flow conditions (Table 2) for the thermal probe. It may be noted that the initial pressure (before arrival of the hypersonic flow) in the test section of the shock tunnel is of the order of 0.1 Pa. This condition leads to very high Knusden number and thus makes it difficult for simulation by using conventional Navier Stokes solver. Therefore, numerical simulations are generally initiated with hypersonic freestream conditions in the computational domain. Such simulations can give only the steady state results and not the temporal variation of any property. The typical computational domain along with the boundary conditions is shown in Fig. 6. In addition, no-slip boundary, isothermal wall condition (300 K) is considered over the hemisphere. In order
754
S. Agarwal et al. / International Journal of Heat and Mass Transfer 114 (2017) 748–757
Fig. 5. (a) Voltage signal from Pitot transducer and thermal probe during shock tunnel experiments; (b) Enlarged view of voltage signals of CSJTs and pitot transducer; (c) Rise in surface temperature history from surface junction probes.
Table 3 Comparison assessment of average values of surface heat fluxes. Types of CSJT
Experiment (Eq.
Stagnation heat flux
(5)) ðW=cm2 Þ
(Eq. (2)) ðW=cm2 Þ
E T J
53 ± 2% 57 ± 2% 89 ± 2%
36 ± 7%
Numerical simulation ðW=cm2 Þ 50.6
to carry out the essential mesh independence study, different meshes of size 280 100, 320 140, 360 180 are employed for simulations as mentioned in Table 4. The convergence histories obtained on all three grids are shown in Fig. 7, where steadystate convergence is assumed when the residual becomes lesser than 106. The stagnation heat fluxes are computed for each of the mesh sizes and encouraging match in values is noticed with
medium level mesh of size 320 140. Subsequently, the temperature and Mach number contours on this grid are plotted in Figs. 8 and 9, respectively. No alteration is chemical composition due to chemical reactions, is observed in these simulations. Temperature contour (Fig. 8) clearly depict this fact where maximum temperature in the domain is seen to be lower than the probable temperature of dissociation reactions. For this mesh size (320 140), the average value of stagnation point heat flux is obtained as, 50.6 W/cm2. The results from the in-house solver are compared with the final outcome of the experiment. 4.3. Recovery of surface heat flux during experiment While predicting surface heat flux from transient temperatures in shock tunnel tests, the standard approach governing one-
755
S. Agarwal et al. / International Journal of Heat and Mass Transfer 114 (2017) 748–757
Fig. 6. Computational domain and associated boundary conditions for flow over thermal probe.
Fig. 8. Temperature contours obtained from numerical simulation.
Table 4 Comparison assessment of average values of surface heat fluxes. Grid size 280 100 320 140 360 180
DYmin (m) 5
6.84 10 6 105 5.35 105
DXmin (m) 5
2.7 10 1.2 105 9 106
Stagnation point heat flux (W/cm2) 39.52 50.57 51.48
Fig. 9. Mach number contours obtained from numerical simulation.
When the thermal properties of the sensing junction are treated as constant, the time-dependent solution of surface heating rates ðq_ s Þ is given by Eq. (3).
Fig. 7. Convergence history for the different mesh sizes.
dimensional heat conduction formulation has been used [1,14– 16,25,26]. The physical model assumes the fact that lateral heat conduction during the experimental time scale (1 ms) is negligible and the heat is conducted only in the direction normal to the surface. In the present case, all the surface junction probes are prepared with metals which are good conductors and the formed junction has a thickness of the order of few micrometer (20– 24 mm). Moreover, the thermal penetration distance during the experimental run-times is negligible as compared to the linear dimension of the thermal sensor (10 mm). Thus, it is quite reasonable to assume that the rise in mean temperature is same as that of surface temperature history fT s ðtÞg as acquired by junction probe.
" # Z b T s ðtÞ 1 t T s ðtÞ T s ðsÞ q_ s ¼ pffiffiffiffi pffiffi þ d s ; 2 0 ðt sÞ3=2 p t
b¼
qffiffiffiffiffiffiffiffi qck
ð3Þ
Here, q is the density, c is the specific heat capacity and k is the thermal conductivity of the substrate material and b is defined as the ‘‘effective thermal product (TP)” of the substrate material. In the case of CSJTs, the sensing surface is prepared through plastic deformation of two different metals and it is unique for each type of CSJT. For an experimental estimate, the average thermal properties values ðq; c and kÞ for the junction materials can be taken from literature reported values [30]. In the present investigation, the ‘‘thermal product ðbÞ ” (for E and J type) is determined experimentally by ‘‘water plunging technique” as reported by authors’ earlier Ref. [31]. However, for T-type probe, the authors repeated the same experiments to determine the value of b. For each surface junction probe, these values are mentioned in Table 1.
756
S. Agarwal et al. / International Journal of Heat and Mass Transfer 114 (2017) 748–757
In order to use Eq. (3) for estimating surface heat flux, the experimental transient temperature data from CSJTs (Fig. 5b), need to be discretized. For the present case, cubic-spline based discretization technique is applied to obtain the correlated experimental temperature data [26,32].
1 fT s ðsÞgspline ¼ a1;i þ a2;i ðs si Þ þ a3;i ðs si Þ2 2 1 3 þ a4;i ðs si Þ ðsi 6 s 6 siþ1 ; i ¼ 1; 2; 3; . . . ; MÞ ð4Þ 6 When the discretized temperature data from Eq. (4) incorporated in Eq. (3), the final expression for surface heat flux is obtained (Eq. (5)).
2 2bffiffiffi p
6 p q_ s ¼ 6 4
2bffiffiffi p
p
M1 Xn
1=2
V i ðPi
1=2
3=2
Ri Þ W3 i ðPi
i¼1
1=2
3=2
a
5=2
4;i V M P M W3M PM þ 10 PM
3=2
Ri Þ þ
a4;i 10
Pi ¼ sMþ1 si ; Ri ¼ sMþ1 siþ1 ; F i ¼ a1;i þ a2;i Pi þ
5=2
ðPi
a3;i 2
o 3 5=2 Ri Þ þ 7 7 5
P2i þ
a4;i 6
P3i ;
2
i ; W i ¼ dds2F i V i ¼ dsdFMþ1
Mþ1
ð5Þ A numerical algorithm is developed in-house for discretization of temperature data (Eq. (4)) and subsequent computation of surface heat flux (Eq. (5)). A time step of the order of 1 ms was considered for discretization of temperature signal and subsequent computation of surface heat flux. All the temperature signals obtained from the surface junction probes during shock tunnel experiments are analyzed to infer stagnation heat flux. The transient temperature distributions obtained from temperature discretization technique are compared with experimental temperature history. All the signals show excellent recovery of analytical distribution of surface temperature rise through cubic spline fitting for a closed form solution of temperature history. The comparison of experimental signals and discretized data is shown in Fig. 10. Subsequently, surface heat fluxes predicted by Eq. (5), are plotted in Fig. 11. The average values of surface heat fluxes computed with test flow duration of 1 ms, for three repeated tests are given in Table 1. The surface heating rates predicted by temperature discretization data show excellent agreement within an experimental uncertainty of ±2%.
Fig. 10. Comparison of experimental signals from surface junction probes with discretized temperature data.
Fig. 11. Experimental heat flux recovery from surface junction probes during from shock tunnel tests.
5. Performance assessment of surface junction probes Considering the flow phenomena from Pitot pressure history (Fig. 5a), it is observed that there is a steady test flow duration of 1 ms, during which all the surface junction probes show a parabolic rise in surface temperature. During this time scale, the estimated values of the average heat flux are found to be, 53 W/cm2, 57 W/cm2 and 89 W/cm2, for E, T and J-types probes, respectively. Referring to Fig. 8, the heat flux histories are seen to have a steady region (1 ms) for E and T-type probes, analogous to experimental test times observed in the pitot pressure history. When average values on this time scale are compared, the recovery in surface heat fluxes has an excellent match for E-type probe (53 W/cm2) and T-type probe (57 W/cm2) with respect to numerical simulation (50.6 W/cm2). Under same test flow conditions, the ‘‘Fay and Riddell” expression for stagnation heat flux under-predicts the values (37 W/cm2). When the similar comparison is made for the J-type probe, there is significant over-prediction of average surface heat flux (89 W/cm2) during the shock tunnel experimental time scale. Moreover, J-type probe sees only a rising trend during same time scale and no steady region is found for average surface heat flux estimation. It essentially means that ‘‘thermal product of surface junction” does not fall in the line of shock tunnel time scale of experiments. It may be noted here that the diameter of outer element of thermocouple material is 2.5 mm for J-type probe in contrast with 3.25 mm for other two probes. Being slight oversize in diameter and with bridge thickness of 20–24 mm, it may not be possible to achieve adequate response time with respect to shock tunnel flows. Further, the effective proximity of the junction to the insulation between thermocouple materials and their purity during its construction process can also be a source of uncertainty in thermal product estimation. In the end, these issues leads to discrepancy in heat flux signals for J-type probe in comparison with its other counterparts. Moreover, it has been emphasized that slight variation in size and effective junction depth can affect the ‘‘response time and thermal product” of surface junction probes [33]. In the ultra-short duration of time scales, it is also reported that the actual value of b can be 30% smaller as compared to its theoretical estimates mentioned in Table 1. In this investigation, the authors explore a slightly higher size due to lack of availability of same size materials locally.
S. Agarwal et al. / International Journal of Heat and Mass Transfer 114 (2017) 748–757
Thus, from the present study, it may be inferred that E and Ttype surface junction probes are found to be suitable for shock tunnel experiments while measuring transient temperatures and predicting surface heat fluxes. In this case, the experimental temperature histories during shock tunnel tests reproduce transient surface heat fluxes through analytical technique as well by numerical simulation with reasonable accuracy. Moreover, it bears higher sensitivity values and an admissible ‘‘thermal product” for the ultra-short duration of time scales prevailing in shock tunnel experiments. For J-type of probe, the performance indicators (such as rise and response times) seem to be quite promising while acquiring transient temperature data for shock tunnel application. However, size of the thermocouple elements and adequate method of creating plastic deformation at the junction surface plays a crucial role for shock tunnel experiments. Further, the determination of b values through appropriate calibration methods needs to revisited while recovering surface heat fluxes from temperature histories during ultra-short duration time scale of experiments with step/impulse heat loads. 6. Conclusions Three surface junction probes (E, J and T-type) have been constructed in-house for their usage in capturing highly transient temperatures in ultra-short duration shock tunnel flows. They are mounted simultaneously on a rake along with the Pitot pressure transducer in the test section of the shock tunnel. When these stagnation probes and Pitot transducer experience the same free stream flow in the shock tunnel, the time responses from CSJTs are compared with Pitot signal. The quick response time of the surface junctions (21 ms for E-type, 29 ms for J-type and 24 ms for Ttype) reveal their suitability in capturing transient temperature data for millisecond duration flows in a shock tunnel. While recovering surface heat flux from transient temperature history, the Etype and T-type probe are found to be most suitable as it bears suitable thermal product for ultra-short duration time scales prevailing in shock tunnel experiments. The inconsistency in surface heat flux prediction from the J-type probe is mainly due to the discrepancy in the theoretical estimate of the effective thermal product of the sensing junction. The ruggedness, ease of fabrication and cost effectiveness, make the CSJTs more lucrative for experiments in shock tunnels. Moreover, they can be flush mounted on miniature aerodynamic model shapes due to their small sizes while measuring transient temperatures. The intention behind the research is to bring about the comparative behavior of inhouse developed surface junction probes for the shock-tunnel experiment. The performance indicators of surface junction probes are quite encouraging and very promising with respect to their future applications in the shock tunnels if the purity of junction elements, bridge thickness and thermal product estimation of junction probes are addressed properly. Acknowledgement The financial support received from ‘‘Aeronautical Research and Development Board (AR & DB)”, and ‘‘Gas Turbine Materials and Processes (GTMAP)”, Government of India (New Delhi) is highly acknowledged. References [1] D.L. Schultz, T.V. Jones, Heat Transfer Measurements in Short Duration Hypersonic Facilities, AGARD-AG-165, 1973. [2] F.K. Lu, D.R. Wilson, Survey of Short Duration Hypersonic and Hypervelocity Facilities, AIAA Paper, 1994, pp. 94–2491.
757
[3] N.M. Reddy, K. Nagashetty, G. Jagadeesh, K.P.J. Reddy, Review of hypersonic research investigations in the IISc shock tunnel, Proc. Ind. Acad. Sci. 21 (6) (1996) 741–773. [4] R.J. Stalker, A. Paull, D.J. Mee, R.G. Morgan, P.A. Jacobs, Scramjets and shock tunnels – the Queensland experience, Prog. Aerosp. Sci. 41 (2005) 471–513. [5] H. Babinsky, J.A. Edwards, Automatic liquid crystal thermography for transient heat transfer measurements in hypersonic flow, Exp. Fluids 21 (1996) 227– 236. [6] G.T. Robert, R.A. East, Liquid crystal thermography for heat transfer measurement in hypersonic flows – a review, J. Spacecraft Rockets 33 (6) (1996) 761–768. [7] J.P. Hubner, B.F. Carroll, K.S. Schanze, Heat-transfer measurements in hypersonic flow using luminescent coating techniques, J. Thermophys. Heat Transfer 16 (4) (2002) 516–522. [8] S. Ohmi, H. Nagai, K. Asai, K. Nakakita, Effect of temperature-sensitive-paint thickness on global heat transfer measurement in hypersonic flow, J. Thermophys. Heat Transfer 22 (3) (2008) 373–381. [9] T. Liu, Z. Cai, J. Lai, J. Rubal, J.P. Sullivan, Analytical methods for determination of heat transfer fields from tempartaure sensitive paint measurements in hypersonic tunnels, 2009, AIAA Paper 2009-736. [10] D. Peng, L. Jiao, Z. Sun, Y. Gu, Y. Liu, Simultaneous PSP and TSP measurements of transient flow in a long-duration hypersonic tunnel, Exp. Fluids 57 (12) (2016) 188. [11] D.A. Stewart, Y.K. Chen, Hypersonic convective heat transfer over 140-deg blunt cones in different gases, J. Spacecr. Rockets 31 (5) (1994) 735–743. [12] G. Jagadeesh, N.M. Reddy, K. Nagashetty, K.P.J. Reddy, Fore body convective hypersonic heat transfer measurements over large angle blunt cones, J. Spacecrafts Rockets 37 (1) (2000) 137–139. [13] N. Sahoo, V. Kulkarni, S. Saravanan, G. Jagadeesh, K.P.J. Reddy, Film cooling effectiveness on a large angle blunt cone flying at hypersonic speed, Phys. Fluids 17 (3) (2005) 1–11. [14] R. Kumar, N. Sahoo, V. Kulkarni, A. Singh, Laser based calibration technique of thin film gauges for short duration transient measurement, ASME J. Therm. Sci. Eng. Appl. 3 (4) (2011) 44504–445049. [15] V. Menezes, S. Bhat, A coaxial thermocouple for shock tunnel applications, Rev. Scientific Instrum. 81 (10) (2010) 104905. [16] H. Mohammed, H. Salleh, M.Z. Yusoff, Design and fabrication of coaxial surface junction thermocouples for transient heat transfer measurements, Int. Commun. Heat Mass Transfer 35 (7) (2008) 853–859. [17] S.R. Sanderson, B. Sturtevant, Transient heat flux measurement using a surface junction thermocouple, Rev. Sci. Instrum. 73 (7) (2002) 2781–2787. [18] H.A. Mohammed, H. Salleh, M.Z. Yusoff, Fast response surface temperature sensor for hypersonic vehicles, Instrum. Exp. Tech. 53 (1) (2010) 153–159. [19] R. Kumar, N. Sahoo, Dynamic calibration of a coaxial thermocouples for short duration transient measurements, ASME J. Heat Trans. 135 (12) (2013) 12450. [20] F.R. Caldwell, Thermocouple materials, in: Temperature measurement and control in science and industry (Ed. C.M. Herzfeld), Appl. Method Instrum., 1962. 3, pp. 81–134. [21] J.D. Anderson Jr, Modern compressible flow with historical perspective, 3rd ed., McGraw-Hill, New York, 2004. [22] J.D. Anderson, Hypersonic and High Temperature Gas Dynamics, AIAA, 2000. [23] S.L.N. Desikan, K. Suresh, K. Srinivasan, P.G. Raveendran, Fast response co-axial thermocouple for short duration impulse facilities, Appl. Therm. Eng. 96 (2016) 48–56. [24] K. Srinivasan, S.L.N. Desikan, R. Saravanan, A. Kumar, P.K. Maurya, Fore-body and base heat flux measurements on a typical crew module in short duration impulse facilities, Appl. Therm. Eng. 103 (2016) 842–854. [25] K.J. Irimpan, N. Mannil, H. Arya, V. Menezes, Performance evaluation of coaxial thermocouple against platinum thin film gauge for heat flux measurement in shock tunnel, Measurement 61 (2015) 291–298. [26] J. Taler, Theory of transient experimental technique for surface heat transfer, Int. J. Heat Mass Transf. 39 (1996) 3733–3748. [27] J.A. Fay, F.R. Riddell, Theory of stagnation point heat transfer in dissociated air, J. Aeronaut. Sci. 25 (2) (1958) 73–85. [28] S. Desai, V. Kulkarni, H. Gadgil, Delusive influence of nondimensional numbers in canonical hypersonic nonequilibrium flows, J. Aerospace Eng. 29 (5) (2016) 04016030. [29] B. John, V. Kulkarni, G. Natarajan, Investigation of ramp-induced shock wave boundary layer interactions in laminar hypersonic flows, Int. J. Heat Mass Transf. 70 (2014) 81–90. [30] H. Mohammed, H. Salleh, M.Z. Yusoff, The transient response for different types of erodable surface thermocouples using finite element analysis, Therm. Sci. 11 (4) (2007) 49–64. [31] S. Agarwal, N. Sahoo, R.K. Singh, Experimental techniques for thermal product determination of coaxial surface junction thermocouples during short duration transient measurements, Int. J. Heat Mass Transf. 103 (2016) 327– 335. [32] N. Sahoo, R.K. Peetala, Transient temperature data analysis for a supersonic flight test, J. Heat Transfer 132 (8) (2010) 084503. [33] D.R. Buttsworth, Assessment of effective thermal product of surface junction thermocouples on millisecond and microsecond time scales, Exp. Thermal Fluid Sci. 25 (2001) 409–420.
Contents lists available at ScienceDirect
International Journal of Heat and Mass Transfer journal homepage: www.elsevier.com/locate/ijhmt
Comparative performance assessments of surface junction probes for stagnation heat flux estimation in a hypersonic shock tunnel Sumit Agarwal a, Niranjan Sahoo a,⇑, K.J. Irimpan b, Viren Menezes b, Siddesh Desai a a b
Department of Mechanical Engineering, Indian Institute of Technology Guwahati, Guwahati 781 039, India Department of Aerospace Engineering, Indian Institute of Technology Bombay, Powai, Mumbai 400 076, India
a r t i c l e
i n f o
Article history: Received 4 March 2017 Received in revised form 20 May 2017 Accepted 23 June 2017
Keywords: Heat probe Coaxial surface junction thermocouples Hypersonic flow Shock tunnel Surface heat flux
a b s t r a c t The aerodynamic environment in the test section of a hypersonic shock tunnel is very hostile as the high enthalpy flows prevail over the test model for a very short test time (1 ms). It calls for the requirement of the high-speed thermal sensors to survive in this harsh environment. Coaxial surface junction thermocouples (CSJTs) have the characteristics features of very high response times that can be used to capture highly transient temperature data during short duration hypersonic flows in the shock tunnel. The present study aims at the comparative performance assessment studies of three different surface junction probes (E, J and T-type) in a hypersonic shock tunnel at a flow Mach number of 8.2. All of them are mounted simultaneously in a rake along with a pitot-probe, in the test section of the tunnel where they experience a step heat load through a slug of test gas prevailing for 1 ms flow duration. Subsequently, the transient temperature histories are recorded at stagnation point of the junction probe and surface heat fluxes are predicted through one-dimensional heat conduction modeling. Side by side, stagnation heat flux is calculated independently through numerical simulations and analytical methods under same experimental flow conditions. The surface heat flux is recovered within a reasonable accuracy for E and T-type probes when experimental results are compared with numerical simulation and analytical solutions. In contrast, substantial deviations in surface heat flux value are observed for a J-type probe, which is mainly due to inconsistency in ‘‘thermal product (TP)” value of CSJT. However, they are quite sensitive in acquiring transient data during short time scale of measurements prevailing in the shock tunnel. The performance indicators such as ‘‘sensitivity and response time” for E-type CSJTs are found to be 58.96 lV/°C and 21 ls, respectively as compared to 43.82 lV/°C and 29 ls for J-type and 28.47 lV/°C and 24 ls for T-type, respectively. Ó 2017 Elsevier Ltd. All rights reserved.
1. Introduction The ability to capture transient temperatures in short duration time scale is very challenging which warrants toward the need for a highly sensitive/fast response sensor [1]. Due to several constraints, the test flow durations are limited to few hundreds of microseconds or less, in the ground based impulse facilities (e.g. shock tubes/shock tunnels/expansion tubes). Again, the stagnation enthalpies of test gases in these facilities are very high, for which the thermal survival of measuring devices becomes difficult [2–4]. At the same time, measuring surface temperature and the heat transfer rate are the dominant issues for the design of high-speed flight vehicles. In many other applications, the experimental data in ground-based impulse facilities need to be supported by the ⇑ Corresponding author. E-mail address: [email protected] (N. Sahoo). http://dx.doi.org/10.1016/j.ijheatmasstransfer.2017.06.109 0017-9310/Ó 2017 Elsevier Ltd. All rights reserved.
design of heat flux sensors. The classical approach of measuring surface temperatures on wind tunnel models in hypersonic flows is based on surface coating with encapsulated thermochromic liquid crystals [5,6]. These crystals reflect light depending on the detected surface temperatures and the reflections are visible as a color response based on different temperatures. The unsteady temperatures are also measured on the aerodynamic surfaces in many transient facilities by employing temperature sensitive paints (TSP) on the model surface [7–10]. The fundamental operating principle of TSPs is the oxygen quenching of luminescence from the paint. The light intensity emitted by the paint, is measured by a photo detector and subsequently correlated with the transient variations of temperatures. In recent years, a typical choice for the researchers to capture short duration transient surface temperatures is the use of platinum/nickel thin film gauges (TFGs) for which the response times fall in the range of 1 ms with a thickness of film the range of few micrometers. These sensors are made out of base
S. Agarwal et al. / International Journal of Heat and Mass Transfer 114 (2017) 748–757
749
Nomenclature c cP k Pr pe p1 q q_ s Rn S T
specific heat of the thermocouple material (J/kg K) specific heat capacity of the test gas at constant pressure (J/kg K) thermal conductivity (W/m K) Prandtl number static pressure at the edge of the boundary layer (Pa) free stream static pressure (Pa) stagnation point heat flux (W/cm2) surface heat flux (W/cm2) nose radius of the blunt body (m) sensitivity (mV/°C) free stream total temperature (K)
metals in the form of ink/paste supported on an insulating substrate [10–14]. The resistance of the metal changes with a temperature that give rise to transient variations in the temperature when the sensor is energized by the suitable power source. Being residues of paints, all these measuring techniques become more prone toward wear and tear on the sensing surface due to the impact of high-speed test gases. In many cases during experiments in shock tunnels, the thin film sensors lose their adequate resistances due to unexpected heat loads. So, frequent replacement is essential during experiments, which becomes cumbersome as it affects the accuracy and repeatability of measurements. An alternative approach to overcome the limitations of TFGs is to employ a ‘‘surface junction probe” instead of metal-insulating substrate combination. Such types of probes can be prepared from metallic elements known as ‘‘coaxial surface junction thermocouple (CSJT)” for which response times are in the similar range as that of TFGs [14–19]. In a CSJT, the sensing element is a surface junction of few micron thickness while it is a point junction in the case of a conventional thermocouple. The CSJT is fabricated by using two metallic elements through a surface junction formed between them obtained by grinding process. The small scale plastic deformation allows the fast response characteristics and makes it suitable for short-duration transient measurements [16,18]. Moreover, the variation in the temperature change in the highspeed flow leads to a voltage change across the metallic elements that can be captured by data acquisition system. The robust design, small size, ease in fabrication, fast response characteristics, ability to be flush-mounted on model surfaces and cost-effectiveness, are some of the distinct advantages of a CSJT as compared to its counterparts. Emphasizing on the need for inevitable short duration time scale measurements and harsh test flow conditions in impulse facilities, the surface junction probes are preferred candidates for transient temperature measurements as compared to other techniques. The design features and various methods of fabrication of different types of thermocouples have been discussed in open literature [15–18]. With a view point of for measuring surface temperature history on aerodynamic bodies in shock tunnel facilities, an E-type (chromel-constantan) and K-type CSJTs have been employed separately [15,17,18]. However, the effective comparisons of different types of CSJTs with simultaneous exposure to high temperature slug of test gases in shock tunnel flows can provide a better insight about the surface junction probe’s characterization (such as response time, prediction of surface heat flux). In fact, knowledge of these parameters is very vital for an accurate estimate of surface heat flux and they are rarely attempted in the reported literatures. In the present work, the authors explore different types of surface junction probes and address their comparative performances in shock tunnel experiments. All the probes experience step heat loads simultaneously and the corresponding
Tw T s ðtÞ t DT s =Dt
q qe le
b
model wall temperature (K) surface temperature (K) time (s) temperature gradient (K/s) density of the thermocouple material (kg/m3) static density at the edge of the boundary layer (kg/m3) dynamic viscosity at the edge of the boundary layer (N s/m2) thermal product (J m2 s0.5 K1)
transient temperature responses are acquired to address their comparative performances. As discussed above, attention in the present study is focused toward three different types of surface junction probes (E-type, T-type and J-type), in terms of their characterization (sensitivity, thermal product and response times) and predicting surface heat fluxes. The experiments are performed in a hypersonic shock tunnel having effective test time of 1 ms at flow Mach number of 8.2. For this purpose, a pitot probe and a hemispherical model housing the CSJTs are mounted simultaneously in the test section of the tunnel, where all the probes experience step rise of heat loads within the test flow duration. Prior to the shock tunnel testing, the sensitivity of each thermocouple (i.e. the variation of voltage with respect to temperature) is determined through the calibration experiments. The transient temperatures from shock tunnel experiments are recorded from each probe and subsequently surface heat fluxes are obtained through one-dimensional heat conduction analysis. Simultaneously, the experimental surface heat fluxes from all the probes are compared with analytical expressions as well as by numerical simulation for stagnation point heat flux. The fabrication details of CSJTs, calibration, shock tunnel experiments and surface heat flux computations are described in the subsequent section.
2. Fabrication and calibration methods The duration of test gas flow in impulsive experimental facilities prevails in the test section only for a few milliseconds or less. In such cases, the technique used to capture the transient temperature through surface junction probes must have a fast response time and suitable enough for rapidly changing flow conditions. Thus, the CSJTs should withstand the adverse test flow environment of very high stagnation enthalpies and be able to trace the transient temperatures within the short test window [1]. Being surface temperature detector, the CSJTs qualify for capturing the highly transient temperature signals. They are usually formed from two elements; the inner wire of the negative element and the outer annulus wire with the positive element. Both the elements are electrically insulated along their lengths. The CSJT has a unique feature of having a surface junction which is formed in its construction process through plastic deformation caused by eroding both the elements. The micro scratches are artificially created by abrading one thermocouple material over the other material by using scalpel blade that makes it suitable to respond in short duration time scale [16]. The important factors such as high electromotive force, fast response time, stability over the specified temperature range and resistant to oxidation, are considered while selecting the materials.
750
S. Agarwal et al. / International Journal of Heat and Mass Transfer 114 (2017) 748–757
2.1. Fabrication of surface junction probes In the present investigation, an E-type (chromel-constantan), Ttype (copper-constantan) and J-type (iron-constantan) surface junction probes have been considered for shock tunnel experiments. For all the types, the inner element is the constantan wire of 0.8128 mm diameter while the outer element has a diameter of 3.25 mm for chromel/copper and 2.5 mm for iron. The wires for the outer elements (i.e. chromel, copper and iron) are of 10 mm long and inner element (i.e. constantan wire) is kept slightly longer than the outer element. A hole (1 mm diameter) is drilled in the outer element with a drill bit and the inner element is then inserted with a layer of epoxy resin into the machined outer element. An electrical insulation (epoxy-resin) is maintained so as to prevent any electrical contacts between the two thermocouple wires (Fig. 1a). The sensing junction is created by slightly abrading one thermocouple material over the other, so that a plastic deformation of one material over the other material is formed, thereby, bridging the gap between the two thermocouple materials. This plastic deformation forms a microscopic junction on the surface (thickness 20–40 mm) with a very low thermal mass of inertia [16]. After the formation of the surface junction, adequate resistance is ascertained with connecting wires. The coaxial thermal sensors fabricated in-house in the above process are shown in Fig. 1(b). More details on different fabrication techniques have been reported in the literature [15–18]. 2.2. Oil-bath calibration for surface junction probes An oil-bath based experimental unit is being used for calibration in which the gradual step variation in the temperature is achieved in a controlled manner [19]. It is mainly performed to check the linearity between the change in voltage signals with the corresponding changes in temperature across the sensing the material during heating and cooling process. For a temperature difference ðT T Þ, if there is a change in voltage ðV V Þ across the junction probe, then the sensitivity ðSÞ of the thermocouple can be given by the following expression;
S¼
V V T T
ð1Þ
The value of S is one of the ‘‘performance indicator” that emphasize the magnitude of the voltage change with respect to temperature in a particular application. Its value will change with different thermocouple material and is obtained experimentally. In the present case, all the surface junction probes are calibrated by following the standard procedure as reported in the literature [14,19]. The setup consists of a heater, oil, beaker, a scientific thermometer, an ice bath to compensate for reference temperature and a digital multimeter (Fig. 2a). One of the junctions of the thermo-
couple is dipped into a beaker and the other ends are kept inside an ice-bath. The oil is heated by a heating element so that the temperature of air in the beaker is gradually regulated in the range of 30 –75 °C. The voltage and temperature data are recorded in step of 5 °C, mainly for two cases; one with a gradual increase of temperature of air from 30 °C to 75 °C (heating process) followed by a decrease in temperature from 75 °C to 30 °C (cooling process). The entire procedure is repeated for three times in order to check the repeatability of the recorded data. Using the data acquisition system (DAS), the variations in voltage are noted against the temperature change of the junction for the heating and cooling process. The calibration curves plotted in Fig. 2(b–d), show the linearity in voltage change corresponding to the changes in temperature during heating and cooling processes, respectively. The slope of the ‘voltage-temperature’ curve gives the ‘sensitivity’ of the surface junction probes and the average values are found to be about 58.96 mV/°C, 28.47 mV/°C and 43.82 mV/°C for E-type, Ttype and J-Type CSJTs, respectively with an uncertainty of ± 3%. These values are mentioned in Table 1 and compare well as reported in literature [20]. 3. Shock tunnel experiments The intended experiments with surface junction probes are carried out in the hypersonic shock tunnel facility of IIT Bombay (Fig. 3). With a flow Mach number of 8.2 at high stagnation enthalpy, these temperature probes experience ‘‘step heat loads” from the test gas. The facility comprises of two major parts i.e. ‘‘shock tube section” separated by a paper diaphragm with a converging-diverging nozzle followed by rectangular test domain (300 mm 300 mm cross-section and 450 mm long) and a dump tank (1 m diameter) attached to a high efficiency multistage vacuum pump (up to 106 mbar). The shock tube portion of the tunnel has a driver section (high pressure) and driven section (lowpressure region) separated by a metallic diaphragm (typically, aluminum alloy of 1.2 mm thickness). During the operation of the tunnel, the driver section is pumped so that the diaphragm ruptures suddenly. Thus, it generates a shock wave that propagates into driven section of the tube, thereby compressing the test gas inside it. As a result, the test gas gets heated due to the sudden rise in temperature across the shock wave. After reaching the end of the driven section, the primary shock undergoes reflection and propagates back into a medium where the test gas is already at elevated temperature and pressure. This leads to further enhancement in temperature and pressures of the test gas due to the reflected shock. Since, the gas motion behind the reflected shock is almost zero, the slug of the test gas in the driven section experiences a momentary reservoir of high pressure and temperature. Subsequently, the paper diaphragm ruptures and the complete slug of test gas expands in the nozzle while establishing a free
Fig. 1. (a) Schematic representation of surface junction probes fabricated in the laboratory; (b) photograph of the fabricated sensors.
751
S. Agarwal et al. / International Journal of Heat and Mass Transfer 114 (2017) 748–757
3.5
Sensitivity=58.96 µV/oC
3
Voltage (mV)
2.5 2 1.5 1 Avg H
0.5
Avg C
0 30
40
50
60
Temperature
(a)
80
(b)
2.5
1.6
Sensitivity = 43.82 µV/oC
2
Sensitivity= 28.47 µV/o C 1.2
Voltage(mV)
Volatage (mV)
70
(oC)
1.5
1
0.8
AVG H
0.4
0.5
AVG H
AVG C
AVG C 0
0 30
40
50
60
70
30
80
40
50
60
70
80
Temperature(oC)
Temperature (oC)
(c)
(d)
Fig. 2. Experimental calibration and voltage-temperature variation of surface junction probes: (a) schematic of oil-bath experimental set-up; (b) calibration curve (E-type); (c) calibration curve (J-type); (d) calibration curve (T-type). Notations: Average value during heating process (AVG H); Average value during cooling process (AVG C).
Table 1 Characteristics and performance indicators for surface junction probes. Tunnel characteristics (based on Pitot signal, Fig. 5a): Steady time = 1 ms; test time = 1 ms Types of CSJT
Materials and size
Thermal product [31] (J m2 K1 s0.5)
Sensitivity (Fig. 2) (mV/°C)
Response time (Fig. 5b) (ms)
Experimental surface heat flux (Eq. (5)) (W/cm2)
E
Chromel (outer element; 3.25 mm diameter) Constantan (inner element; 0.813 mm diameter) Iron (outer element; 2.5 mm diameter) Constantan (inner element; 0.813 mm diameter) Copper (outer element; 3.25 mm diameter) Constantan (inner element; 0.813 mm diameter)
8452
59
21
53 ± 2%
9585
44
29
89 ± 2%
7046
28.5
24
57 ± 2%
J
T
stream flow (Mach 8.2) in the test section for a typical time duration of 1 ms. The test flow conditions generated through this process with ‘helium’ as driver and ‘air’ as test gas are given in Table 2. 3.1. Test model and instrumentation The measurement diagnostics for the shock tunnel involves the determination of shock speed, nozzle-supply pressure and effective time duration of the slug of the gas passing over the test model. For
this purpose, two pressure transducers (Model: 102A08, Make: PCB Piezotronics, USA) are mounted toward the end of the driven section (505 mm apart) to measure the speed of the primary shock and reservoir/nozzle-supply pressure. Another pressure transducer (Pitot probe) placed in the test section, measures the pressure history of the test gas flow. The test model for surface junction probe is a hemispherical model (made out of brass) of radius 12.2 mm and length 35 mm (Fig. 4a). Considering the experimental flow conditions and geometrical shape of the hemispherical model, a
752
S. Agarwal et al. / International Journal of Heat and Mass Transfer 114 (2017) 748–757
Fig. 3. Hypersonic shock tunnel experimental facility.
Table 2 Free stream flow conditions in the test section of the shock tunnel. Mach No. ðM1 Þ
Static pressure P 1 ðPaÞ
Static temperature T 1 ðKÞ
Unit Reynolds number, Re1 ðm1 Þ
Stagnation enthalpy ðMJ=kgÞ
8.2 ± 2.5%
155 ± 5%
90 ± 3%
(1.48 106) ± 1%
1.41 ± 3%
Fig. 4. Hemispherical model housing surface junction probes and pitot pressure transducer for shock tunnel experiments: (a) geometrical details; (b) photograph of the rake.
S. Agarwal et al. / International Journal of Heat and Mass Transfer 114 (2017) 748–757
bow shock is expected for a Mach number of 8.2. As a result, there is a sudden rise in temperature leading to significant increase in surface heat flux in the vicinity of surface junction probes [21,22]. Hence, the hemispherical model housing the CSJTs may also be called as ‘‘stagnation probe”. For the shock tunnel experiments, a rake assembly consisting of hemispherical models for CSJTs and a Pitot probe are mounted in the test section of the tunnel with their leading edges in the same vertical plane at the exit of the nozzle (Fig. 4b). Further, it ensures the fact that all the CSJTs and the Pitot probe encounters same test flow conditions at the same time during the experiments. Moreover, the dimensions of the rake are chosen such that all the measurements are performed in the core flow regions of the test gas. An isothermal ambience has been maintained during the test time by wrapping the connecting leads of the sensor and the cold junction of the thermal sensor in the hemispheres. An op-amp instrumentation amplifier (INA 128) is used to amplify the output obtained from all the thermal sensors viz., E, T and J-types, respectively. The amplifier has a gain factor of 500 with an operating bandwidth frequency in the range of 1– 40 kHz, which is found to be sufficient for capturing the voltage signals in the shock tunnel flow environment (typically 1 kHz frequency). After suitable amplifications, all the signals from the sensors are acquired by using a PC-based data acquisition system armed with NI-PCI-6115 S-series cards, having a sampling rate of about 1 MS/s. The acquired signals are post-processed further through a 4th order IIR low-pass filter having a cut-off frequency of about 10 kHz. 3.2. Signal processing and data interpretation The quality of the flow at free stream Mach number of 8.2 in the test section can be ascertained by measuring transient responses simultaneously from the Pitot probe and surface junction thermal probe. This Pitot pressure transducer has a response time of 1 ms, as specified by the manufacturer. At the free stream test flow conditions (Table 2), it is expected to have bow shock wave formation over the hemispherical model [21]. Along the axis of each probe, a normal shock behavior can be assumed so that the Pitot probe experiences stagnation pressure drop across the normal shock. The effective test duration of the flow in the tunnel is the time during which the Pitot signal is steady (1 ms) as shown in Fig. 5(a). In the same plot, the typical time histories of surface temperature change ½T s ðtÞ during test flow duration (1 ms) recorded from each surface junction probe are presented so that the experimental data can be interpreted during test flow duration. All these signals follow a parabolic trend depicting the characteristics feature of thermal sensors in shock tunnel flow environment. By comparing the time domain for thermal probes and Pitot probe, it is found that the ‘‘response time” of all the surface junction probe is in the range of few microseconds (Table 1). Since all the sensors are mounted in single plane (Fig. 4) and keeping the Pitot probe as a benchmark reference sensor, the response time of each CSJTs are calculated (E-type: 21 ms; J-type: 29 ms; T-type 24 ms) using the time lag as shown in Fig. 5(b). These numbers seem to be adequate for ultrashort duration experiments [15,23–25]. Further, the parabolic trend of voltage history from surface junction probes is observed with almost similar time duration from the Pitot history. This fact justifies the usability and applicability of CSJTs for recovering transient temperature change during microsecond time scale duration in hypersonic flow environment. With known values of ‘‘sensitivity” of each probe (Table 1), the temperature rise experienced by the thermocouples during the test time are plotted in Fig. 5(c). From this plot, the temperature gradient ðDT s =DtÞ is obtained as, 1779 K/s (E-type) and 1951 K/s (T-type) and 1862 K/s (J-type). It justifies the fact that microscopic thickness of surface junction (30 mm) probes have adequate thermal inertia on shock tunnel
753
flow time scales. Moreover, the calorimetric principle of heat flux gauges are utilized for surface heat flux prediction, as the thermal penetration of heat into the probe is negligible during the shock tunnel experimental time scale [1,26]. However, the actual temperature rise is only in the range of 3–5.5 K during the test time, despite the test gas temperature at the stagnation point being about 1780 K. 4. Computations of surface heat flux 4.1. Prediction based on correlation The stagnation enthalpy of test gases in the shock tunnel is very high while the flow encounters ‘‘step rise” in surface heat flux during experimental test window of 1 ms at the surface junction probe. The heat flux at the stagnation point of the model can be estimated by theoretical approach by ‘‘Fay and Riddell” correlation [22,27]. Moreover, there exists modified version of this correlation at high and enthalpy test conditions in the literature [23,24]. In this case, the authors use Eq. (2) to calculate the stagnation heat flux corresponding to test flow conditions in the shock tunnel, which are also adopted in the literatures [27].
1=4 1 2ðpe p1 Þ qo ¼ 0:763Pr0:6 ðqe le Þ0:5 pffiffiffiffiffi cp ðT o T w Þ qe Rn
ð2Þ
Here Pr is Prandtl number of the test gas (= 0.71 for air), pe; qe and le are the static pressure, static density and dynamic viscosity, respectively at the edge of the boundary layer. Further, p1 is the free stream static pressure, cp is the specific heat capacity of the test gas at constant pressure, T o and T w are the free stream total temperature, and model wall temperature, respectively and Rn is the nose radius of the blunt body. For the present study, the average value of stagnation heat flux is estimated as, 37 W/cm2 with repeatability margin of ±3 W/cm2 (Table 3). 4.2. Numerical computations for surface heat flux prediction In order to compliment the experimental studies, a numerical non-equilibrium compressible flow solver has been used to perform the numerical simulation with a second order accuracy [28]. Essentially, it is an upgraded version of ‘‘unstructured solver for hypersonic aerothermodynamics simulations (USHAS)” developed in-house. The governing equations (mass, momentum, energy and species conservation) are solved for two-dimensional axisymmetric laminar flow in finite volume formulation where eleven prominent chemical reactions of dissociation of air and associated five species (N2, O2, NO, N, O) are accounted. This solver employs AUSM scheme for inviscid flux computations and second order accuracy has been achieved by using Venkakrishnan limiter [29]. More details of the calculations for high-temperature reaction kinetics, transport properties and thermodynamic properties may be found in Refs. [28,29]. Simulations are performed to compliment the present experimental studies with the freestream flow conditions (Table 2) for the thermal probe. It may be noted that the initial pressure (before arrival of the hypersonic flow) in the test section of the shock tunnel is of the order of 0.1 Pa. This condition leads to very high Knusden number and thus makes it difficult for simulation by using conventional Navier Stokes solver. Therefore, numerical simulations are generally initiated with hypersonic freestream conditions in the computational domain. Such simulations can give only the steady state results and not the temporal variation of any property. The typical computational domain along with the boundary conditions is shown in Fig. 6. In addition, no-slip boundary, isothermal wall condition (300 K) is considered over the hemisphere. In order
754
S. Agarwal et al. / International Journal of Heat and Mass Transfer 114 (2017) 748–757
Fig. 5. (a) Voltage signal from Pitot transducer and thermal probe during shock tunnel experiments; (b) Enlarged view of voltage signals of CSJTs and pitot transducer; (c) Rise in surface temperature history from surface junction probes.
Table 3 Comparison assessment of average values of surface heat fluxes. Types of CSJT
Experiment (Eq.
Stagnation heat flux
(5)) ðW=cm2 Þ
(Eq. (2)) ðW=cm2 Þ
E T J
53 ± 2% 57 ± 2% 89 ± 2%
36 ± 7%
Numerical simulation ðW=cm2 Þ 50.6
to carry out the essential mesh independence study, different meshes of size 280 100, 320 140, 360 180 are employed for simulations as mentioned in Table 4. The convergence histories obtained on all three grids are shown in Fig. 7, where steadystate convergence is assumed when the residual becomes lesser than 106. The stagnation heat fluxes are computed for each of the mesh sizes and encouraging match in values is noticed with
medium level mesh of size 320 140. Subsequently, the temperature and Mach number contours on this grid are plotted in Figs. 8 and 9, respectively. No alteration is chemical composition due to chemical reactions, is observed in these simulations. Temperature contour (Fig. 8) clearly depict this fact where maximum temperature in the domain is seen to be lower than the probable temperature of dissociation reactions. For this mesh size (320 140), the average value of stagnation point heat flux is obtained as, 50.6 W/cm2. The results from the in-house solver are compared with the final outcome of the experiment. 4.3. Recovery of surface heat flux during experiment While predicting surface heat flux from transient temperatures in shock tunnel tests, the standard approach governing one-
755
S. Agarwal et al. / International Journal of Heat and Mass Transfer 114 (2017) 748–757
Fig. 6. Computational domain and associated boundary conditions for flow over thermal probe.
Fig. 8. Temperature contours obtained from numerical simulation.
Table 4 Comparison assessment of average values of surface heat fluxes. Grid size 280 100 320 140 360 180
DYmin (m) 5
6.84 10 6 105 5.35 105
DXmin (m) 5
2.7 10 1.2 105 9 106
Stagnation point heat flux (W/cm2) 39.52 50.57 51.48
Fig. 9. Mach number contours obtained from numerical simulation.
When the thermal properties of the sensing junction are treated as constant, the time-dependent solution of surface heating rates ðq_ s Þ is given by Eq. (3).
Fig. 7. Convergence history for the different mesh sizes.
dimensional heat conduction formulation has been used [1,14– 16,25,26]. The physical model assumes the fact that lateral heat conduction during the experimental time scale (1 ms) is negligible and the heat is conducted only in the direction normal to the surface. In the present case, all the surface junction probes are prepared with metals which are good conductors and the formed junction has a thickness of the order of few micrometer (20– 24 mm). Moreover, the thermal penetration distance during the experimental run-times is negligible as compared to the linear dimension of the thermal sensor (10 mm). Thus, it is quite reasonable to assume that the rise in mean temperature is same as that of surface temperature history fT s ðtÞg as acquired by junction probe.
" # Z b T s ðtÞ 1 t T s ðtÞ T s ðsÞ q_ s ¼ pffiffiffiffi pffiffi þ d s ; 2 0 ðt sÞ3=2 p t
b¼
qffiffiffiffiffiffiffiffi qck
ð3Þ
Here, q is the density, c is the specific heat capacity and k is the thermal conductivity of the substrate material and b is defined as the ‘‘effective thermal product (TP)” of the substrate material. In the case of CSJTs, the sensing surface is prepared through plastic deformation of two different metals and it is unique for each type of CSJT. For an experimental estimate, the average thermal properties values ðq; c and kÞ for the junction materials can be taken from literature reported values [30]. In the present investigation, the ‘‘thermal product ðbÞ ” (for E and J type) is determined experimentally by ‘‘water plunging technique” as reported by authors’ earlier Ref. [31]. However, for T-type probe, the authors repeated the same experiments to determine the value of b. For each surface junction probe, these values are mentioned in Table 1.
756
S. Agarwal et al. / International Journal of Heat and Mass Transfer 114 (2017) 748–757
In order to use Eq. (3) for estimating surface heat flux, the experimental transient temperature data from CSJTs (Fig. 5b), need to be discretized. For the present case, cubic-spline based discretization technique is applied to obtain the correlated experimental temperature data [26,32].
1 fT s ðsÞgspline ¼ a1;i þ a2;i ðs si Þ þ a3;i ðs si Þ2 2 1 3 þ a4;i ðs si Þ ðsi 6 s 6 siþ1 ; i ¼ 1; 2; 3; . . . ; MÞ ð4Þ 6 When the discretized temperature data from Eq. (4) incorporated in Eq. (3), the final expression for surface heat flux is obtained (Eq. (5)).
2 2bffiffiffi p
6 p q_ s ¼ 6 4
2bffiffiffi p
p
M1 Xn
1=2
V i ðPi
1=2
3=2
Ri Þ W3 i ðPi
i¼1
1=2
3=2
a
5=2
4;i V M P M W3M PM þ 10 PM
3=2
Ri Þ þ
a4;i 10
Pi ¼ sMþ1 si ; Ri ¼ sMþ1 siþ1 ; F i ¼ a1;i þ a2;i Pi þ
5=2
ðPi
a3;i 2
o 3 5=2 Ri Þ þ 7 7 5
P2i þ
a4;i 6
P3i ;
2
i ; W i ¼ dds2F i V i ¼ dsdFMþ1
Mþ1
ð5Þ A numerical algorithm is developed in-house for discretization of temperature data (Eq. (4)) and subsequent computation of surface heat flux (Eq. (5)). A time step of the order of 1 ms was considered for discretization of temperature signal and subsequent computation of surface heat flux. All the temperature signals obtained from the surface junction probes during shock tunnel experiments are analyzed to infer stagnation heat flux. The transient temperature distributions obtained from temperature discretization technique are compared with experimental temperature history. All the signals show excellent recovery of analytical distribution of surface temperature rise through cubic spline fitting for a closed form solution of temperature history. The comparison of experimental signals and discretized data is shown in Fig. 10. Subsequently, surface heat fluxes predicted by Eq. (5), are plotted in Fig. 11. The average values of surface heat fluxes computed with test flow duration of 1 ms, for three repeated tests are given in Table 1. The surface heating rates predicted by temperature discretization data show excellent agreement within an experimental uncertainty of ±2%.
Fig. 10. Comparison of experimental signals from surface junction probes with discretized temperature data.
Fig. 11. Experimental heat flux recovery from surface junction probes during from shock tunnel tests.
5. Performance assessment of surface junction probes Considering the flow phenomena from Pitot pressure history (Fig. 5a), it is observed that there is a steady test flow duration of 1 ms, during which all the surface junction probes show a parabolic rise in surface temperature. During this time scale, the estimated values of the average heat flux are found to be, 53 W/cm2, 57 W/cm2 and 89 W/cm2, for E, T and J-types probes, respectively. Referring to Fig. 8, the heat flux histories are seen to have a steady region (1 ms) for E and T-type probes, analogous to experimental test times observed in the pitot pressure history. When average values on this time scale are compared, the recovery in surface heat fluxes has an excellent match for E-type probe (53 W/cm2) and T-type probe (57 W/cm2) with respect to numerical simulation (50.6 W/cm2). Under same test flow conditions, the ‘‘Fay and Riddell” expression for stagnation heat flux under-predicts the values (37 W/cm2). When the similar comparison is made for the J-type probe, there is significant over-prediction of average surface heat flux (89 W/cm2) during the shock tunnel experimental time scale. Moreover, J-type probe sees only a rising trend during same time scale and no steady region is found for average surface heat flux estimation. It essentially means that ‘‘thermal product of surface junction” does not fall in the line of shock tunnel time scale of experiments. It may be noted here that the diameter of outer element of thermocouple material is 2.5 mm for J-type probe in contrast with 3.25 mm for other two probes. Being slight oversize in diameter and with bridge thickness of 20–24 mm, it may not be possible to achieve adequate response time with respect to shock tunnel flows. Further, the effective proximity of the junction to the insulation between thermocouple materials and their purity during its construction process can also be a source of uncertainty in thermal product estimation. In the end, these issues leads to discrepancy in heat flux signals for J-type probe in comparison with its other counterparts. Moreover, it has been emphasized that slight variation in size and effective junction depth can affect the ‘‘response time and thermal product” of surface junction probes [33]. In the ultra-short duration of time scales, it is also reported that the actual value of b can be 30% smaller as compared to its theoretical estimates mentioned in Table 1. In this investigation, the authors explore a slightly higher size due to lack of availability of same size materials locally.
S. Agarwal et al. / International Journal of Heat and Mass Transfer 114 (2017) 748–757
Thus, from the present study, it may be inferred that E and Ttype surface junction probes are found to be suitable for shock tunnel experiments while measuring transient temperatures and predicting surface heat fluxes. In this case, the experimental temperature histories during shock tunnel tests reproduce transient surface heat fluxes through analytical technique as well by numerical simulation with reasonable accuracy. Moreover, it bears higher sensitivity values and an admissible ‘‘thermal product” for the ultra-short duration of time scales prevailing in shock tunnel experiments. For J-type of probe, the performance indicators (such as rise and response times) seem to be quite promising while acquiring transient temperature data for shock tunnel application. However, size of the thermocouple elements and adequate method of creating plastic deformation at the junction surface plays a crucial role for shock tunnel experiments. Further, the determination of b values through appropriate calibration methods needs to revisited while recovering surface heat fluxes from temperature histories during ultra-short duration time scale of experiments with step/impulse heat loads. 6. Conclusions Three surface junction probes (E, J and T-type) have been constructed in-house for their usage in capturing highly transient temperatures in ultra-short duration shock tunnel flows. They are mounted simultaneously on a rake along with the Pitot pressure transducer in the test section of the shock tunnel. When these stagnation probes and Pitot transducer experience the same free stream flow in the shock tunnel, the time responses from CSJTs are compared with Pitot signal. The quick response time of the surface junctions (21 ms for E-type, 29 ms for J-type and 24 ms for Ttype) reveal their suitability in capturing transient temperature data for millisecond duration flows in a shock tunnel. While recovering surface heat flux from transient temperature history, the Etype and T-type probe are found to be most suitable as it bears suitable thermal product for ultra-short duration time scales prevailing in shock tunnel experiments. The inconsistency in surface heat flux prediction from the J-type probe is mainly due to the discrepancy in the theoretical estimate of the effective thermal product of the sensing junction. The ruggedness, ease of fabrication and cost effectiveness, make the CSJTs more lucrative for experiments in shock tunnels. Moreover, they can be flush mounted on miniature aerodynamic model shapes due to their small sizes while measuring transient temperatures. The intention behind the research is to bring about the comparative behavior of inhouse developed surface junction probes for the shock-tunnel experiment. The performance indicators of surface junction probes are quite encouraging and very promising with respect to their future applications in the shock tunnels if the purity of junction elements, bridge thickness and thermal product estimation of junction probes are addressed properly. Acknowledgement The financial support received from ‘‘Aeronautical Research and Development Board (AR & DB)”, and ‘‘Gas Turbine Materials and Processes (GTMAP)”, Government of India (New Delhi) is highly acknowledged. References [1] D.L. Schultz, T.V. Jones, Heat Transfer Measurements in Short Duration Hypersonic Facilities, AGARD-AG-165, 1973. [2] F.K. Lu, D.R. Wilson, Survey of Short Duration Hypersonic and Hypervelocity Facilities, AIAA Paper, 1994, pp. 94–2491.
757
[3] N.M. Reddy, K. Nagashetty, G. Jagadeesh, K.P.J. Reddy, Review of hypersonic research investigations in the IISc shock tunnel, Proc. Ind. Acad. Sci. 21 (6) (1996) 741–773. [4] R.J. Stalker, A. Paull, D.J. Mee, R.G. Morgan, P.A. Jacobs, Scramjets and shock tunnels – the Queensland experience, Prog. Aerosp. Sci. 41 (2005) 471–513. [5] H. Babinsky, J.A. Edwards, Automatic liquid crystal thermography for transient heat transfer measurements in hypersonic flow, Exp. Fluids 21 (1996) 227– 236. [6] G.T. Robert, R.A. East, Liquid crystal thermography for heat transfer measurement in hypersonic flows – a review, J. Spacecraft Rockets 33 (6) (1996) 761–768. [7] J.P. Hubner, B.F. Carroll, K.S. Schanze, Heat-transfer measurements in hypersonic flow using luminescent coating techniques, J. Thermophys. Heat Transfer 16 (4) (2002) 516–522. [8] S. Ohmi, H. Nagai, K. Asai, K. Nakakita, Effect of temperature-sensitive-paint thickness on global heat transfer measurement in hypersonic flow, J. Thermophys. Heat Transfer 22 (3) (2008) 373–381. [9] T. Liu, Z. Cai, J. Lai, J. Rubal, J.P. Sullivan, Analytical methods for determination of heat transfer fields from tempartaure sensitive paint measurements in hypersonic tunnels, 2009, AIAA Paper 2009-736. [10] D. Peng, L. Jiao, Z. Sun, Y. Gu, Y. Liu, Simultaneous PSP and TSP measurements of transient flow in a long-duration hypersonic tunnel, Exp. Fluids 57 (12) (2016) 188. [11] D.A. Stewart, Y.K. Chen, Hypersonic convective heat transfer over 140-deg blunt cones in different gases, J. Spacecr. Rockets 31 (5) (1994) 735–743. [12] G. Jagadeesh, N.M. Reddy, K. Nagashetty, K.P.J. Reddy, Fore body convective hypersonic heat transfer measurements over large angle blunt cones, J. Spacecrafts Rockets 37 (1) (2000) 137–139. [13] N. Sahoo, V. Kulkarni, S. Saravanan, G. Jagadeesh, K.P.J. Reddy, Film cooling effectiveness on a large angle blunt cone flying at hypersonic speed, Phys. Fluids 17 (3) (2005) 1–11. [14] R. Kumar, N. Sahoo, V. Kulkarni, A. Singh, Laser based calibration technique of thin film gauges for short duration transient measurement, ASME J. Therm. Sci. Eng. Appl. 3 (4) (2011) 44504–445049. [15] V. Menezes, S. Bhat, A coaxial thermocouple for shock tunnel applications, Rev. Scientific Instrum. 81 (10) (2010) 104905. [16] H. Mohammed, H. Salleh, M.Z. Yusoff, Design and fabrication of coaxial surface junction thermocouples for transient heat transfer measurements, Int. Commun. Heat Mass Transfer 35 (7) (2008) 853–859. [17] S.R. Sanderson, B. Sturtevant, Transient heat flux measurement using a surface junction thermocouple, Rev. Sci. Instrum. 73 (7) (2002) 2781–2787. [18] H.A. Mohammed, H. Salleh, M.Z. Yusoff, Fast response surface temperature sensor for hypersonic vehicles, Instrum. Exp. Tech. 53 (1) (2010) 153–159. [19] R. Kumar, N. Sahoo, Dynamic calibration of a coaxial thermocouples for short duration transient measurements, ASME J. Heat Trans. 135 (12) (2013) 12450. [20] F.R. Caldwell, Thermocouple materials, in: Temperature measurement and control in science and industry (Ed. C.M. Herzfeld), Appl. Method Instrum., 1962. 3, pp. 81–134. [21] J.D. Anderson Jr, Modern compressible flow with historical perspective, 3rd ed., McGraw-Hill, New York, 2004. [22] J.D. Anderson, Hypersonic and High Temperature Gas Dynamics, AIAA, 2000. [23] S.L.N. Desikan, K. Suresh, K. Srinivasan, P.G. Raveendran, Fast response co-axial thermocouple for short duration impulse facilities, Appl. Therm. Eng. 96 (2016) 48–56. [24] K. Srinivasan, S.L.N. Desikan, R. Saravanan, A. Kumar, P.K. Maurya, Fore-body and base heat flux measurements on a typical crew module in short duration impulse facilities, Appl. Therm. Eng. 103 (2016) 842–854. [25] K.J. Irimpan, N. Mannil, H. Arya, V. Menezes, Performance evaluation of coaxial thermocouple against platinum thin film gauge for heat flux measurement in shock tunnel, Measurement 61 (2015) 291–298. [26] J. Taler, Theory of transient experimental technique for surface heat transfer, Int. J. Heat Mass Transf. 39 (1996) 3733–3748. [27] J.A. Fay, F.R. Riddell, Theory of stagnation point heat transfer in dissociated air, J. Aeronaut. Sci. 25 (2) (1958) 73–85. [28] S. Desai, V. Kulkarni, H. Gadgil, Delusive influence of nondimensional numbers in canonical hypersonic nonequilibrium flows, J. Aerospace Eng. 29 (5) (2016) 04016030. [29] B. John, V. Kulkarni, G. Natarajan, Investigation of ramp-induced shock wave boundary layer interactions in laminar hypersonic flows, Int. J. Heat Mass Transf. 70 (2014) 81–90. [30] H. Mohammed, H. Salleh, M.Z. Yusoff, The transient response for different types of erodable surface thermocouples using finite element analysis, Therm. Sci. 11 (4) (2007) 49–64. [31] S. Agarwal, N. Sahoo, R.K. Singh, Experimental techniques for thermal product determination of coaxial surface junction thermocouples during short duration transient measurements, Int. J. Heat Mass Transf. 103 (2016) 327– 335. [32] N. Sahoo, R.K. Peetala, Transient temperature data analysis for a supersonic flight test, J. Heat Transfer 132 (8) (2010) 084503. [33] D.R. Buttsworth, Assessment of effective thermal product of surface junction thermocouples on millisecond and microsecond time scales, Exp. Thermal Fluid Sci. 25 (2001) 409–420.