GT2017-63053

EXPERIMENTAL STUDY OF IMPINGEMENT HEAT TRANSFER CHARACTERISTICS IN A CONVERGENT CHANNEL WITH PIN FIN Yang Xu Northwestern Polytechnical University Xi an 710072 China Hui ren zhu Northwestern Polytechnical University Xi an 710072 China Wei jiang Xu Northwestern Polytechnical University Xi an 710072 China Cun liang Liu Northwestern Polytechnical University Xi an 710072 China Hai ying Lu AVIC Shenyang Aircraft Engine Design Institute Shenyang 110000 China ABSTRACT The protection of the inlet components of an aircraft engine from the adverse effects of ice accretion has been a crucial design problem since the very early years of flight Therefore the configuration with efficient heat transfer is the focus on the design of a hot air anti icing system in an aero engine This paper studies experimentally on the heat transfer in a convergent channel with pin fin within the strut Experiments are carried out by using a transient liquid crystal technique The dimensionless lateral distance S d where S is the distance between the leading edge and pin fin d is the diameter of the impingement hole the dimensionless vertical distance S1 d where S1 is the distance between pin fins as well as the pitch ratio D d where D is the diameter of the pin fin of the pin fin and impingement hole are respectively investigated to study the heat transfer on the surface of convergent channel The Reynolds number based on the hydraulic diameter of the impingement hole ranges from 6300 to 12700 Within the experimental range the result shows that the Average Nusselt Number increases at first and then decreases as the dimensionless lateral distance S d increases The Average Nusselt Number has the highest value when S d 2 5 When the dimensionless vertical distance S1 d increases the Average Nusselt Number goes up at first and then reduces The Average Nusselt Number has the highest value when S1 d 3 The increasing pitch ratio D d leads to a lower Average Nusselt Number The Average Nusselt Number has the highest value when D d 1 3 The Average Nusselt Number increases as the Reynolds number increases INTRODUCTION The protection of the inlet components of an aircraft engine from the adverse effects of ice accretion has been a crucial design problem since the very early years of flight Ice formation on the inlet of an aircraft engine has a great influence on the aircraft engine s working condition Therefore the configuration with efficient heat transfer is the focus on the design of a hot air anti icing system in an aero engine Kissling et al 1 described the ground facility and evaluation technology of engine anti icing system tests He identified advantages and disadvantages of four means of evaluation and of instrumentation and equipment Lacey et al 2 briefly presented some methods of protecting the engine against icing and proposed a method of automatic anti iceing operation The adequacy of facilities for various types of icing tests was discussed by Olsen et al 3 Downs et al 4 investigated the boundary layer profile downstream of the anti icing air exhaust slot of a two dimensional aero engine inlet section by means of a pitot rake and indicated the injection of anti icing air produced a sub layer downstream of the anti icing air exhaust plane Riley et al 5 investigated various exhaust slot geometry designs in an experimental study to make sure the engine intake air upstream of the compressor could be used efficiently downstream of the exhaust slot to effect heating of the downstream surface acoustic liner to facilitate the prevention of ice build up in the inlet of an aircraft engine He found the most significant variables were assembly length exit plane width exhaust angle and slot depth A 2D Navier Stokes CFD code was used by Saeed et al 6 to simulate jet impingement on a a flat plate and b the inner surface a slat of a multi element airfoil He coupled the CFD code to an ice accretion and anti icing simulation code the CANICE and 1Copyright 2017 ASME Proceedings of ASME Turbo Expo 2017 Turbomachinery Technical Conference and Exposition GT2017 June 26 30 2017 Charlotte NC USA GT2017 63053 discussed the merits of using the CFD tool in conjunction with the CANICE code Plsnqusrt et al 7 presented experimental study on a full scale mock up of the leading edge section of a jet aircraft slat and dealt with the mapping of the convective heat transfer in a multijet anti icing system by application of the quantitative infrared thermography technique He also performed tridimensional numerical simulations with the commercial CFD code FLUENT Results showed that the heating performance of such a multijet system depended on the jet Reynolds number the distance of the supply duct and the skin and the spanwise and chordwise jet arrangement Recently the numerical method was used by Bisnco et al 8 to analyze an anti icing system based on hot air impinging jets on internal wing surface in order to check the efficiency of the system in order to find the optimal geometrical configuration to avoid the ice formation on the external wing surface An anti icing system was analyzed by Andreozzi et al 9 based on hot air impinging jets on internal wing surface in order to check the efficiency of the system Pellissier et al 10 used the finite element Navier Stokes applications package to compute the external and internal airflows of hot bleed air anti icing systems And he presented the optimal geometric configuration of the Piccolo tube A hot air anti icing system of a gas turbine engine inlet was analyzed by Kamel et al 11 numerically and used a three dimensional potential flow code to determine the flow field in and around the inlet And he developed a particle trajectory code by using a local linearization technique Dong et al 12 14 used the computational method to study the heat transfer distribution of an aero engine strut and a turboshaft engine inlet strut under icing conditions He also performed an experimental study on the performance of hot oil anti icing system of aero engine strut in icing wind tunnel In this study we selected the strut of the inlet components of an aircraft engine as the research object Due to the strut is similar to a convergent channel a further study on the heat transfer characteristics of the convergent channel with pin fin was performed to obtain the optimal pin fin arrays within the scope of this study This would enrich the theoretical research results of the heat transfer in the convergent channel with pin fin NOMENCLATURE c discharge coefficient kJ kg K d diameter of the impingement hole mm D diameter of the pin fin mm h heat transfer coefficient W m2 K LCD liquid crystal display m mass flow rate kg s Nu Nusselt number R Molar gas constant J mol 1K 1 Re Reynolds number S lateral distance mm S1 vertical distance mm S2 distance of leading edge area mm t the heating time s T temperature K T0 initial temperature of the subject Y vertical direction Greek Symbols thermal conductivity density dynamic viscosity coefficient Subscripts i initial airflow temperature g gas flow aw adiabatic wall ave average EXPERIMENTAL SETUP A model of convergent channel with pin fin as shown in figure 1 was designed and experimented in this paper where S is the distance between the leading edge and pin fin S1 is the distance between pin fins Since the area where the ices are iced is mainly the leading edge area and the area 1 5d from the leading edge covers the area where the leading edge needs to be heated S2 1 5d was selected as the leading edge area where d is the diameter of the impingement hole The leading edge area can be recognized by the black line in each contour of Re 9500 The experimental system is shown in figure 2 First air entered the mass flowmeter to measure the mass flow rate of air went into the heater which can increase the mainstream temperature by 15 in one second then supplied air to the test section through the inflow pipeline An entrance with two impingement holes was supplied to the test section FIGURE 1 SKETCH OF CONVERGENT CHANNEL 2Copyright 2017 ASME FIGURE 2 SCHEMATIC OF TEST SETUP MEASUREMENT THEORY The transient liquid crystal measurement technique was used to measure a complete Nu distribution on heated surface Initially the air sent from the centrifugal blower was heated then diverted directly into the atmosphere by the bypass port where provided with a thermocouple to monitor temperature of the airflow When the air flow temperature met the requirements the bypass port was shut down and the intake port of experimental system was open by solenoid valve controller to let air flow into the test section Simultaneously temperature data was collecting as well as the LCD color changing process on the experimental wall to calculate the heat transfer coefficient of inner wall The liquid crystal had been calibrated before the LCD was used to collect the temperature data in the experiment which could determine the location of the camera and lights Then the standardization of relation between hue and temperature can be carried out LCD temperature measurement principle and image processing technique can be found in Reference 15 By the results of previous studies 16 17 when the measuring plates material was plexiglass whose thickness is 15mm testing time was within 2min the interior heat transfer to the perpendicular direction of wall was the most important the impact of the side direction to the transient heat transfer coefficient measuring was very small so the impact of the lateral heat transfer was negligible The thickness of the test section was taken as 15mm The testing time would be controlled within 2min Conducted by semi infinite one dimensional transient heat conduction theory 18 the relationship between time and the subject s surface temperature at a certain point was presented 2 1 2 00 exp1 c th erfc c th TTTT gw 1 Where Tw was the wall temperature at some time thus the Thermochromic liquid crystal color temperature in the experiment K T0 was the initial temperature of the subject K Tg was the temperature of the heated fluid K h was the local heat transfer coefficient W m2 K t was the heating time s was the density of subject kg m3 was the thermal conductivity of subject W m K c was the specific heat capacity of subject J kg K PARAMETER DECLARATION Some of the main parameters were defined as follows Reynolds number Re at the outlet of each impingement hole 4 Re m d 2 Where m was the mass flow of each impingement hole kg s d was the hydraulic diameter of the impingement hole was the air dynamic viscosity kg m s Nusselt number hd Nu 3 Where h was the heat transfer coefficient of the convergent channel W m2 K d was Diameter of the impingement hole m was the thermal conductivity W m 2 K 1 TEST CONDITIONS Experiments were carried out to study the effects of the geometry the pin fin size spacing etc on the Nusselt number The experimental conditions were listed in Table 1 Exp S d S1 d D d Re 1 1 5 3 1 3 6300 2 1 5 3 1 3 9500 3 1 5 3 1 3 12700 4 2 5 3 1 3 6300 5 2 5 3 1 3 9500 6 2 5 3 1 3 12700 7 3 5 3 1 3 6300 8 3 5 3 1 3 9500 9 3 5 3 1 3 12700 10 2 5 2 1 3 6300 11 2 5 2 1 3 9500 12 2 5 2 1 3 12700 13 2 5 5 1 3 6300 14 2 5 5 1 3 9500 15 2 5 5 1 3 12700 16 3 5 3 2 3 6300 17 3 5 3 2 3 9500 18 3 5 3 2 3 12700 19 3 5 3 1 6300 20 3 5 3 1 9500 21 3 5 3 1 12700 TABLE 1 LIST OF TEST CONDITIONS 3Copyright 2017 ASME EXPERIMENTAL UNCERTAINTY The uncertainty of the mass flow In the experiments the mass flow rate m was obtained from volume flow rate Qv and the relationship between mass flow and volume flow rate was v p mQ RT 4 Based on the random error formulas of function dY 1 222 2 12 12 n n fff dxdxdx xxx 5 The uncertainty of the mass flow was 2 1 2 22 T dT p dp Q dQ m dm v v 6 The accuracy of the instruments used in experiments was as follows dQv Qv 2 5 dp p 0 5 dT T 0 03 we could figure out the uncertainty of the mass flow dm m 2 55 The uncertainty of the heat transfer coefficient Defining dimensionless temperature w aw 7 wwi TT 8 awawi TT 9 In the experiment the adiabatic wall temperature Taw Tg the airflow temperature For the semi infinite wall the steps figuring out heat transfer coefficient are f 10 2 1 x fe erfc 11 t h 12 c 13 And 1 2 2 1 df ff d 14 The uncertainty of the heat transfer coefficient was 222 11 22 ht PPPP ht 15 hT h aw PP h 16 12 2 2 1 h ff f 17 The relationship between h and could be figure out in Fig 3 according to 10 11 13 and 16 And h could be figure out based on 7 The uncertainty of the heat transfer coefficient Ph h could be figure out based on 15 at last FIGURE 3 THE RELATIONSHEIP BETWEEN h AND In the experiment the temperature of the main flow was set from 50 to 75 The temperature of the original wall changed from 18 to 24 according to local atmosphere The thermocouples used to calculate h were calibrated using a glycol bath so the bias error in the gas temperature measurements was less than 0 2 The range of the was from 0 24 to 0 75 The uncertainty of the heat transfer coefficient Ph h was less than 10 RESULTS AND DISCUSSION The experimental results of the heat transfer distribution in the convergent channel with different pin fin array configurations are shown in this section First the heat transfer distribution under different distances between the pin fin and the leading edge are shown and the leading edge area can be recognized by the black line in each contour of Re 9500 Next the heat transfer distribution under different pitch ratios of the pin fin and impingement hole is presented In the meantime the effects of the Reynolds number on the strut are also discussed In the end the thermal performance of the pin fin arrays is evaluated by the spanwise averaged Nusselt numbers in the leading edge area and the area averaged Nusselt numbers Effects of the lateral distance Fig 4 shows the distribution of the Nusselt number with smooth channel From Fig 4 it can be observed that the strong vortices which enhance the heat transfer are generated on the two sides of the impingement hole due to the impact of the jet At the same time only a small amount of the air flows out through the outlet since the velocity of outlet jet is high and concentrated while most of the air is still moving along the flow direction Due to the suction effect of the outlet the turbulence of air at the downstream of the outlet increases which leads to the enhancement of the heat transfer at the downstream of the outlet As the jet impingement becomes weaker the Nusselt number decreases gradually from the downstream of the outlet to the leading edge of the plate At the 4Copyright 2017 ASME same time with the increase of the Reynolds number the heat transfer area increases as well as the peak value a Re 6300 b Re 9500 b Re 12700 FIGURE 4 DISTRIBUTIONS OF THE NUSSELT NUMBER IN SMOOTH CHANNEL Fig 5 shows the distribution of the Nusselt number in the convergent channel at different dimensionless lateral distance S d when S1 d 3 D d 1 3 and Re 6300 9500 12700 Y is the vertical direction the distance between the leading edge and pin fin is S and the diameter of the impingement hole is defined as d It can be noted that the heat transfer distribution in the convergent channel with pin fin is similar to that in the smooth channel Compared with the smooth channel the turbulence intensity of air flow is strengthened by the pin fin arrangement which increases the distribution of the heat transfer along the spanwise of the channel and also makes the distribution more uniform As a result the Nusselt number in the channel increases When the dimensionless distance S d 1 5 the distance between the pin fin and the impingement hole is so far that the effect of the turbulence caused by the pin fin is low and weak for the leading edge area as can be seen in Fig 5 It is observed that since the distance between the pin fin and the impingement hole becomes closer the effect of the turbulence is enhanced when the dimensionless distance S d 2 5 which leads to the increase of the Nusselt number It is obvious that when the dimensionless distance S d 3 5 there is the closest distance between the pin fin and the impingement hole The effect of the turbulence caused by the pin fin is thus stronger especially at the area around the pin fin However if the pin fin is quite close to the impingement hole it can become more difficult for the jet flow flowing through the pin fin which weakens the flow and reduces the Nusselt number in the leading edge area Fig 6 shows the distribution of the spanwise averaged Nusselt number which is obtained from the leading edge area in the convergent channel when S1 d 3 and D d 1 3 It can be noted that the spanwise averaged Nusselt number increases since the pin fin is arranged on the convergent channel which leads to the increase of the intensity turbulence Because of the disturbance effects and obstruction of the pin fin the spanwise averaged Nusselt number increases first and then decreases with the increasing of lateral distance Besides the spanwise averaged Nusselt number reaches its maximum when S d 2 5 As the subjacent impingement hole is opposite to the pin fin the jet will be directly impinge on the front of the pin fin and the turbulence intensity becomes stronger which resulting in a stronger heat transfer However the upper impingement jet impinges on the side of the pin fin resulting in a lower disturbance which leads to a lower heat transfer In the further study if the spacing of the pin fin can be changed by ensuring that the pin fin is arranged at the center of the impingement jet the optimal heat transfer enhancement effect can be obtained As the Reynolds number increases the spanwise averaged Nusselt number increases a Re 6300 S d 1 5 b Re 9500 S d 1 5 c Re 12700 S d 1 5 d Re 6300 S d 2 5 e Re 9500 S d 2 5 5Copyright 2017 ASME f Re 12700 S d 2 5 g Re 6300 S d 3 5 h Re 9500 S d 3 5 i Re 12700 S d 3 5 FIGURE 5 DISTRIBUTIONS OF NUSSELT NUMBER WHEN Re 6300 9500 12700 AND S d 1 5 2 5 3 5 Fig