1、Volume 16, Number 1, January 2010An International Journal of Heating, Ventilating,Air-Conditioning and Refrigerating ResearchAmerican Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc.Volume 16, Number 1, January 2010HVAC accepted October 9, 2009A novel cooling design featuring a
2、 two-stage expansion process is proposed in this study. With-out any help from insulation, the design can minimize or even entirely eliminate condensate formation outside the cold-plate surface even for a very low evaporation temperature. The design incorporates a double-pipe inlet/outlet and a two-
3、container cold plate, and its perfor-mance is compared to a conventional cold plate. For the conventional cold plate, the outlet pressure/temperature hold quite steadily at a light heat load, and the outlet pressure/temperatureshows a substantial rise when the refrigerant is completely evaporated. T
4、he outlet pressure/ temperature of the condensate-free cold plate shows an opposite trend at a heavy heat loading. The benefit of this characteristic is that the wall surface temperature and the outlet tempera-ture of the cold plate can be maintained comparatively more steady than the conventional c
5、old plate. INTRODUCTIONCurrently complementary metal-oxide semiconductor (CMOS) chip technologies act as the major class of integrated circuits and are widely applied in microprocessors, microcontrollers, static random-access memory (RAM), and other digital logic circuits. With the ongoing rise of c
6、apability, there is a strong demand for cooling to ensure the performance of microprocessors, microcontrollers, static RAM, and other digital logic circuits. In particular, it is known that the performance of CMOS can be drastically improved if the temperature can be further reduced. There are many
7、advantages (Ghibaudo et al. 1992), e.g., higher carrier mobility, higher satura-tion velocity, better turn-on capabilities (sub-threshold slope), latch-up immunity, improved reliability due to activated degradation processes, reduced power consumption, a decrease in leakage currents, a lowering of i
8、nterconnection resistance, increased thermal conductivity, and a reduction of thermal noise, in addition to the low temperature operation. Moreover, it is well known that operating semiconductor devices at lower temperatures leads to conspicuously improved performance (Taut et al. 1997). This is bec
9、ause of faster switching times of semiconductor devices and increased circuit speeds due to lower electrical resistance of interconnecting materials at low-temperature operations (Balestra and Ghibaudo 1994). Depend-ing on the doping characteristics, attainable performance improvements range from 1%
10、 to 3% for every 10C (50F) lower transistor temperature (Phelan 2001). However, in addition to the phys-ical limit of shrinking the size of the integrated circuit, the accompanied heat generation becomes Yu-Lieh Wu is an assistant professor and Min-Yi Chen is a graduate student of the Department of
11、Refrigeration, Air Conditioning and Energy Engineering, National Chin-Yi University of Technology, Taiping City, Taiwan. Kai-Shing Yang is an assistant professor of the Department of Electro-Optical and Energy Engineering, MingDao University, Chan-ghua, Taiwan. Chi-Chuan Wang is a senior lead resear
12、cher of Energy hence, alternatives such as heat pipes, liquid immersion, jet impingement and sprays, thermoelectrics, and refrigeration (Tru-tassanawin et al. 2006) must be considered. Of the available alternatives, only thermoelectrics and refrigeration can provide a sub-ambient operation that is q
13、uite attractive for high-flux appli-cations. In practice, refrigeration is capable of operating at a high-temperature ambient, yet its coefficient of performance (COP) is well above the present thermoelectrics system. There are also other advantages for exploiting refrigeration cooling (Taut et al.
14、1997), such as maintenance of low junction temperatures while dissipating high heat fluxes, potential increases in micropro-cessor performance at lower operating temperatures, and increased chip reliability. Investigations reported for cooling of electronic devices via refrigeration were mainly rela
15、ted to the fundamental system performance, such as junction to ambient air thermal resistance, system COP of the refrig-eration system (Phelan and Swanson 2004), and transient response behavior (Nnann 2006). Some refrigeration cooling systems for electronics are already available (e.g., see Schmidt
16、and Noto-hardjono 2002, Thermaltake 2009, and Bash et al. 2002). However, as pointed out by Agwu Nnanna (2006), there are two major concerns when using refrigeration systems to cool electronics. The first is associated with the condensation on the sur-faces subject to sub-ambient operation, and the
17、second is the systems lagging response to applied load at the evaporator. Note that condensation takes place when the temperature is below the dew-point temperature of the surrounding air. The presence of water condensate can bring hazards to the electronic system and must be avoided at all times. T
18、ypical solutions may involve clumsy insulation or use an additional heater to vaporize condensate outside the cold plate (Asetek 2009). The former requires considerable space that is often quite limited in prac-tice and is apt to reduce the overall system performance due to blockage of the airflow.
19、The lat-ter design not only raises problems in control but also incurs additional energy consumption. In view of the shortcomings of these two common solutions, the present study offers a novel design to entirely eradicate the influence of condensate. Performance of the proposed concept is then comp
20、ared with the conventional cold plate.NOVEL CONCEPT TO MINIMIZE/ELIMINATE CONDENSATEA refrigeration system normally has four major components, namely a compressor, a con-denser, an expansion device, and an evaporator (a cold plate placed above the heat source for heat removal). For this study, the r
21、efrigeration system includes a variable-speed drive compres-sor, a water-cooled condenser, a metering valve and a capillary tube as the expansion device, and a cold plate as the evaporator. R-134a is used as the working refrigerant. The most crucial design of this concept lies in the process of the
22、pressure-reducing device. During the pressure-reducing process, refrigerant temperature is also reducing. For the conven-tional design, a low temperature well below the ambient temperature is achieved through an expansion device (capillary tube, expansion valve, short tube, and the like), resulting
23、in a surface temperature lower than the corresponding dew-point temperature. Therefore, condensation takes place on the surface if there is no thermal insulation. To tackle this problem, we have proposed a two-stage expansion design that is capable of minimizing or even entirely eliminating the form
24、a-tion of condensate along the cold plate without the help of insulation. The basic principle of this novel idea can be seen from the T-S diagram shown in Figure 1. The idea incorporates the two-stage expansion in the cold plate. For the first stage of expansion, a capillary tube (or expan-sion valv
25、e) with a shorter length is placed after the condenser, thereby giving rise to a moderate pressure drop of the refrigerant from the condenser. For the second stage of expansion, a two-container cold plate as shown in Figure 2a is fabricated. As seen, the temperature of the 2010, American Society of
26、Heating, Refrigerating and Air-Conditioning Engineers, Inc. (www.ashrae.org). Published in HVAC its schematic is shown in Figure 2b. Figure 4 shows the difference between this novel design and the conventional design during operation. Tests were conducted at the same environmental con-ditions. The a
27、mbient condition was maintained at Tdb = 27C1C (80.6F1.8F) and RH =55%5%, where Tdbis the dry-bulb temperature and RH denotes the relative humidity. As clearly seen in Figure 4a, humidification occurs at the outer surface of the conventional cold plate with condensate (or droplets) being easily iden
28、tified. As seen in Figure 4b, no condensate forms outside the cold plate surface in the present two-stage design. In fact, with the double-pipe design, complete elimination of the condensate at the outlet connection piping is achieved.For a more detailed comparison of the performance of the conventi
29、onal cold plate with the present condensate-free cold plate design, one can see the associated variations (pressure/ temperature) at the outlet vs. heat load in Figure 5. As seen in the outlet pressure in Figure 5a, there is a distinct difference between the two-stage cold plate and the conventional
30、 cold plate. For instance, the outlet pressure of the conventional cold plate holds rather steady when the sup-plied heat is lower than 100 W (0.095 Btu/s), yet it starts to rise for a further increase of supplied load. Notice that the length of the capillary tube is fixed (ID = 1.0 mm 0.0394 in., l
31、ength = 300 mm 11.81 in.) for the conventional cold plate, thereby the refrigerant mass flow rate is roughly fixed for the conventional cold plate. In this regard, both the low-side pressure and the outlet temperature are increased provided the refrigerant is completely evaporated in the cold (a) (b
32、)Figure 4. Typical test results of the present two-stage expansion concept relative to the conventional cold plate: (a) conventional design and (b) present design. 2010, American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc. (www.ashrae.org). Published in HVAC the temperatur
33、e of the refrigerant is then lowering further in the second container to perform heat exchange with the heat source. For complete removal of condensate formation from the whole refrigeration system, the present design uses a double-pipe inlet/outlet for the cold plate with the warmer refrigerant flo
34、wing into the cold plate being placed at the outer tube while the colder refrigerant vapor leaving the cold plate is placed in the inner tube. Through this design, the entire refrigeration system, including piping, is free from condensation formation.A detailed comparison is also made between the co
35、nventional cold plate and the present condensate-free cold plate. For the conventional cold plate, the outlet pressure/temperature hold quite steady when the heat load is light, yet the outlet pressure/temperature show a substantial rise when the refrigerant is completely evaporated. In the meantime
36、, the outlet pressure/temperature of the two-stage cold plate shows an opposite trend at a heavy heat load. The benefit of this characteristic is that the wall surface temperature and outlet temperature can be maintained comparatively steady.ACKNOWLEDGMENTSThe authors are indebted to the financial s
37、upport from the Bureau of Energy, the Ministry of Economic Affairs, Taiwan, and the 10% innovative funding from Industrial Technology Research Institute.NOMENCLATURECOP = coefficient of performance, dimensionlesscp= specific heat, J/kg/K (Btu/lbm/F)icond= enthalpy change across con-denser, J/kg (Btu
38、/lbm)dP/dz = two-phase pressure gradient, Pa/m (psi/ft)= cooling water mass flow rate, kg/s (lbm/s)= total refrigerant mass flow rate, kg/s (lbm/s)P = pressure, Pa (psi)P = pressure drop, Pa (psi)= heat transfer rate at condenser, W (Btu/s)= heat transfer rate at evaporator, W (Btu/s)= supplied heat
39、, W (Btu/s)= heat loss, W (Btu/s)RH = relative humidity, %S = entropy, J/K (Btu/F)T = temperature, C (F)Tc,out = outlet temperature of cold plate, C (F)Tw = wall temperature of cold plate, C (F)Tdb = dry-bulb temperature, C (F)Twater,in = inlet temperature of cooling water, C (F)Twater,outy= outlet
40、temperature of cooling water, C (F)= two-phase friction multiplier for total flow assumed liquid, dimensionlessmwaterm134atotalQcondQevapQinQlossLO2 2010, American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc. (www.ashrae.org). Published in HVAC accepted August 12, 2009This
41、paper presents a new strategy for robust temperature control of variable-air-volume (VAV) air-handling units (AHUs) that can deal with dynamic variations and the associated constraints in a straightforward manner. The dynamics of VAV AHUs is described by a first-order-plus-time-delay model, and the
42、dynamic variations of the process gain, time con-stant, and time delay are described using uncertainty sets. These uncertainty sets are converted into an uncertainty polytope, and then an offline robust model predictive control (MPC) algo-rithm is employed for robust control design. The design proce
43、dure is illustrated step by step. An operating mode identification scheme is also developed based on the operating characteristics of VAV AHUs in order to reduce the size of uncertainty sets and, hence, to improve the control performance. Case studies are performed on a simulated VAV AHU. Results ar
44、e presented to show that the proposed control strategy is able to enhance robustness without much user inter-vention, reduce the control activities, and satisfy constraints when implemented in the tempera-ture control of VAV AHUs. INTRODUCTIONAir-handling units (AHUs) are widely equipped in modern c
45、ommercial buildings to treat the supply air for maintaining a desired indoor environment. There are two basic types of AHU: constant air volume (CAV) and variable air volume (VAV). VAV AHUs provide supply air at a predefined temperature setpoint. Since VAV AHUs vary airflow according to load conditi
46、ons and are efficient in energy consumption, they have become increasingly popular. A typical VAV AHU is illustrated in Figure 1, in which the cooling/heating coil serves as a heat exchanger. The supply fan moves the mixed air from outside through the heat exchanger, and the air is cooled (or heated
47、) by the chilled (or hot) water through the coil. The conditioned air is then delivered to occupied space. VAV AHUs typically include temperature, humidity, pressure, and flow rate controls. Among these, the temperature control of VAV AHUs is not trivial, because a typical VAV AHU is characterized b
48、y (1) dynamic and static nonlinearities and frequent changes in operating point, (2) information-poor data due to few and low accuracy sensors, and (3) poor resolution from A/D and D/A converters (Salsbury 2005).Gongsheng Huang is an assistant professor in the Division of Building Science and Techno
49、logy, City University of Hong Kong, Kowloon, Hong Kong; Shengwei Wang is department chair and a professor in the Department of Building Ser-vices Engineering, Hong Kong Polytechnic University, Kowloon, Hong Kong; and Xinhua Xu is a professor in the De-partment of Building Environment and Services Engineering, School of Environmental Science and Engineering, Huazhong University of Science and Technology, Wuhan, China. 2010, American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc. (www.ashrae.org). Published in HVAC Gouda e
