1、NASA Technical Paper 1369 LOAN COPY: RET kWL TECHNICAL KIRTLAND AFB, Transient Shutdown Analysis of Low-Temperature Thermal Diodes Richard J. Williams MARCH 1979 NASA Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-TECH LIBRARY KAFB, NM 0334380 NASA
2、Technical Paper 1369 Transient Shutdown Analysis of Low-Temperature Thermal. Diodes Richard J. Williams Ames Research Center Mofett Field, California National Aeronautics and Space Administration Scientific and Technical Information Office 1979 Provided by IHSNot for ResaleNo reproduction or network
3、ing permitted without license from IHS-,-,-TRANSIENT SHUTDOWN ANALYSIS OF LOW-TEMPERATURE THERMAL DIODES Richard J. Williams* Ames Research Center The various thermal diodes available for use in cryogenic systems are described. Two diode types, liquid-trap and liquid-blockage diodes, were considered
4、 to be the most attractive, and thermal models were constructed to predict their behavior in the reverse mode. The diodes, which used spiral artery wicks, were of similar size and throughput and were examined experimentally in a parallel test setup under nominally identical conditions. Their charact
5、eristics were ascertained in terlfrs of forward-mode and reverse-mode conductances, shutdown times and energies, and recovery to forward-mode operation with ethane as the working fluid in the temperature range I70 K to 220 K. Test data compare well with the data obtained with single heat pipe testin
6、g. Results show that the liquid-blockage diode is the quicker of the two diodes to shut down from the forward mode and that it transfers less energy to its evaporator during shutdown (8 min and 296 J as opposed to 10 min and 1150 J). However, the liquid-blockage diode has a larger reverse-mode condu
7、ctance which results in a greater overall evaporator temperature rise. The selection of the relative size and heat inputs to the condenserlreservoir configuration of the liquid-blockage diode was shown to be an important factor in the operation of the diode if the evaporator is to be protected from
8、a rapid increase in temperature after a reversal. Also included are data that show that the reinitiation of heat-piping action during recovery to forward-mode operation cannot be guaranteed if a limit in cool-down rate of the condenser is exceeded. This limit was found to be I Klmin for the liquid-t
9、rap diode and 2 Klmin for the liquid-blockage diode. General guidelines for the choice of a particular diode for an actual application are also given. Heat pipes are continuing to be developed to meet increasingly difficult requirements of spacecraft thermal control. Although extensive studies of bo
10、th active and passive variable conductance heat pipes for fine temperature control have been carried out, many applications exist wherein the ability of the heat pipe to conduct heat efficiently in one direction is of primary importance. Such heat pipes are described as thermal diodes (refs. 1, 2).
11、These diodes are attractive for use in the cryogenic temperature range. In the near future a large number of cryogenic payloads are due to be flown. In one proposed appli- cation, diode heat pipes are used to extract heat from a low-temperature sensor, such as an infrared detec- tor, and to thermall
12、y disconnect the sensor to prevent overheating should the radiator be exposed to a sudden high external heat flux. Using this concept, low-temperature cooling can be provided in low sub- solar Earth orbits where radiator cooling was never before considered possible. Several diode techniques have bee
13、n identified (ref. 1). These include the use of noncondensable gas, liquid-flow control, and freezing of the working fluid. *NRC Resident Research Associate More recently the concept of employing a foil reed in the vapor space to shut off the diode in the reverse mode has also been suggested. From t
14、hese diodes, two types of liquid flow control were considered to be the most attractive: the liquid-trap diode and the liquid-blockage diode. The liquid-trap concept employs a reservoir situ- ated at the normal evaporator end of the pipe which does not communicate with the wick. In normal- mode oper
15、ation the trap contains no liquid, and the diode performs as a normal heat pipe (cf. fig. 1). During reverse-mode operation the trap becomes the cold end of the pipe and condensation of the working fluid occurs inside the trap. The wick is thus depleted of working fluid and a rapid reduction of tran
16、sport capability results until all the fluid is condensed in the trap. Throughput is then limited to conduction heat transfer along the wall and wick. When condi- tions again reverse themselves, the trap becomes the hot end of the pipe and acts as an evaporator until all the liquid is expelled and n
17、ormal heat pipe action is resumed. With the liquid-blockage technique the heat pipe is charged with excess fluid. In normal-mode operation this excess fluid collects in a reservoir situated at the condenser end of the pipe (fig. 1). Under reverse- mode operation this excess liquid migrates to the co
18、ld Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-BLOCKING ORIFICE BLOCKED SOURCE / !:iiRESERVolR SINK I SOURCE HEAT HEAT HEAT NORMAL MODE REVERSE MODE RESERVOIR/TRAP EMPTY RESERVOIR/TRAP I n- I I“ tttt 4444 CCJi tttt NORMAL MODE REVERSE MODE Figure
19、 1 .- Liquid-trap and liquid-blockage concepts. end and there occupies a volume sufficient to block the vapor space of the normal evaporator together with a portion of the transport section. Due to its low thermal conductivity, the liquid very effectively limits the heat transfer in this section of
20、the device. This diode technique is attractive for cryogenic appli- cations where the normal-mode evaporator is rela- tively short compared with the condenser. This configuration minimizes the excess liquid required for blockage and thus minimizes the reservoir size, whereas with the liquid trap the
21、 reservoir must be sized to hold the majority of fluid in the pipe. This results in the trap being larger than the reservoir required for liquid blockage, and thus the liquid blockage diode has a weight advantage over the trap diode. However, an important factor in cryogenic diodes is the pipe press
22、ure under ambient conditions; for constant outside diameter and a given wick, the highest specific volume and therefore the lowest pres- sure are obtained with the liquid-trap technique. Thus, there still exists much conjecture and debate about the relative performance characteristics of the two dev
23、ices. The purpose of this study was, therefore, to examine both mechanisms under identical condi- tions to ascertain their characteristics in terms of shutdown times and energies, reverse-mode conduc- tance, and recovery to forward-mode operation. It is hoped that the results obtained from this inve
24、stiga- tion will provide valuable guidelines for designers in the choice of a diode to meet their requirements. NOMENCLATURE? A cf FT9 FB h k L P 6 Q R r T t V area forward-mode conductance Q/MCp ratios (eqs. (2) and (3) heat-transfer coefficient thermal conductivity length pressure heat input rate
25、heat input ratio condensation rate in evaporator to con- densation rate in trap bubble radius temperature time volume Subscripts: C condenser DO dryout e evaporator Ext external (area) Int internal (area) J joint R reservoir SD shutdown t trap 2 Provided by IHSNot for ResaleNo reproduction or networ
26、king permitted without license from IHS-,-,-DIODE CONSTRUCTION AND TEST SETUP Both arterial wicks and axial grooves are suitable for the capillary system of a liquid-trap diode. Axial grooves, although more reliable, are inferior in trans- port capacity to similarly sized arterial wicks; until recen
27、tly they have only been produced from alumi- num (ref. 3) which results in a higher reverse-mode conductance than that of a stainless steel pipe. The liquid-blockage diode, on the other hand, is almost entirely confined to an arterial wicking system. An axial-groove diode incorporating a plug in the
28、 evapo- rator vapor space has been suggested for spacecraft applications. However, the difficulties that would be encountered in a 1 -g test situation, such as draining of the upper grooves due to liquid communication between the grooves in the blocked portion, have precluded any development of such
29、 a diode. The artery diode must be designed to insure that the vapor space in the blocked portion is small enough so that the capillary force will support the pressure head of the liquid slug in 1-g ground tests. This is necessary for the vapor space to prime and remain filled in the reverse mode. T
30、hese small vapor spaces produce large vapor pressure drops in the normal mode of operation and thus restrict the capac- ity of the heat pipe. To circumvent this limitation, which is particularly severe at low temperatures where the fluids have low capillary-rise charac- teristics, a new geometry has
31、 been developed (ref. 4). An orifice plate is inserted in the heat pipe at the blocking meniscus, the opening of the plate being at the bottom of the pipe as shown in figure 2. The increased vapor flow area gained by the large evapo- rator vapor space more than compensates for the additional vapor p
32、ressure loss introduced by the orifice. BLOCKED+UNBLOCKED BLOCKING ARTERY I f ION A-A To allow performance comparisons, one diode of each type was constructed, each with a spiral artery wick and identical diameter and effective lengths to give comparable forward-mode conductances. The liquid-blockag
33、e design includes the aforementioned orifice plate - the liquid-trap heat pipe therefore has a slightly higher throughput, the difference represent- ing the blocking orifice pressure loss. The fabricated diodes consist of four sections: an evaporator, a transport section, a condenser, and either a l
34、iquid trap or liquid reservoir. The wicks were formed by wrapping 250 mesh stainless steel and 0.04-cm-diameter spaces on a mandril (see fig. 3). A transition section at the evaporator end of the liquid- trap diode is integral with a cylindrical reservoir having an inner core of aluminum channel. Si
35、milarly, a transition section at the condenser end of the liquid-blockage diode connects a reservoir to the heat pipe. No liquid communication, that is, no capilla:y connection, was provided between the reservoir and wick of the liquid-trap diode or between the reservoir and wick of the liquid-block
36、age diode. Liquid com- munication was achieved between the arteries and the condenser and evaporator walls with three equally spaced scroll-type webs manufactured from 250 mesh stainless steel screen. Circumferential grooves (63/cm) were used in both the evaporator and condenser sec- tions. Detailed
37、 design dimensions of the pipes are summarized in table 1. 0.051 cm diarn I Figure 2.- Blocking orifice - liquid blockage. Figure 3.- Wick geometry. 3 Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-TABLE 1.- HEAT PIPE DESIGN DATA Liquid trap Liquid
38、blockag sngths, cm Evaporator 10.16 10.16 Transport Blocked 0 10.16 Unblocked 28.26 18.10 Condenser 30.48 30.48 Reservoir/trap 15.56 6.03 Effective 48.57 48.57 )iameters, cm Pipe, 0.d. .635 .635 Pipe, i.d. .493 .493 Artery, 0.d. .300 .300 Solid tunnel, 0.d. .05 1 .05 1 Reservoir/trap, 0.d. 1.588 1.5
39、88 Ither pertinent design information includes: Pipes 304 - 1/8 HD stainless steel Screening 250-mesh 304 stainless steel Circumferential Reservoir/trap 6061 aluminum laminates grooves 63/cm (16Olin.) 0.239 cm thick with 0.127 cm wide X 0.127 cm deep axial machined grooves. Core machined to 1.448 cm
40、 0.d. for press fit into 304 - 1/8 HI stainless steel cylindrical shell The condensers of both diodes were enclosed in an aluminum block (fig. 4) which mated to a liquid nitrogen sink. Evaporator masses of 0.168 kg were attached to each evaporator to simulate a detector assembly and aluminum masses
41、were also attached to the trap (0.423 kg) and to the liquid-blockage reser- voir (0.182 kg). Strip heaters were attached to the evaporators and to the liquid trap to simulate forward-mode heat loads; rod heaters in the con- denser block controlled the forward-mode tempera- ture and, together with st
42、rip heaters on the liquid reservoir, provided a means for diode reversal. Both the liquid trap and the liquid reservoir were in con- tact with an LN2 cooling loop to facilitate a rapid cool-down and a wide range of temperature control. The locations of the thermocouples used to moni- tor the local t
43、emperature of the heat pipes are shown in figure 4. The rod heaters and a thermocouple embedded in the LN2 sink were connected to a unit that provides temperature control for the condenser sink. Once assembled, the diode package was wrapped in 30 layers of multilayer insulation (MU) before being ins
44、erted into a vacuum chamber. The MLI density was 30 layerslcm. All testing was performed in the vacuum chamber, which had ambient tempera- ture wells. TRANSIENT SHUTDOWN MODELS Liquid Trap If the liquid-trap diode is shut down directly from the forward mode of operation, there will be a temperature
45、gradient from the evaporator to the con- denser. Therefore, immediately after the initiation-of shutdown the diode will continue to operate in the forward mode until the condenser temperature is greater than both the evaporator and vapor tempera- ture. The temperature difference between the evapo- r
46、ator and the condenser and the sensible heat stored in the evaporator tend to retard the onset of shut- down. This retardation is linearly dependent on the forward-mode heat throughput. This time delay has not been accounted for in the thermal model to be detailed below. After the onset of shutdown,
47、 that is, with the condenser temperature greater than the evaporator temperature, the heat flow in the diode was reversed. Liquid was evaporated from the condenser and the vapor was condensed in the evaporator and trap. While the fluid that condensed in the trap remained there, the fluid that conden
48、sed in the evaporator was wicked back to the condenser to be reevaporated and maintained an adverse heat piping action. The tran- sient shutdown process of the diode therefore depended on the ratio of evaporator condensation rate to trap condensation rate. This ratio was desig- nated R. A thermal no
49、dal model of the liquid-trap diode system is shown in figure 5. By lumping together several parameters and neglecting the thermal inertia of the vapor, the con- denser can be coupled directly to the evaporator and trap by the overall forward-mode conductance of the pipe. In this manner the simplified model of figure 6 was obtained. The maximum heat transport capability of the diode was a function of liquid inventory and was 4 Provided by IHSNot
