Combining Technologies for a High-Power Solution
Liquid cooling is a time-proven approach to maintaining safe operating temperatures of high heat flux electronics. It’s effective, capable of rejecting high heat loads and well understood. Although liquid cooled approaches are active, meaning they consume power and have mechanical components, they are often the best solution when power and temperature requirements demand it. In some cases, the liquid cooling lines need to be routed close to the heat-generating components to maintain thermal goals. This can, however, present a significant risk to the system in the event of a leak. If there were to be a fluid leak in the system, it would result in catastrophic failure to the electronics system.
Risk Mitigation to mitigate fluid leaks
One way to mitigate the fluid leak scenario is by decoupling the ultimate heat sink (liquid) from the heat-generating devices. By adding heat pipes to transfer heat at low thermal resistance, you can increase reliability with a minimal performance impact. The added distance between electronics and coolant could lead to a thermal penalty.
In the example below, ACT designed, developed, and tested a thermal solution that is capable of transferring heat from within an electronics chassis to an external liquid cooling loop. This included the heat pipe heat spreader design as well as the series/parallel flow channel liquid cold plate that acted as the ultimate heat sink. The electronics assembly, referred as a slat, has a fixed thermal resistance that is set by the customer’s electronics design. This slat is mounted to a heat pipe embedded, or HiK™ heat spreader, which transfers the heat through a series of interfaces and aluminum conduction paths to the external liquid cold plates which are mounted on the top and bottom of the electronics chassis.
Each HiK™ heat spreader in this assembly transfers 180W. The power is split between the two cooling loops with each handing 90W of power. This is important to note since heat pipes work both gravity assisted, and against gravity. Since the heat pipes were designed to work in both of orientations, and have a fixed thermal resistance, the electronics were able to maintain symmetric thermal gradients using both the upper and lower HiK™ heat spreaders. There is a combination of 16 HiK™ plates within this system for a total heat load of 2,880W.
Customer Collaboration- What does the design process look like?
ACT began the effort with a detailed thermal analysis that mapped the thermal resistance network from the heat-generating components to the chilled water supply. Several design iterations were required to determine a cold plate design that met both temperature uniformity and pressure drop goals. The final solution was a serpentine tube-in-plate design to ultimately meet necessary thermal requirements. Each serpentine removed heat from four of the HiK™ plates, four serpentines were used in parallel to reject heat from the 16 HiK™ plates within the system as seen in the Figure 1 CAD drawing below.
Figure 1: CAD Rendering of a HiK™ plate enhanced electronics chassis
Manufacturing
Following the thermal analysis of the proposed system and the customers’ approval, the ACT team was ready to move forward with manufacturing the prototype. The first step in production was manufacturing the heat pipes, aluminum components and the manifold assembly. Commercial off-the-shelf cold plates were used in lieu of an all-aluminum assembly in order to simplify and expedite the prototype lead time. Following assembly of the prototype, heat loads were simulated using film heaters that matched the heat input footprint provided by the customer and chilled water was supplied using a commercial chiller. An image of the test setup is shown in Figure 2.
Figure 2: HiK™ Chassis prototype in thermal performance testing
Testing
Results of the testing in comparison to the thermal predictions are shown in Figure 3. The max junction temperature was predicted to be 118.0°C whereas the assembly measured 118.2°C. The temperature drop highlighted in the red boxes on the left of Figure 3 represents the junction-to-case temperature drop within the electronics slot. This information was provided by the customer and is fixed in the comparison. There are minor differences between the predicted and measured interface materials. The predicted temperature drop of the cold plate was 11.7°C and measured a 14.5°C. One key temperature drop to highlight is for the HiK™ plate. The prediction estimated a 4.1°C temperature drop, whereas the assembly measured only 2.9°C.
Figure 3: Predicted Thermal Results (Top) and Measured Results (Bottom)
Summary
It’s not uncommon for large chassis to have such high waste heat that it requires liquid cooling. While this is less reliable than conduction or air-cooled solution, there are methods to increase reliability by separating the liquid flow and potential leak points from the high-value electronics. By using heat pipes to transfer the heat from the source to the liquid-cooled sink, you can vastly reduce thermal resistance in this conduction path and keep the liquid loop a safe distance from the electronics. This hybrid approach can provide both high thermal performance and high reliability when compared to alternative solutions.
HiK™ Chassis Images
Discuss your application with ACT engineers today!
Experimental And Modeling Analysis Of A Large-Scale Two Phase Loop Thermosyphon
Debraliz Isaac Aragones1,2, Chien-Hua Chen1,∗, Justin A. Weibel2, David M. Warsinger2, Richard W. Bonner1
ABSTRACT
Liquid pumping requires large quantities of electrical energy, including about 7%of the energy of building heating, ventilation, and air conditioning (HVAC) systems. To replace pumped condenser-cooling tower water loops with a passive alternative system, we implemented a commercial-scale two-phase loop thermosyphon (TPLT). The unit consists of a 13 m riser integrated with a commercially available cooling tower and circulation water heater that simulates heat loads up to 25 kW. In addition to providing passive cooling capabilities, the cooling tower unit is also maintenance-free, reliable, and can operate in both dry and wet modes. This study characterizes the performance (temperature difference between the evaporator and condenser) and the two-phase flow behavior of the loop under various refrigerant charges. Sight glasses installed throughout the loop are used to identify the operating flow regimes in the riser and downcomer. Over the range of operating conditions, we identified that there is an optimal refrigerant charge range for a specific heat load at which near-isothermal operation can be achieved. We further developed a model to predict the operating flow rate and gravitational height in the downcomer and compared it with the experimental data. The results show that the model agrees with the experimental data, in particular the threshold gravitational head height at which there will be subcooled liquid at the condenser exit, indicating that it can be used as a predictive tool for identifying the optimal loop charge for a given heat load.
Keywords:
Gravity-driven two-phase loop thermosyphon, pressure drop, refrigerant charge mass.
NOMENCLATURE
Subscripts
𝑓 Liquid
𝑔 Gas
INTRODUCTION
Current heating, ventilation, and air-condition (HVAC) systems typically use open-loop evaporative cooling towers that require a condenser pump. This condenser pump accounts for about 7% of the total power consumption for building HVAC systems [1]. Eliminating the need for this pump would result in significant energy savings potential. This study presents the use of a two-phase loop thermosyphon (TPLT) to deliver high heat loads (up to 25 kW) to a cooling tower over a long height of 13 m without the use of a pump. A TPLT is a natural circulation heat transfer device driven by a fluid density difference and gravity. A schematic diagram of a typical TPLT is shown in Figure 1. A TPLT consists of an evaporator, riser, condenser (placed higher than the evaporator), and downcomer. A heat load is applied to the working fluid in the evaporator and the exiting two-phase mixture (or single-phase vapor) is driven upward by buoyancy forces. The working fluid travels up through the riser and into the condenser where heat is removed. The condensate returns to the evaporator through the downcomer due to gravity. Because a TPLT can ideally operate near- isothermally at the fluid saturation temperature, it has an excellent potential heat transfer capability of offering minimal thermal resistance. Additionally, it has no moving parts, making it more reliable in comparison to pumped cooling systems.
The performance of TPLTs for HVAC applications has been studied theoretically and experimentally at smaller scales. One study investigated the factors affecting the liquid head height in the downcomer as a function of the temperature difference between the evaporator and condenser, as well as refrigerant charge [2]. It was found that the fluid in the downcomer is not always a continuous column of single-phase liquid; increasing the evaporator-condenser temperature difference and the refrigerant charge will raise the liquid head. Another study explored the optimum refrigerant charge to minimize the temperature difference between the evaporator and condenser sections in a microchannel TPLT with respect to evaporator airflow rates and outdoor air conditions [3]. The cooling capacity of a TPLT integrated with a vapor compression cycle has also been studied experimentally [4, 5]. The current study aims to understand the TLPT operational flow regimes corresponding to various refrigerant charges and power inputs, and the correspondence between these regimes and system performance in terms of the evaporator-condenser temperature difference. Using the insight from the experimental evidence, a physics-based two-phase flow model was developed to predict the TPLT operating flow rate and liquid head.
EXPERIMENTAL METHODS
A TPLT system was constructed and implemented with commercial HVAC equipment, with a total height of 13 m. For each test at a fixed refrigerant charge, the power input was first set to 15 kWand then increased in 2.5 kWincrements up to 25 kW. The system used R-134a as the refrigerant at multiple different charge levels from 27 kg (60 lbs) to 49.5 kg (110 lbs) (in 4.5 kg (10 lb) increments); data were collected for each charge.
2.1 Flow Loop
Figure 2 shows a diagram of the system used to perform the experiments. The water heater operates using resistance coils and can be adjusted to deliver the desired power input up to a maximum of 25 kW. The heater is set up to pump the water to a flat plate heat exchanger (SWEP B50H) that serves as the evaporator (Figure 2 bottom left). The evaporator has four ports,
Q = 15 kW 17.5 kW 20 kW 22.5 kW 25 kW
two of which are connected to the water heater and the other two are connected to the TPLT system refrigerant. Both the water and the refrigerant enter through the bottom of the evaporator and exit from the top ports in a co-current fashion. The heated refrigerant becomes less dense and the bubbles formed are driven upward by buoyancy forces to the condenser through the riser.
The condenser used is a commercially available unit (Baltimore Air Coil, VC1-10-DM) with a 24.62 kW cooling capacity. The condenser consists of several parallel serpentine coils where the refrigerant is condensed back to the liquid phase. The condenser can accomplish the heat rejection in a dry mode, without the use of a water spray, or in a wet mode, with the use of a water spray. In this work, the condenser was operated in the wet mode. Gravitational forces return the refrigerant from the condenser to the evaporator through the downcomer. The refrigerant thereby cycles around the TPLT, passively rejecting the total heat load. Temperature and pressure measurements are taken at the inlets and outlets of the riser and downcomer, as labeled in Figure 2.
Sight glasses were also installed every approximately 1 m (3 ft) throughout the loop to allow for the identification of flow regimes.
2.2 Experimental Procedure
To begin each experiment at a set refrigerant charge, the water heater, pump, and condenser are turned on. The water heater is set to the desired power input. The system is given 30-60 min for the temperature and pressure profiles to reach a steady state before increasing the power input. The heat load is then incremented and the process is repeated up to the maximum power input (25 kW).
A turbine flowmeter (Omega FTB 1303) measures the volumetric flow rate in the downcomer. A representative example of the temperature measurements at the downcomer outlet (evaporator inlet) for the case with a 40.5 kg (90 lb) charge is shown in Figure3. The sharp increases in temperature correspond to an increase in power input (at the times indicated by the dashed vertical blue lines). For the experimental data analysis, the constant sections of the temperature profile (indicated by horizontal dashed red lines) are averaged over the corresponding time period. This results in a single time-averaged data point for each power input.
MODELING APPROACH
A parametric study was conducted using an engineering equation solver (EES) to investigate the effect of change in charge and heat load on the volumetric flow rate and liquid head height in the TPLT. A one-dimensional steady-state model was constructed using energy, mass, and momentum conservation equations. The heat load into the evaporator was assumed to be equal to the heat removed by the condenser. Under this assumption, the accelerational pressure drop terms cancel out for a complete cycle around the loop. The frictional pressure drop through the system lines was considered, but the pressure drop of the relatively shorter flow lengths through the evaporator and condenser was neglected. Minor frictional losses through the pipe fittings were included.
The mass balance was found using
𝑚̇ = 𝜌.𝑢𝐴 (1)
where 𝑚̇ is the mass flow rate of the loop, 𝜌. is the mixture density, 𝑢 is the flow velocity, and 𝐴 is the cross-sectional area.
The mixture density is found by
𝜌. = 𝜌𝑙 (1 − 𝛼) + 𝛼𝜌𝑔 (2)
where 𝜌𝑙 and 𝜌𝑔 are the density of the fluid in the liquid and gas phase, respectively, and 𝛼 is the void fraction parameter. The energy balance across the evaporator and condenser was found to neglect sensible heating of the fluid.
𝑄= 𝑥𝑚̇ 𝐻𝑓 𝑔 (3)
where ̇ 𝑄 is the power input, 𝑥 is the vapor quality, and 𝐻𝑓 𝑔 is the enthalpy of vaporization of the fluid. The pressure drop of the system is derived from a momentum balance considering effects due to acceleration (A) of the flow (heating or cooling), friction(F), and gravity (G):
(︃𝑑𝑃/𝑑𝑧)︃= −(︃𝑑𝑃/𝑑𝑧)︃𝐹−(︃𝑑𝑃/𝑑𝑧)︃𝐴−(︃𝑑𝑃/𝑑𝑧)︃𝐺 (4)
where,
(︃𝑑𝑃/𝑑𝑧)︃𝐺= 𝜌.𝑔ℎ𝑓 (5)
(︃𝑑𝑃/𝑑𝑧)︃𝐹= 𝑓2𝜙12 𝜌.𝑢2 (6)
𝑔 is the gravitational constant, ℎ𝑙𝑖𝑞𝑢𝑖𝑑 is the height of the liquid column in the downcomer, and 𝑓2𝜙 is the two-phase friction factor [6]. The Friedel correlation was used for 𝑓2𝜙 based on the recommendations in [7]. A flow chart summarizing the modeling calculations is provided in Figure 4. The user inputs the loop dimensions, fluid properties, and a maximum height constraint boundary condition for the distance between the evaporator and condenser. This condition is crucial to the model to capture the physical limitation of the liquid head height in the downcomer reaching the condenser. Also, inputs are the parametric variables of charge mass and heat load. The model then takes the first set of variables and solves for the riser quality, liquid height, and volumetric flow rate. The mass balance in this calculation only considers two-phase flow in the condenser. If the height constraint has not been reached, the solution is viable and the model proceeds to predict the same output quantities for the next set of variables. However, if the height constraint has been reached for some input conditions, this height is then fixed and the mass balance equation is adjusted to account for liquid backing into the condenser. The model then outputs only the volumetric flow rate and vapor quality (with the liquid head height being at the maximum). The mass balance of the condenser had a significant impact on the results because the condenser accounts for over half of the total volume of the TPLT.
RESULTS AND DISCUSSION
4.1 Sight Glass Observations
The sight glass observations allowed the identification of three different flow regimes dependent on the refrigerant charge amount and operating conditions. A summary of the observations can be found in Figure 5. A lower charge in the system resulted in single-phase vapor in the riser and a two-phase mixture in the downcomer (Figure 5a, Regime I). The downcomer is suspected to have a very low liquid head height in this regime. Increasing the charge of the system will increase the liquid column height in the downcomer. This ‘transitional’ regime consists of two-phase flow in the riser and both two-phase flow and single-phase liquid flow in the downcomer (Figure 5b, Regime II). When the system is charged such that the liquid head in the downcomer reaches the condenser, there is then a single-phase liquid flow in the downcomer (Figure 5c, Regime III).
4.2 Experimental and Modeling Results
As the target of the TPLT system is isothermal operation, the performance is evaluated using the largest temperature difference in the loop. That is the temperature difference between the evaporator outlet and condenser outlet.
Δ𝑇𝑙𝑜𝑜𝑝 = 𝑇3 − 𝑇1 (7)
This temperature difference is plotted as a function of the refrigerant charge in Figure 6a for each power input. These results indicate that an increase in charge increases the overall temperature difference across the loop. The pressure difference across the TLPT is used to explain the trends with power input at each of each charge because the loop operates with the refrigerant in a saturated state, where the temperature and pressure are directly related. At 27 kg (60 lb), in Regime I, Δ𝑇 is seen to increase with increasing charge. This is because the gravitational pressure head in the downcomer is low enough that the frictional pressure drop in the riser dominates. When the input power increases, the frictional pressure increases, resulting in larger static pressure differences and therefore larger (saturation) temperature differences. At 31.5 kg (70 lb) and 36 kg (80 lb), in Regime II , the temperature difference decreases slightly with an increase in power input. This is because the gravitational head is more dominant at these charge levels and the riser gravitational head (or liquid column height) decreases as the power input increases. At 40.5 kg (90 lb), there is a transition from Regime II to Regime III, depending on the heat input. For 15 kWand 17.5 kW, the trend matches the behavior in Regime II for 31.5 kg (70 lb) and 36.5 kg (80 lb); however, further increases in power result in subcooling at the condenser outlet and a considerable increase in the temperature difference. Observations using the sight glasses revealed that at high enough charges the liquid column backs into the condenser, characteristic of Regime III as described above. This is identified to be the cause of problematic subcooling at high charges. At 45 kg (100 lb) and 49.5 kg (110 lb), operation in Regime III subcooling across all heat inputs. However, it is interesting to see that, at each given heat input, the temperature differences do not become more severe from 45 kg (100 lb) to 49.5 kg (110 lb). This could be explained by the fact that the condenser has a large manifold and the additional 4.5 kg (10 lb) of charge is most likely floods the manifold first, therefore causing a negligible amount of additional subcooling. It is speculated that if the charge were to continue to increase further, there would be additional subcooling as the condenser manifold filled up and the liquid started to go into the condensing tubing. The measured volumetric flow rate for each test case can be seen in Figure 6b. For each charge, the flow rate expected increases with an increase in power input. For the highest charges of 36-49.5 kg (90-110 lb), operating in Regime III, the flow rates all collapse. This is attributed to the driving liquid head reaching the limit of the condenser height.
Overall, while the best performance occurs at lower charges, having some amount of excess charge does not have a severe negative impact on the system performance for the range investigated.
At 27 kg (60 lbs) the TPLT achieved a Δ𝑇 of less than 0.5◦𝐶 with corresponding flow rates of less than 0.75 𝑚3/ℎ (12.5 LPM). This is quite an improvement compared to conventional open-loop cooling towers which typically have a Δ𝑇 of 5◦𝐶 with a pumped water flow rate of approximately 0.66 𝑚3/ℎ (11 LPM)[8].
Regarding the model predictions, the liquid column height in the downcomer is calculated using the pressure and volumetric flow rate from the experiments (Figure 7a). The model is able to capture the overall trend of an increasing column height with charge till Regime III. The calculations of the liquid column height are expected to have a mismatch in the quantitative values because the mass flow rate is calculated using a single-phase liquid density, but there is a two-phase flow in the downcomer.
The modeling prediction of the flow rate compared with the experimental measurements is presented in Figure 7b. The model is able to predict the flow rate very accurately at charges of 36 kg (80 lb) and greater where the downcomer is a single-phase liquid. However, there is some discrepancy at lower charges. This is thought to be because the fluid in the downcomer is two-phase at the lower charges. The model calculates the volumetric flow rate in the downcomer using the liquid density rather than a mixture density and therefore underpredicts the flow rate.
CONCLUSION
An experimental and modeling analysis of a large-scale two-phase loop thermosyphon (TPLT) with a 13 m tall riser was conducted. The system was tested at seven different charges of R134a (27 kg (60 lb) to 49.5 kg (110 lb), in 4.5 kg (10 lb) increments) for five power inputs (15 kW to 25 kW, in 2.5 kW increments). Sight glasses installed throughout the loop enabled the identification of three operating flow regimes under various operating conditions. Regime I occurs at lower charges and results in single-phase vapor in the riser and a combination of a two-phase mixture and a single-phase liquid column in the downcomer. This regime resulted in the best performance with an evaporator condenser.
Δ𝑇 < 0.5◦𝐶. Increasing the refrigerant charge results in single-phase vapor and a two-phase mixture in the riser and increases the liquid column height in the downcomer (Regime II). Continuing to increase the charge results in the liquid column backing into the condenser and causes undesirable subcooling in the condenser (Regime III). The results show that the modeling predictions agree with the regime-specific trends in the experimental data for the system flow rate and liquid head height.
ACKNOWLEDGMENTS
This work is supported by DOE NETL under the contract DE-FE0031657. The authors thank Mr. Chad Burkholder, Mr. DennisWinters, and Mr. Phil Martin at Advanced Cooling Technologies, Inc. for their assistance in performing the experiments.
REFERENCES
[1] Morrison, Frank. “Saving water with cooling towers.” ASHRAE Journal Vol. 57 No. 8 (2015): pp. 20–33.
[2] Zhang, Penglei, Wang, Baolong, Shi, Wenxing, and Li, Xianting. “Experimental investigation on two-phase thermosyphon loop with partially liquid-filled downcomer.”
Applied Energy Vol. 160 (2015): pp. 10–17. DOI 10.1016/j.apenergy.2015.09.033.
[3] Ling, Li, Zhang, Quan, Yu, Yuebin, Liao, Shuguang, and Sha, Zhengyong. “Experimental study on the thermal characteristics of microchannel separate heat pipe respect to different filling ratio.” Applied Thermal Engineering Vol. 102 (2016): pp. 375–382. DOI 10.1016/j.applthermaleng.2016.03.016.
[4] Han, Linjun, Shi, Wenxing, Wang, Baolong, Zhang, Penglei, and Li, Xianting. “Development of an integrated air conditioner with thermosyphon and the application in mobile phone base station.” International Journal of Re- frigeration Vol. 36 No. 1 (2013): pp. 58–69. DOI 10.1016/j.ijrefrig.2012.09.012.
[5] Lee, Sunil, Kang, Hoon and Kim, Yongchan. “Performance optimization of a hybrid cooler combining vapor compression and natural circulation cycles.” International Journal of Refrigeration Vol. 32 No. 5 (2009): pp. 800–808. DOI 10.1016/j.ijrefrig.2008.12.008.
[6] Ghajar, Afshin J. and Bhagwat, Swanand M. Frontiers and Progress in Multiphase Flow I (2014). DOI 10.1007/978-3- 319-04358-6.
[7] Thome, John R. “Engineering Data Book III Two-Phase Pressure Drops.” (2006): pp. 13.1–13.34.
[8] “Water Piping and Pumps Technical Development Program.” Technical report no. Sigler Commercial. 2017. URL http://siglercommercial.com/wp-content/ uploads/2017/10/02-Chilled-Water-Piping-Pumps.pdf.
COOLING A POWER ELECTRONICS CONTROL CABINET
A power electronics enclosure for a test setup is full of high-power electronic components generating massive waste heat
This power electronics enclosure for a test setup is full of high-power electronic components generating massive waste heat. A customer found that the cabinet’s internal temperature was rising above acceptable temperatures and came to ACT because they were worried about overheating electronics causing downtime during testing.
The cabinet is set up inside a dusty warehouse with an average ambient temperature of 21°C. The 15 heaters inside of the cabinet are 100W each and create approximately 450W of waste heat. Before implementing a cooling solution, the internal cabinet temperatures were reaching 48°C, on average. This was putting the internal power electronics above their maximum operating temperature of 43°C and causing stress on the electronics; over time this can result in a power failure. The customer was proactive and reached out to ACT before the excess temperatures led to overheating of the components in the cabinet, potentially ruining the test setup. Adding a heat exchanger helped regulate the internal temperature, and the choice of an ACT Sealed Enclosure Cooler was ideal because NEMA rated sealing gasket ensures that the cabinet is not impacted by the dusty warehouse environment. An added benefit is that by choosing a NEMA 4X-compliant model, washdown operations are possible.
Impact of Installing an ACT-HSC-22
ACT’s HSC-22 sealed enclosure cooling unit provided the customer with a sealed cooling solution for their power electronics cabinet.
After installing the ACT-HSC-22, the average internal temperature of the power electronics cabinet dropped to 35°C, which is far below the maximum operating temperature of 43°C. After operating the test equipment for 2 minutes the internal cabinet temperature readout was 38°C, remaining the average temperature throughout the test.
The customer was pleased with this solution that provided the needed cooling while preventing worries about filters and regular cleaning inside of the cabinet. They also expressed satisfaction that the unit was quieter than an air conditioner.
Overview of the Solution
Cooling solution overview of power electronics cabinet temperature with and without ACT’s Sealed Enclosure Cooler
After installing the ACT-HSC-22, the average internal temperature of the power electronics cabinet dropped to 35°C, which is far below the maximum operating temperature of 43°C.
ACT-HSC-22 Sealed Enclosure Cooler for Power Electronics Cooling
Heat Sink Design for IGBT Cooling
Figure 1. Simulated Heat Sink Temperatures
Advanced Cooling Technologies (ACT) was selected by a leading company to develop an innovative design that would cool twelve IGBT modules below their maximum juncture temperature. In addition to the junction temperature requirement, ACT had to work within strict size, noise and temperature uniformity requirements as well. As is often the case with paralleling IGBTs, load sharing can be difficult to achieve without intelligent thermal designs.
Using ACT’s in-house code and CFD simulations, the design team paired a custom heat sink design with a fan that easily integrated into the customer’s assembly. The wide fin pitch selected for this design allowed for higher air flow rates from the fans, which was critical to achieving the temperature uniformity requirement. After running thermal simulations on heat sinks with and without heat pipes, it was decided that at this time, the customer would be able to meet their thermal requirements with a custom heat sink without heat pipes.
Utilizing its engineering experience and thermal software capabilities, ACT was able to design the customer a low-cost heat sink with excellent thermal performance. The analysis was also able to show that if IGBT power levels increase in the future, the benefits gained by upgrading this heat sink design to include heat pipes, are already known, and the change can be affected without needing changes to the rest of the system.
The key benefits for the customer of this production ready, fully integrated solution provided by ACT, are the temperature uniformity, low operational noise, and small footprint. Providing guidance for the next level thermal solution, for when power levels rise, is the type of added value gained from working with ACT’s experienced engineering staff.
Rhombus Energy Solutions: Cooling an IGBT Module
Figure 1. Modeling of Evaporator Thermal Performance
Rhombus Energy Solutions came to ACT looking to replace their liquid cooling system with a passive heat transfer solution. The primary heat source in the system was a large IGBT module, which produces a maximum of 4kW of heat. The goal was to design a solution that provided cooling to keep the IGBT module case temperatures below 85°C, while fitting inside Rhombus’s existing packaging.
To achieve the program goals, ACT developed a custom loop thermosyphon assembly comprised of a custom brazed aluminum evaporator, microchannel aluminum condenser, and fluid transport lines.
Figure 2. Angled Microchannel Condenser
The fluid in a loop thermosyphon is driven around the loop by gravity; this means that the pressure drop of the loop has to be lower than in a pumped system. ACT iterated through multiple evaporator, condenser, and transport line designs to determine the best combination, taking into account the performance, cost, and packaging of each design. Ultimately, an angled condenser was chosen to provide a large airflow area, while keeping the overall height of the assembly to a minimum. ACT designed-in flexible lines between the evaporator and condenser to improve the installation of the cooling system. The final solution is predicted to keep the IGBT case temperature below 79°C, providing ample cooling for the customer.
The key benefits of ACT’s solution are an increase in both the reliability and power consumption of the cooling system. This will increase Rhombus’s efficiency rating, and improve the reliability and maintenance requirements for their system.
Loop Thermosyphon for Power Electronics / Power Grid Applications
The power grid is facing ever-mounting challenges as the demand for residential and commercial electricity increases. North Carolina State University is committed to modernizing the electric grid and creating energy leaders. NCSU’s FREEDM System Center was involved in a recent project developing a Medium Voltage Solid-State Transformer (SST). Their goal was to increase energy density, simplify system design, reduce costs, and improve efficiency system reliability. Of course, these drive use and dissipate kW levels of heat so finding a thermal solution that achieved their program objectives was a challenge.
A major thermal puzzle was how to effectively and efficiently dissipate 2-10Kw of heat from up to 6 power modules on a single heat sink assembly. The proposed layout is seen in Figure 1.
Figure 1. ACT’s Loop Thermosyphon for NCSU’s FREEDM System Center
For the thermal experts at ACT, there was an obvious solution: ACT’s Loop Thermosyphon. This rugged, reliable and essentially passive solution was specifically designed for these Power Electronics applications.
Working with the NCSU team, ACT designed and developed the Loop Thermosyphon seen below in Figure 2.
Figure 2. ACT’s Loop Thermosyphon
The device was capable of dissipating 2 kW of heat and achieved a maximum device junction temperature of 85°C with a coolant temperature of 60°C.
NCSU was delighted with the results and pleased with the performance of working with the ACT. Anup Anurag, a key person on the team, commented:
“In my opinion, ACT was one of the best companies we have worked with (not just in terms of thermal design). We got the products delivered as promised (before the promised time) and they were very helpful and professional in keeping us updated about progress in the design and manufacturing. We will definitely do business with ACT in the future.”
Got a difficult thermal challenge? Give us a call. ACT’s thermal experts design solutions to thermal challenges on projects in diverse markets and has particular expertise with emerging technology for the power grid. ACT is proud to help to create a more reliable technology for the future. Contact ACT for your power electronics and power grid applications today.
A True Partnership in Support of Our Customer’s Success
Successful Silicon Valley tech startups have high aspirations, tight deadlines and limited resources. For over two years now, ACT has proven its agility in meeting the dynamic needs for one such company.
Smart Wires Inc., an innovative technology company located in the San Francisco Bay Area, has developed a revolutionary product, the Power RouterTM, which adjusts the flow of power across grid lines by controlling impedance levels. This control enables power companies to push or pull power as required, offering customers a more reliable source of electricity. The product utilizes power-electronics devices which need well-engineered thermal-management solutions for enhancing reliability. That’s where ACT came in.
In the beginning, the tech startup brought in 50 experts in some 30 different disciplines from all over the world for an on-site brainstorming session to review the many features of their initial design. ACT was invited as a thermal management expert. From the first brainstorming session, ACT’s role grew. ACT’s engineers took responsibility for both thermal and mechanical design and the manufacturing of critical components. With ACT’s support, the Smart Wires Product Development team completed 11 different prototype designs in just over 18 months. ACT has since delivered over 50 prototypes of the final design. A significant portion of ACT’s work was performed and completed on site at Smart Wires. ACT engineers became true members of the team, even surfing and eating Ethiopian food with Smart Wires engineers.
According to Haroon Inam, Chief Technology Officer at Smart Wires, “We knew in the early development stages that we would need to address thermal issues and it would require a partner with industry-leading thermal-management engineering expertise. Within days of our request, ACT sent out an engineering team to help us pinpoint any thermal issues. They then designed and delivered effective solutions and have become a key part of our team. ACT has been a highly reliable and responsive partner.”
According to Devin Pellicone, Lead Engineer for ACT, “Working with the Smart Wires team has been an exciting challenge for us. They set, and met ambitious deadlines and goals that required us to rapidly develop innovative thermal solutions. It’s been very rewarding to help them position themselves for great commercial success. We believe this work has established a strong collaboration that will only get stronger.”
PUMPED TWO-PHASE COOLING FOR THERMAL MANAGEMENT OF HIGH HEAT FLUX DEVICES
As the power density of electronic devices and lasers continue to increase, new cooling strategies such as pumped two-phase (P2P) cooling are needed to maintain their temperatures below maximum operating limits and provide a necessary high degree of isothermality to optimize performance and lifetime. ACT has developed bench-scale and standalone two-phase cooling systems that can handle several hundred Watts per square centimeter of thermal power and provide excellent temperature control and uniformity.
Figure 1. This stand-alone, completely self-contained two-phase cooling system is capable of handling heat transfer rates ~ 300W/cm2, even when incorporating multiple evaporators that have different heat loads.
Figure 1 shows a stand-alone two-phase cooling system developed at Advanced Cooling Technologies, Inc. (ACT) that is capable of handling heat loads upwards of 300W/cm2 while maintaining tight temperature control and isothermality over the heat transfer area. In addition, the unit can handle multiple, separate, non-uniform and transient heat loads on the different evaporators. As noted, the unit is self-contained and simply has to be plugged in, setpoint temperatures input and thermal loads mounted to the two-phase evaporators. This apparatus has been used to test advanced two-phase evaporators, some of which are quite large and some of which include the use of microporous coatings to improve boiling performance. For more information on two-phase evaporators and the use of advanced boiling enhancement coatings, kindly refer to the following boiling enhancements page.
In the design of the unit, particular attention has also been given to address flow maldistribution between multiple evaporators and flow and thermal instabilities internal to the evaporator and at the system level. Regarding flow maldistribution, it is important since adequate flow to and within each evaporator is essential to avoid local dry-out and overheating. In addition, internal controls automatically adjust the saturation conditions of the refrigerant prior to the evaporator inlet(s) such that the refrigerant boils as it enters the evaporators providing optimal cooling. A two-phase mixture of liquid and vapor exits each evaporator, which is eventually condensed onboard the unit.
For those unfamiliar with pumped two-phase cooling systems, it should be noted that the key components in the system include a pump, preheater, surge tank, condenser, accumulator and evaporator(s). The surge tank consists of vapor and liquid at saturation; by controlling the pressure in the tank, the saturation (boiling) temperature of the working fluid can be controlled. The preheater heats the subcooled liquid exiting the condenser prior to entry into the evaporator(s) to again adjust saturation conditions and achieve optimal cooling performance.
HiK™ Plates to Improve Size, Weight, and Power (SWaP)
Aluminum HiK™ Plate
High Conductivity (HiK™) heat sinks can also improve the Size, Weight, and Power (SWAP) compared to standard heat sinks. Placing a discreet heat source on a large metal heat sink will produce large thermal gradients as the heat slowly conducts through the aluminum to the fins. Embedding heat pipes in a HiK™ heat sink can increase the thermal conductivity from around 180 W/m K to 500-1,200 W/m K, providing an opportunity to reduce heat sink plate thickness and fin area. This approach has been proven in a variety of weight/volume sensitive applications including: Ruggedized Electronics, UAVs, Handheld/Portable Devices, LEDs and optical devices.
Figure 1. A HiK™ natural convection heat sink reduces the mass by over 34% when compared with an all-aluminum heat sink with the same thermal performance.
Embedded heat pipes can improve performance and reduce the mass of forced and natural convection heat sinks. ACT fabricated a HiK™ heat sink and an all-aluminum heat sink with the same performance; see Figure 1. The total heat dissipation is 150W in both cases. The conventional aluminum heat sink is 12 inches (30.5 cm) long, weighs 9.6 lbs. (4.4 kg) and has a base thickness of 0.6 inch (1.5 cm). Introduction of 5 heat pipes, 3 in close proximity to the heat source and another two a little further out for improved spreading reduced the length to 10 inches (25.4 cm), reduced the thickness to 0.28 in (0.7 cm), and reduced the mass to 6.3 lbs. (2.9 kg) for an overall material reduction of over 34%. Thermal images that demonstrate the improvement are shown in Figure 2. The Hi-K heat sink seen on the right maintains the same source temperature, even though the heat sink is shorter, lighter, and thinner. The improvement is directly attributable to the addition of heat pipes which can be seen as red lines in the picture on the right.
-
Figure 2. Thermal images of the two natural-convection heat sinks show that the HiK™ heat sink has a similar performance to the standard heat sink, with a reduction in mass of over 34%.
Learn more about thermal solutions for power electronics here…
Pumped Two Phase Cooling for High Heat Flux Applications
With support from the National Science Foundation, ACT has developed a pumped two-phase cooling system for high heat flux electronic components and laser diodes and is now working on packing the system in a compact, user friendly, stand-alone platform. The two-phase system efficiently handles fluxes on ~ 300-500W/cm2, requires little pumping power, maintains device temperatures below operating limits and provides for a high degree of isothermality over the heated surface. Meeting these requirements is important in many applications including lasers whose emission wavelengths are temperature-dependent.
Unlike single-phase cooling, two-phase systems take advantage of the phase change (latent heat) of the coolant which enables it to handle higher heat fluxes for a given heat load. Two-phase cooling systems can however be more complex and prone to flow and thermal instabilities. As such, techniques to effectively manage instabilities have been developed and characterized. These include the application of engineered microporous coatings on the heated surface(s), which enhance boiling performance by increasing the number of nucleation sites together with a capillary-driven resupply of the coolant to the heated surface, which prevents or postpones dry-out (extends CHF).
A schematic of a pumped two-phase cooling system is shown in Figure 1. Key components include a pump, preheater, surge tank, evaporator (the heat sink), condenser and accumulator. The surge tank and the preheater differentiate this system from a traditional liquid cooling loop. The surge tank consists of vapor and liquid at saturation; by controlling the pressure in the tank, the saturation (boiling) temperature of the working fluid can be controlled. The preheater heats the subcooled liquid exiting the condenser to a temperature close to the saturation temperature before it enters the evaporator. This is important as boiling heat transfer is most efficient at saturation (minimal subcooling).
Figure 1: Schematic of a pumped two-phase cooling system
Also shown in Figure 2 is a representative copper minichannel heat sink coated with a microporous coating.
Figure 2: Minichannel heat sink with porous sintered power coating
The pumped two-phase cooling system shown in Figure1 was fabricated and evaluated. The heat transfer coefficient (HTC) [W/m2K] and the Incipient Wall Superheat [K] were determined as a function of the input heat flux and coolant mass flux [kg/m2s] using refrigerant R134a and others. Representative results for the HTC for coated and uncoated minichannel heat sinks are shown in Figure 3. Clearly, the HTC is higher for the coated heat sinks and the CHF is extended.
Figure 3: Heat Transfer Coefficient (HTC) associated with Minichannel Heat Sinks noting the increase in CHF for coated heat sinks.
The incipient wall superheat is also shown in Figure 4 as a function of input heat flux on coated and uncoated heat sinks. Again, the microporous coating enhances the thermal performance of the heat sinks as evident by lower values of incipient wall superheat; in other words, a heat-generating device mounted on a two phase heat sink with microporous coating will be maintained at a lower temperature compared to one that is mounted on an identical uncoated heat sink.
Figure 4: Incipient wall superheat decreases with the application of microporous coatings
In addition to improving the heat transfer, the thin copper coatings help suppress two-phase flow instabilities. Figure 5 shows flow instabilities with a bare copper cold plate, where the vapor flow occasionally reverses and enters the input plenum. In contrast, instabilities are suppressed in the cold plate with a thin porous coating, see Figure 6.
Figure 5. High heat fluxes with uncoated mini- and micro-channel systems can have unstable flow. Note the intermittent vapor flow back into the inlet plenum.
Figure 6. The maximum heat flux in two-phase systems can be increased by adding a thin porous layer. An additional benefit is that the flow becomes more stable.
In short, pumped two phase cooling systems have been developed. Flow and thermal stability issues were well managed with the use of porous coatings, which increase the heat transfer coefficient and extend the CHF. Pumping power required is minimal and the application of the coating does not increase the pressure drop in a measurable way as the coating thickness is very small compared to the channel dimensions. Additional work on flow boiling heat transfer in minichannel heat sinks is ongoing with a focus on maximizing performance and understanding the parameters (i.e., the coating properties – thickness for a given particle size, etc.) that affect performance for specific applications.
For more information on Pumped Two Phase Cooling, contact The Thermal Experts at ACT.
Manufacturing Quality Control
A consumer electronics manufacturer sought out ACT to perform an assessment of a test set that was providing high reject rates in production. This test was used by its supplier for quality control in high-volume production of a thermal module. The test set was giving high reject rates for the production thermal modules. The consumer electronics manufacturer sought out ACT for both our expertise in manufacturing and quality control. ACT replicated the test set and repeated multiple series of thermal measurements on the thermal modules. ACT then performed a comprehensive Repeatability and Reproducibility study. ANOVA, Kolmogrov-Sminov, Pearson correlation and other key statistical parameters were calculated. The analysis showed reproducibility errors for the test set to be far higher than the generally acceptable 10-30%.
Problem:
Instrumentation error was suspected because the products were rigorously tested in another, more controlled test apparatus and found to perform well. The root cause for the error was determined to be fixturing in the test set which permitted inconsistent thermocouple contact with the device being tested.
Solution:
ACT recommended a systematic re-design of the test set. ACT also helped develop correlations for other test sets for different operating conditions. The client was extremely satisfied with the results from the projects and stated their process quality control has taken a significant step forward because of ACT’s contributions.
Power Module Heat Pipe Heat Sink
ACT Offers Customized Heat Pipe Heat Sink Solutions
Advanced Cooling Technologies, Inc. can deliver a customized heat sink heat pipe solution for companies in industries such as aerospace, electronics and many others. As a premier thermal management solutions provider with a strong focus on research and development, we can provide heat pipe heat sinks for your specific applications.
We can also handle every aspect of your heat pipe heat sink project from conception through production. As an ISO9001:2008-certified company, you can also count on us to deliver a high-quality heat sink product that will exceed your expectations.
We Take a Problem-Solving Approach to Heatpipe Heatsink Design
At ACT, we take great pride in our ability to solve problems for our customers. Here’s just one example of how we have helped our customers overcome a challenge:
A customer was designing a power module consisting of a series of IGBTs. They were seeking a heat sink design but had many constraints and requirements. In addition to power dissipation (almost 6kW during transient, up to 2.4KW steady state), cooling methodology, heat sink size, IGBT orientation and overall ruggedness requirements were all fixed.
ACT Provided an Effective Heat Sink Pipe Solution
To address the problem, ACT used in house developed analytical software to identify the optimal fin solution for the given geometric and flow constraints. This modeling provided a convection resistance, which was then incorporated into the FEA thermal analysis that included the electronics, thermal interfaces, and heat spreader. After a series of design iterations using strategically located vertical and horizontal heat pipes, a final prototype design proved feasible, satisfying the various performance objectives. Customer was so impressed with the thoroughness of the study that a second project to design and build prototypes was initiated.
Temperature Profile of Power Module with IGBT’s mounted on proposed heat sink design platform, operating under defined power dissipation operating conditions.
Contact Us to Learn More About Our Heat Pipe Heat Sink Design Process
With more than a decade of experience in solving difficult problems for customers in a wide range of industries, ACT can help you meet thermal management challenges in your operation. We’ll work closely with you to develop a customized heat pipe heat sink solution for your applications. Contact us today for more information or to schedule an on-site consultation.