Microchannel Heat Exchangers and Reactors

FIELD OF THE INVENTION

The present invention relates to micro-channel heat exchangers; more particularly, this invention relates to processes for making micro-channel heat exchangers.

BACKGROUND OF THE INVENTION

Microchannel based heat exchangers hold great promise for reducing heat exchanger weights and volumes. Since the heat transfer coefficient increases inversely as the channel diameter, the amount of surface area needed to transfer a given amount of heat decreases directly with the channel diameter. This opens up the possibility of creating extremely compact heat exchangers, of reduced weight, using very small channels. As the size of the channels and overall heat exchanger volumes are reduced, pressure drop is also increased. However, through the control of velocity, a reasonable pressure drop can be achieved in a compact microchannel heat exchanger configuration.

Given the great potential, work is being pursued to fabricate and deploy microchannel heat exchangers, particularly for high energy density microelectronics devices. For these high value components that must be cooled to avoid component damage, high cost microchannel heat exchangers can be justified. However, current microchannel fabrication costs cannot be easily justified for low value heat exchange applications. If a low cost microchannel heat exchanger fabrication technique could be realized, then a whole range of processes, requiring effective heat exchange, could be favorably impacted. In particular, space and weight limited transportation applications could greatly benefit from a low cost microchannel heat exchanger fabrication technique.

Many industrial processes could benefit from inexpensive microchannel heat exchangers. For example, the following industrial processes could benefit from inexpensive microchannel heat exchangers.

Energy and Power: Absorption cooling cycles, condensers, chillers, dehydration, fuel processors, fuel cells, heat recovery

Refining: Air separation, combined reaction/heat exchange

Chemical Processing: Pharmaceuticals, plastics, petrochemicals, acids, alkalis, fertilizers

Hydrocarbon Gas and NGL: Gas processing, liquids recovery, cryogenics

Advanced cooling technology absorption cycles, e.g., as used for cooling electronics, dominated by fluids heat transfer requirements would benefit greatly from highly compact, highly effective and low cost heat exchangers. Another area that should benefit is fuel processors and fuel cells for advanced power systems. It has been shown that there is great potential to shrink power system component sizes, e.g., so that they can be more easily fit within a vehicle. If costs are low, along with volume and weight, a heat exchanger that is cheap to produce and offers a significant reduction in volume and weight over an existing heat exchanger will find broad use in power system heat management.

Microchannel and mini-channel heat exchangers and reactors are being successfully used in many special applications, such as off shore hydrocarbon processing, petroleum refining, chemical processing, semiconductor and automotive air-conditioning. In addition, developments are proceeding for their applications in fuel cell systems, as mentioned above. Microchannel heat exchangers are also under consideration for use in building HVAC applications.

Petrochemical Processing: Heatric Division of Meggitt (UK) Limited (HEATRIC), is a UK-based manufacturer of microchannel heat exchangers, has a chemically etched and diffusion-bonded heat exchanger product line, which they have named PCHE (Printed Circuit Heat Exchanger). These microchannel heat exchangers have been used in the upstream hydrocarbon processing, petrochemical and refining industries since the early 1990s. The weight range of these microchannel heat exchangers can be from 1 Kg to 100 tons, with a maximum surface area of a single unit of about 108,000 ft². Design temperature and pressure ranges of these heat exchangers vary from −450° F. to 1650° F., and from 1 bar up to 650 bar, respectively. A single PCHE modular unit capacity has been as low as a few KW up to as high as about 90 MW. Besides gas processing applications, PCHE's are also being used for chemical processing, refining, power and energy applications. The PCHE manufacturing process uses chemical etching to create microchannels of 0.2 mm to 5.0 mm width on plates of 0.5 mm to >10 mm thick.

Automotive Air Conditioning: Microchannel heat exchangers are also used for the working fluid side of automotive air conditioning systems. An example is the Parallel Flow (PF™) microchannel heat exchanger by Modine Manufacturing Company.

Semiconductor Cooling: There are increasing demands for liquid cooling systems for high heat flux electronics, where space limitations are significant. Some specific applications include power electronics such as FPGA, blade servers, medical lasers and diagnostic equipments, military and aerospace avionics, industrial lasers, analytical instrumentation, telecommunications, high-performance personal computers and supercomputers cooling. One of the most important requirements for high-performance electronic equipment is to maintain the temperature below a limit, and achieve that cooling quietly. Electronics manufacturers have developed factory sealed, liquid-based cooling units for high value products. A typical liquid cooling system utilizes a pumped single-phase thermally conductive liquid to remove the heat from the electronic components. Microchannel cold plates and heat exchangers have been used in these kinds of applications. While microchannel cold plates alone can be used for applications with lower performance requirements, additional liquid-to-air heat exchangers, with higher heat flux capability, are also being used to transfer excess component heat into ambient air.

Building HVAC: Microchannel heat exchangers may also find use in building HVAC applications. A microchannel evaporator design, related to refrigerant flow maldistribution and the ability of the evaporator coil to effectively shed condensate, has been a technical challenge.

Fuel Cell Systems: Microchannel heat exchangers are being used for specific fuel cell system applications such as high-performance stack coolers, hydrocarbon fuel reformer, vaporizers and mixers. By making these reactors and heat exchangers more compact, thermal management is improved and the units can be better packaged for vehicle and mobile power type applications of interest. Microchannel technology is also being used for fuels reforming applications for fuel cells. As with other developers and manufacturers of microchannel heat exchangers, this technology is based on micromachining of plates that are then diffusion bonded to create heat exchanger, or reactor cores.

Various traditional and modern microchannel heat exchanger fabrication techniques are possible. While many channel scales can be covered by these techniques, the focus is on channel heights from 0.5 mm to 0.01 mm.

Miniaturized traditional fabrication techniques include classic machine shop manufacturing processes that are adapted to produce microchannels. Some examples of this equipment include miniature milling machines and sawing, with saw cuts of 0.025 mm width and spacing of 0.004 mm now possible. As an example of these traditional techniques, a heat exchanger can have channels of from about 0.071 mm to 0.14 mm for the larger channels. The device was constructed of micro-machined stainless steel plates stacked and diffusion bonded together. Micromachining, while capable of producing efficient and low volume microchannel heat exchangers, the delicate nature of the tools and the high accuracy and precision required make microchannel heat exchangers in this manner is expensive. Micro-electric discharge machining (Micro EDM) has also been demonstrated, with the use of very fine wires as electrodes. However, this method is relatively slow, which drives up fabrication costs. Laser machining is also a powerful tool that can handle a wide range of materials of interest. Focused ion beam machining can also be used for cutting operations, and can operate in the submicron regime. In both of these methods, the beams require sophisticated positioning technologies to create the many microchannels required for heat exchangers. This drives up fabrication costs.

Modern microchannel heat exchanger fabrication technologies can be classified as serial or batch processes. In serial processes, the part is machined in a step by step approach, which typically results in low material removal rates and low throughput. This is justified for specialized high volume applications, but is too costly for many heat exchanger applications of interest.

Batch micro-fabrication techniques are derived from the semiconductor sector. For this fabrication technique, there are two types: bulk micromachining and surface micromachining. In bulk micromachining, the object is constructed by removing material from a single raw material. These objects have good strength because they are derived from a single block of material. In contrast, surface micromachining creates an object by building up layers of material on a substrate, through a series of deposition, patterning and etching steps, with perhaps different materials used in the layers. The key step in forming specific structures is the etch process. It can be wet chemical, dry or plasma based. Anisotropic wet etching can be used to create V-grooves in silicon. Unfortunately, the process is very slow, and requires many hours. This is not a viable technique for low cost heat exchanger applications. Dry etch techniques are also possible, including the Deep Reactive Ion Etch (DRIE) approach, which uses alternating etch and polymer passivation chemistries. Given the complexity of the approach and the slow processing speed, this technique would also not be viable for low value applications.

High aspect ratio, i.e., channel height to channel width, object fabrication can also be accomplished by the LIGA technique. However, LIGA has failed to become widely accepted because of the difficulty in making masks and the high cost. Micro-fabrication by lamination of machined multiple thin metal sheets is used. Each sheet can be configured by a variety of the above techniques. For commercial HEATRIC microchannel heat exchangers, for example, chemically etched metal sheets are utilized. The machined thin sheets are stacked and diffusion bonded in a high temperature furnace under load. For stainless steel laminations, a temperature of 920 C is required at 4000 psi for over four hours. While this technique can produce useful heat exchangers and reactor cores, costs are higher than desired. It has been indicated that each lamination would cost approximately $500 per pound, which is a high cost for low value heat exchanger application.

In sum, conventional microchannel heat exchangers can transfer large amounts of heat within a small volume as a result of many small channels that facilitate heat transfer. However, costs of these units are high because of the expense to form the many microchannels using conventional micromachining or etch methods. In the known fabrication techniques, either a raw material block has material removed to create the channels, or a substrate has layers added to build up the channels. These processes require the use of precision machines and multi-step processes. This adds time and cost to the fabrication process, and makes these techniques too expensive for low value heat exchanger applications.

SUMMARY OF THE INVENTION

The invention provides a method of manufacturing heat exchangers or heat exchanger cores, particularly well suited for, but not limited to low value applications. The technique forms constructs, stacks these constructs according to the desired core configuration, e.g., a cross-flow core, then bonds the constructs together to form the core, or a building block for a larger core of a heat exchanger. Manifolds may then be attached to the core to complete an assembly.

A construct is a structure that has wires secured to a sheet, e.g., metal wires attached to a metal sheet. The sheet has about the same as, or slightly greater than the length and width of the completed core. The wires are attached to the sheet in a specific alignment reflecting the desired microchannel size. The wire height defines roughly the height of the channel. The hydraulic diameter of wires used in a construct can range from between 0.01 mm and 5 mm. The spacing between wires reflects the desired channel width.

When constructs are stacked together, the structure that will form the core, or a core precursor is formed. A channel for the core, or building block for a core of a heat exchanger, is formed by a sheet and a pair of attached wires of one construct and a sheet of another construct that is placed on the wires. Constructs are stacked together in this manner so that when the stacked constructs are bonded, the sheets and wires form the channels. The bonding technique used will depend on the construct material, which can be metal, non-metal or a combination of the two. For an all-metal construct the bonding is diffusion bonding or load assisted brazing (LAB) or brazing to induce a mechanical and metallurgical bond between the wires and sheets.

According to another aspect of the invention a cross flow heat exchanger is constructed from stacking constructs in alternating directions. This alternating channel direction is repeated to construct a basic “building block” cross-flow core. Once several of these building blocks are created, they can be stacked in any number, and manifolded, to address a needed capacity, or adapt to different fluids. For example, the constructs may be constructed of two types—the first using a first size wire adapted to form a first channel size for a first fluid (liquid) and a second using a second size wire (larger than the first size wire) adapted to form larger channel sizes for a second fluid (gas). Heating fins may also be added to the larger channel. Thus, a construct of the first type is stacked over every other one of the constructs of the second type, to thereby form a network of differently-sized channels to facilitate heat transfer between the liquid and gas.

According to another aspect of the invention a manufacturing method according to the invention uses low-cost material and techniques for assembly of constructs, thereby significantly reducing the manufacturing cost and promoting widespread utilization of microchannel heat exchangers. The following applications refer to conventional microchannel heat exchanger uses, however, it is contemplated that the inventive cores and manufacturing methods are capable of addressing many other applications where conventional heat exchangers are being used because of the higher manufacturing cost of current microchannel heat exchangers.

According to one aspect of the invention, a heat exchanger core constructed in accordance with the invention has a solid area of about 39%, or at most about 39%. The solid area for an etched microchannel is typically higher, e.g., a solid area of 55% for a HEATRIC microchannel heat exchanger. Therefore, for an equivalent heat transfer, the weight of the inventive core would be about 29%, or at least about 29% less than that of HEATRIC's etched microchannel heat exchanger.

As a result of chemical etching a HEATRIC PCHE, for example, has more material, per volume of heat exchanger, compared to a core constructed in accordance with the invention. This is because etching creates rounded channels, which are relatively closely spaced. Also, the sheets required to form such structure need to be about twice the thickness of sheets used to form constructs (assuming the sheet thickness is about the same as the height of a wire). With high unit manufacture, the higher volume of material used will result in higher cost than that for the inventive core. This will particularly be the case when exotic materials are needed to address either temperature or corrosion requirements.

According to another aspect of the invention a core and method of manufacturing a core of a heat exchanger can be used for the working fluid side of automotive air conditioning systems.

According to another aspect of the invention a core and method of manufacturing a core of a heat exchanger can be used for liquid cooling systems for high heat flux electronics, where space limitations are significant. Some specific applications include power electronics such as FPGA, blade servers, medical lasers and diagnostic equipments, military and aerospace avionics, directed energy weapons, industrial lasers, analytical instrumentation, telecommunications, high-performance personal computers and supercomputers cooling. One of the most important requirements for high-performance electronic equipment is to maintain the temperature below a limit, and achieve that cooling quietly. The leading performance electronics manufacturers have developed factory sealed, liquid-based cooling units for their high value products. A typical liquid cooling system utilizes a pumped single-phase thermally conductive liquid to remove the heat from the electronic components. While microchannel cold plates alone can be used for applications with lower performance requirements, additional liquid-to-air heat exchangers, with higher heat flux capability, are being used to transfer excess component heat into ambient air.

According to another aspect of the invention, a method of making a heat exchanger having microchannels includes assembling a plurality of constructs including continuously feeding a plurality of wire and a sheet into a press for securing the plurality of wire to the sheet; combining the plurality of constructs; and bonding the combined plurality of constructs to each other to form the heat exchanger.

According to another aspect of the invention, a method of making a construct for combining with other constructs to form a microchannel heat exchanger core includes providing a continuous supply of wires and sheet material; feeding the wires through a guide to align the wires with the sheet material, wherein a space between each of the aligned wires corresponds to about the width of a microchannel; attaching the aligned wires to the sheet material; and cutting the sheet material and wires to form the construct.

According to another aspect of the invention, a method of making a heat exchanger having microchannels includes providing a plurality of constructs; and stacking the plurality of constructs, wherein a microchannel is formed by the wire and sheet of a first construct and the sheet of a second construct placed on top of the first construct.

According to another aspect of the invention, an apparatus comprises a core precursor comprising a plurality of constructs; wherein a height of the wires is about the height of a microchannel; and wherein the apparatus is adapted for being formed into a plurality of microchannels of the heat exchanger core by bonding the plurality of constructs to each other.

According to another aspect of the invention, a machine configured to form constructs includes continuous supply of wires and sheet material, the wires and sheet material arranged to be continuously fed, or spooled into a press for attaching the wires to the sheet material, wherein the wires pass through a guide that spaces the wires according to microchannel width, and the press includes rollers.

INCORPORATION BY REFERENCE

All publications and patent applications mentioned in this specification are herein incorporated by reference to the same extent as if each individual publication or patent application was specifically and individually indicated to be incorporated by reference. To the extent there are any inconsistent usages of words and/or phrases between an incorporated publication or patent and the present specification, these words and/or phrases will have a meaning that is consistent with the manner in which they are used in the present specification.

BRIEF DESCRIPTION OF THE DRAWINGS

Non-limiting and non-exhaustive embodiments of the invention are described with reference to the following figures, wherein like reference numerals refer to like parts throughout the various views unless otherwise specified.

FIG. 1 shows a perspective view of a cross-flow heat exchanger core fabricated in accordance with the disclosure.

FIG. 2 shows a partial perspective view of a series of alternating constructs used to form the cross-flow heat exchanger of FIG. 1.

FIGS. 3B and 3A show top and side views, respectively, of two constructs of FIG. 2, with the FIG. 3A view being taken at section IIIA-IIIA in FIG. 3B.

FIG. 4 is a flow diagram illustrating a process for making a heat exchanger in accordance with the disclosure.

FIG. 5 is a schematic of a station for making constructs.

FIGS. 6A and 6B are two embodiments of a wire guide and straightening assembly used with the station of FIG. 5.

FIG. 7 shows a plurality of constructs aligned within a frame suited for being placed within a press for maintaining alignment among stacked constructs where the stacked constructs in the frame are a core precursor to a core of a heat exchanger.

FIGS. 8A and 8C show examples of wire meshes used to manufacture a construct according to an alternative method of making a construct.

FIG. 8B is a schematic of a second station for making constructs.

FIG. 9A, 9B, 9C are examples of heat exchangers made using core assembled according to processes described in the disclosure.

FIG. 10 shows a stiffened core structure.

DETAILED DESCRIPTION OF EMBODIMENTS

A “wire” is intended to mean a slender flexible strand or rod, which can be made of a metal, metal alloy, a non-metal or a combination of metal and non-metal. A wire can have a circular, triangular, elliptical, square, trapezoidal or rectangular cross section.

The terms “sheet” or “plate” are used interchangeable in the disclosure. Thus, use of the word “sheet” rather than a “plate” or “plate” verses “sheet” is not intended to exclude or require one type of structure over another, unless stated otherwise.

A “heat exchanger” means a body to exchange heat, whether to remove heat from a mass (fluid or solid) in contact with the heat exchanger or to supply heat to the body as in the case of a heat exchanger used in a reactor core. A microchannel heat exchanger, for purposes of this disclosure, means a heat exchanger having a channel height of between about 0.01 mm to 5 mm. Unless stated otherwise, whenever the word “channel” in the disclosure, it is intended to refer to a channel having a height of between about 0.01 mm to 5 mm, i.e., a “microchannel”.

The term “construct” as used in this description means wire secured to a sheet or plate, e.g., by adhesive, where the width and length of the plate or sheet corresponds to about the length and width of a core assembled using several of the constructs assembled in a stacked relationship, and each of the wires secured to the plate or sheet is spaced from adjacent wire by about the desired microchannel width for the core of a heat exchanger formed from the constructs. Since the wire is secured to the plate or sheet, the constructs may handled individually without risk of the wire separating or moving on the plate. The wire is spaced by “about” the desired microchannel width since it will be understood that the exact microchannel width for a finished core may change somewhat after constructs are bonded together.

The term “core precursor” refers to the structure assembled from constructs before bonding the constructs to each other to form the channels in the core of a heat exchanger. Constructs stacked according to the desired arrangement of channels for the core, e.g., FIG. 1, is a core precursor. Additionally, the total number of constructs that will be used to form the channels in a core, but not necessarily maintained in the stacked arrangement before being bonded to form the core, e.g., constructs held within a frame, is also considered a core precursor.

FIG. 1 is a perspective view of one example of a microchannel heat exchanger 10 fabricated using a method of manufacture according to the disclosure. The heat exchanger core 10 in this example includes a first plurality of channels 12 that extend through the core 10 in direction A and a second plurality of channels 14 that extend through the core 10 in direction B. When fitted with manifolds the core functions as a cross-flow type heat exchanger, which is adapted to transfer heat from a first fluid to a second fluid as the two fluids pass through the respective channels 12 and 14 orientated at 90 degrees to each other in alternating fashion (every row of channels are oriented 90 degrees to the rows of channels above and below it). As the fluids travels through the channels, heat is transferred via thermal conduction across the relatively thin walls separating adjacent channels. Thus, as a cold fluid, e.g., a refrigerant passes through the channels 12 and a hot fluid passes through the channels 14 heat is transferred from the hot to cold fluid by conduction across the channel walls. Each of the channels 12 and 14 extend entirely through the core and terminate at opening 12a and 14a, respectively, on opposite sides. When integrated into a heating or cooling system, the faces 10a, 10b are coupled to manifolds (not shown) to direct the fluid through the respective channels, with manifolds potentially attached at the fluid exiting faces to collect and direct the exiting fluid.

FIG. 2 shows a corner of a core precursor 99 that will be used to form the heat exchanger core of FIG. 1. Here is shown a plurality of stacked constructs 22, which when bonded together form the cross-channels for the core 10 in FIG. 1. Each construct 22 includes a foil, sheet or plate having spaced wires applied to one side of the foil, plate or sheet. Wire width and spacing can be equal or can vary. Wire width on the edges of sheet or plate can be extra wide to form a good sealing surface between sheet or plate and edge located wire.

FIGS. 3A and 3B show side and top views, respectively, of two such constructs 22a, 22b (FIG. 3A is a view taken at cross-section section IIIA-IIIA in FIG. 3B) from FIG. 2. IN this example the constructs 22a and 22b are identical to each other and formed from the same wire and sheet stock by a process according to the disclosure. Construct 22b is oriented 90 degrees relative to construct 22a. The wires 26a, 26b are evenly spaced over each sheet 24a, 24b and secured to the respective sheet (round, flat or other cross-section wires may be used). The two stacked constructs 22a, 22b form channels 28. A third construct (not shown) having wires oriented in the direction of wires 26a of construct 22a and placed on construct 22b will form channels between it and construct 22b that extend perpendicular to the channels 28 in FIG. 3A. This process of stacking constructs, one on top of another with wires arranged at 90 degrees to the wires of an immediately adjacent construct, produces the core precursor 99 having stacked constructs shown in FIG. 2. When bonded together, e.g., via diffusion bonding the stacked constructs form the channels of the heat exchanger. The constructs 22a and 22b after bonding and finishing of the edges will each have a length and width of LC and WC, respectively, which is the length and width for the core 10. When initially cut, however, their lengths are longer, i.e., LC+δL, as illustrated in FIG. 3B. The extra length (δL) is added when using a preferred cutting step in the manufacturing process, as explained below.

Referring to FIG. 3A, the wire used for each construct has a height H (in this example a square wire of height H) and the sheet used has a thickness t. The wires are evenly spaced by a distance W so that the channels 28 formed by the stacked constructs 22a, 22b have an aspect ratio (AR) of W:H. Thus, for 0.07 inch wire spaced 0.42 inches apart the aspect ratio for the channels 28 is 6:1 (after these constructs are bonded together, as explained below, the AR may be slightly higher).

As will be greater appreciated in view of the disclosure, in an alternative embodiment the construct 22a may use a different size wire than the wire used for construct 22b and/or the material used for wire 26a may be made of a different material that wire 26b. Such embodiments may be desirable for applications in which air passes through one channel and a liquid passes through the second channel.

The constructs 20 in FIG. 2 may be bonded into a solid structure in a high temperature, controlled atmosphere furnace. Therefore, a heat exchanger manufacture process according to one aspect the disclosure includes a process step similar to that currently utilized for conventional heat exchanger fabrication. Additionally, the wire and sheet or plate material used to make the constructs include stock material that is commonly used in a broad range of applications and are manufactured in high volume at low cost. And the fabrication process associated with stacking constructs on top of each to form channels, and ability to use flexible wire provides a modular building-block for exchangers of wide variety. The cores of bonded constructs can be joined together, with proper manifolding, to produce a range of heat exchanger capacities and types (e.g. cross-flow, counter-flow, and co-flow).

Referring again to FIG. 3A, in one embodiment constructs 22a, 22b consist of two thin metal plates, on the order of 0.5 mm thickness, separated by parallel oriented fine wire (flat wire in this case) of a similar composition, on the order of 1.0 mm diameter. The individual wires are separated by approximately six diameters, thereby producing rectangular channels with AR of 6:1 for flow. These channels therefore have H=1.0 mm and W=6×1.0=6.0 mm. Hundreds of these channels can be formed by stacked constructs, as described below, to form a basic microchannel building block core.

Heat exchanger thermal efficiency, or heat transfer to pressure drop coefficient ratio, is an important consideration for achieving very compact heat exchangers, with reasonable pressure drops. The higher the thermal efficiency, the lower the pressure drop, for given heat transfer. For microchannel heat exchangers, the channel Reynolds numbers, are very low, and the channel flow is laminar. Based on laminar flow theory, and proven by experimentation, thermal efficiency will vary with channel aspect ratio (AR). This is given in Table 2

TABLE 2

Thermal efficiency for different channel cross-section geometries

Cross-section

Nu (H)/
Thermal

geometry
Nu (H)
Nu (T)
f xRe
Nu (T)
Efficiency

equilateral triangle
3.11
2.47
13.33
1.26
0.263

rectangle (AR = 1)
3.61
2.98
14.2
1.21
0.286

Circle
4.384
3.66
16
1.19
0.307

rectangle (AR = 4)
4.384
4.44
18.3
1.20
0.328

rectangle (AR = 8)
5.33
5.60
20.6
1.16
0.355

rectangle (AR = ∞)
8.235
7.54
24
1.09
0.386

Where Nu (H) is the Nusselt number for constant heat transfer, Nu (T) is the Nusselt number for constant temperature, Re the Reynolds number for the channel, f the friction factor. As shown, the heat transfer coefficient increases substantially with increasing channel width, with an 8 to 1 channel having an 88% higher heat transfer coefficient than the 1 to 1 case (AR=1). However, as shown, the friction factor also increases, but not as strongly, resulting in a higher thermal efficiency as channel width increase. As shown, as the channel aspect ratio is increased, the thermal efficiency, given by StPr^2/3/f, increases, with the optimal configuration being an infinitely wide channel. However, this channel would not have the needed structural strength, for a reasonable separation plate thickness. Examples of material that may be used for construct fabrication are shown in Table 3.

TABLE 3

examples of foil, sheet, plate and wire material

Material
Form
Size Range (mm)

Stainless Steel 304
Foil/Sheet
Thickness .013-3

Aluminum 6061-T6
Foil/Sheet
Thickness .017-3

Copper 110
Sheet
Thickness .025-3

Inconel HX
Sheet
Thickness .5-3

Stainless Steel 304
Wire
Diameter .13-3

Aluminum 1100
Wire
Diameter .32-3

Copper 110
Wire
Diameter .13-3

Inconel HX
Wire
Diameter .078-3

For liquid/gas heat exchangers, the liquid side can have very small channels with the gas side having much larger channels to account for the much lower density gas for similar liquid and gas mass flows. In this case, the large gas channels formed by constructs can be filled with corrugated fins that add surface area and enhance gas heat transfer. In this embodiment, a heat exchanger constructed in accordance with the invention can therefore be made into a “radiator” type heat exchanger.

Heat exchangers constructed in accordance with the invention may be made from material other than metal, or may use combinations of metals and non-metals. In these cases, methods other than metal bonding methods (e.g. diffusion bonding, brazing, soldering) can be utilized, including adhesives and heat activation to make a heat exchanger from stacked constructs. Indeed, it is contemplated that the concept of stacked constructs using low-cost material according to the disclosure has considerable flexibility to address many temperature, pressure, fluid type and corrosion/oxidation conditions. For example, channel walls can be coated with washcoats and catalysts to promote simultaneous reaction and heat transfer. In the following discussion examples of constructs made from metal wire mesh and sheet, and round and flat wire and sheet cases are described in connection with manufacturing processes according to the disclosure.

Constructs made in accordance with the disclosure can provide very high heat transfer rates. As such, a comparatively small block of bonded constructs can produce about the same heat transfer as much larger conventional plate and fin heat exchangers. Also, these blocks will have very high effectiveness (i.e. approaching the theoretical thermal maximum) as a result of very high heat transfer rates. For example, it is projected that a 53 liter volume conventional plate and fin heat exchanger would be reduced in volume by 98%. Therefore, a block of constructs would be tens of millimeters per side, versus several hundreds of millimeters per side for a conventional plate and fin heat exchanger. Given this volume reduction, many more core precursors can be thermally bonded within a furnace, which reduces fabrication and furnace costs.

According to one aspect of the disclosure constructs 22 depicted in FIGS. 3A-3B are formed by securing tensioned wire to sheet material using a temporary adhesive. Two examples of this process are discussed below, which refer (for convenience purposes only, not to limit the scope of the disclosure) to using a flat or round wire to form constructs and with a core being a cross-flow design for a heat exchanger.

Once assembled with wires spaced as desired and corresponding to about the microchannel size for the heat exchanger, the constructs are stacked to form one or more core precursors. Next, the constructs are bonded to each other to form the channels of the core. In the illustrated examples both the wire and sheet material are metal. The bonding process is diffusion bonding or a Load Assisted Brazing (LAB) process to induce both mechanical and metallurgical bonding among the stacked constructs. After bonding the core is finished by cleaning sides to remove excess material from edges. After this step, one or more of the core structures will form the heat exchanger. The cores may be individually used as a heat exchanger (HEX) or stacked together to form a larger core from the combined building block cores. Manifolds and additional coupling structure for the fluid sources may then be attached to the core as desired. Several examples are discussed below. The foregoing process for making a HEX is summarized in the flow diagram of FIG. 4. Optional steps in this process are to secure stiffeners or plates to the core, which is discussed below. Depending on the environment of the HEX, stiffeners and/or support plates may not be needed.

Laminator

FIG. 5 is a schematic representation of a station for securing metal wire 30 to metal sheet material 40 to form the constructs 22. In this example the metal wire 30 is dispensed from spools 31, i.e., spooled, and aligned by a wire guide and straightening assembly 60 as it is advanced towards a press 50. The wire guide and straightening assembly 60 can have the three portions 62, 64 and 66 as shown, which assembly 60 both spaces and straightens the wire 30 (according to the spacing desired for microchannel width) to configure the wire 30 for the desired spacing on the sheet 40 disposed below the wire 30. The wire guide and straightening assembly 60 in FIG. 5 is shown in greater detail in FIG. 6B. For round wire the guide and straightening assembly 70 depicted in FIG. 6A may be used in place of the assembly 60 in FIG. 5.

Referring to FIG. 5, the sheet 40 is a continuous sheet having a width (WC, as shown in FIG. 3B) and sealing ribbons run along the length of the sheet 40 (left to right in FIG. 5) as it is dispensed from the roll 41. Upstream from the wire guide and tension assembly 60 is an adhesive which is disposed between the wire 30 and sheet 40. In a preferred embodiment the adhesive is a heat-activated adhesive dispensed from a roll 43.

As can be appreciated from FIG. 5 the wire 30, sheet 40 and adhesive 43 are continuously fed from left to right towards a press 50 to thereby facilitate a high volume production of constructs 22 on a continuous basis. After being brought together so that the wire 30 is aligned and spaced over the sheet 40, the wire 30 and sheet 40 pass between rollers 52a and 54c. The press 50 (having upper and lower belts 55, 53 that are moved using roller belts 54a, 54b, 52a and 52b as shown) applies a pressure to the combined structure (i.e., wire 30 and sheet 40) as it is moved left to right in FIG. 5. As the combined structure translates towards the right within the press 50, the adhesive 43 is heat activated at 43a, cooled at 43b and allowed to set before exiting at 51b. At the exit 51b of the press 50 the wire 30 is secured to the sheet 40 so that the combined structure can be later handled without risk of wires separating or moving on a sheet. At 59 a laser cutter may then cut the wire-sheet structure into a desired construct 22 size illustrated in FIG. 3B. Since the wire 30 is secured to the sheet 40, the cut constructs 22 can be handled, packaged, stacked, etc. for later use or to assemble core precursors as desired.

FIGS. 6A and 6B show the wire guide and straightening assemblies 70 and 60 for round wire and flat wire embodiments, respectively. Referring to FIG. 6B the sheet 40 and adhesive is moved under the flat wire 30 fed from upstream spools (not shown in this figure). The flat wire 30 is spaced by the guide 62. An additional guide 64 is used here to assist with straightening wire and maintaining the proper orientation of the flat wire 30 (e.g., a guide that straightens the flat wire as it is dispensed from spools 31). The flat wire 30 may be directed from guide 64 down to a plate 66 by an angle θ₁to place the wire on the sheet 40 surface. The plate face 66a may be adjusted by screws 66b.

Referring to FIG. 6A, a guide and straightening assembly 70 is shown for a round wire 30. In this example there are two guides 72, 74 and straightening pegs 76 disposed between the guides. Since the wire 30 is round, the straightening may be done by directing the wire 30 around the pegs 76, e.g., the illustrated single wire 30 is directed around the two pegs 76a, 76b in FIG. 6A. Alternatively, the round wire 30 may be spaced and straightened using the guide and straightening assembly 60 used for flat wire 30 (round wire 30 can use either of the assemblies).

As will be appreciated from FIG. 5 and the foregoing, the guide and straightening assembly 60/70 maintains upstream alignment and wire spacing as the combined structure passes between rollers 52a, 54c. As the wire 30 and sheet 40 pass between these rollers, the wire 30 is tensioned and held to the sheet 40. The wire 30 and sheet 40 are held together in proper alignment while the adhesive 43 is applied and allowed to set before exiting the press 50.

Bonding Process

After the constructs 22 are formed using the station 50, a plurality of the constructs 22 are stacked according to the number of channels and desired core type, e.g., 20 constructs stacked in alternating directions, one on top of the other to form a cross-flow core having 20 channels (an extra sheet is added on top or bottom to close the 20th channel space. Additionally or alternatively top and bottom reinforcing plates are added to the stacked constructs 22). The constructs 22 are held in this position securely before being bonded. The constructs 22 arranged in this manner, i.e., stacked one on top of the other, are ready for bonding. For purposes of holding the constructs together, a frame 80 (FIG. 7) separate from a press or furnace is preferably used, so that there can be mechanical and metallurgical bonding among the constructs while the constructs are maintained in alignment within the frame.

The frame 80 for core precursor 99 is depicted in FIG. 7 (the extra sheets and/or reinforcing plates mentioned above removed to show construct orientation in frame 80). The frame 80 is sized accordingly to the basic construct 22a, 22b length and width (FIG. 3B), and desired number of channels (i.e., height of the core). The frame 80 includes a base plate 82 that can be fixed in a furnace or a clamp. Pins 86a, 86b, 86c, 86d, 86e, 86f, 86g and 86h are received in holes of the base plate 82. Pins 86a, 86c, 86e and 86g are positioned at corners formed by the stacked constructs 22a, 22b of the core precursor 99. Pins 86b, 86d, 86f and 86h are located at the ends of the core precursor 99. These pins hold the core precursor 99 (and top and bottom sheets not shown) in alignment. At the two edges of each construct 22a, 22b of the core precursor 99 are securing ribbons 81 (rather than wire 30) to facilitate the sealing of edges. These securing ribbons 81 may be attached to edges of the sheet of the construct when the wires 30 forming the microchannel widths are attached, e.g., a thick, flat wire 81 dispensed from a spool 31 and continuously attached to edges of the sheet using the process described in connection with FIG. 5.

In one embodiment, an upper compression plate (not shown) is placed over the frame 80 and lies on the upper sheet that was placed over the upper construct 22a. The four slots of plate 82 are also present on the upper plate. Bolts (not shown), each having a threaded end, are placed in each of the upper plate and base plate 82 slots with a nut (not shown) used to hold the upper plate and base plate 82 together. The upper plate and base plate 82 apply a compression load on the core precursor 99. E.g., a press having a circular ram applies the compression load on the core precursor 99 and while this load is being the nuts are tightened down to maintain the compressive load through the upper plate and base plate 82; the frame 80 with core precursor 99 can then be removed from the press while the compressive load is maintain through the upper plate, base plate 82 and tightened-down nuts. This assembly is then placed in a furnace for bonding. The faces of the upper plate and base plate 82 that contact the core precursor 90 may have a ceramic, or non-stick surface to prevent adherence, sticking or bonding between the upper and lower sheets and respective faces of the core precursor 99 during bonding in the furnace. Prior to bonding residual adhesive can be removed from the constructs so that the adhesive does not contaminate the furnace. For example, a Methyl alcohol bath at 85-120F temperature was found to be effective to remove almost all adhesive. During furnace processing, any residuals from the cleaning process will vaporize at less than 800 F under the vacuum conditions in a furnace.

In one example, the assembly described above (i.e., FIG. 7 with upper plate and tightened-down nuts) is bonded into building block cores using the VPEI Load Assisted Brazing (LAB) process. LAB is a hybrid joining process that involves both solid-state and liquid-state bonding. The LAB process is usually employed in cases where the integrity of a diffusion bond is desired at a low cost. The process also allows for smaller loads to be applied, compared to pure diffusion bond designs, resulting in reduced deformation of less robust parts. Alternatively, a LAB process is used to keep intricate braze assemblies aligned and with zero gap size so that minimal braze filler metal is necessary. Reducing the amount of braze alloy in the joint can be beneficial, by reducing base metal erosion, unwanted alloy flow and fill, overall assembly weight, and cost. The practice of Load Assisted Brazing includes a force (or load) applied to the assembly (i.e., FIG. 7 with upper plate and tightened-down nuts) bring the base materials into intimate contact at a high temperature. When intimate contact is achieved at the base material interfaces, under the right conditions, mechanical and metallurgical bonding will proceed. The properties of the bonded interface can be brought to near base metal levels by dwelling at an elevated temperature and pressure appropriate to the materials and geometries of the part. The time and temperature dwell allows for a diffusion-controlled process to recrystallize and homogenize interfacial grain structure, while also minimizing or eliminating entrapped voids. LAB processing may be carried out in a controlled atmosphere or in vacuum. A vacuum atmosphere can provide a very clean processing environment and presents no additional cost. The choice of vacuum is standard in bonding/brazing of stainless steel. Large and small VPEI furnaces have cold wall vacuum chambers with refractory metal heating elements and all-metal hot zones, providing low out-gassing rates and relatively fast cycle times. A LAB process requires an applied force and this may be performed using either a dynamic hydraulic load or a static mass load. To achieve the high vacuum levels necessary, a cryogenic pump is used to capture the chamber gasses by freezing them onto very low temperature arrays. This pumping process effectively lowers the chamber pressure to a billionth of an atmosphere, or on the order of 10E-6 torr. Because of out gassing from the furnace internals upon heating, the practical service vacuum pressures are expected to be in the range to 10E-5 to 10E-4 torr. For stainless base material and nickel braze material, this pressure range will give satisfactory results at processing temperature of ≧950° C. The test bonding furnace, as well as other VPEI vacuum furnaces, has automated controls via PID-based digital control programmers. These programmers allow for time and temperature segments to be programmed for setting ramps and dwells. For temperature feedback, the furnaces use thermocouples.

An example of a thermal profile for bonding constructs of a core precursor to a form the core channels is summarized below:

1. Apply a holding force to the part stack

2. Ramp up at 10° C./min (50° F./min)

3. Above 350° C. (662° F.), hold to allow out gassing if the pressure spikes to greater than 3E-4

4. Hold at 850° C. (1562° F.) for 15 minutes to help homogenize the stack temperature

5. Soak at 980±10° C. (1796±50° F.) for 2 hours. Begin soak time when all part TC's are in range.

6. Apply a load at the beginning of the soak

7. Ramp down at 10° C./min (50° F./min)

8. At ≦100° C. (212° F.), the chamber may be vented to atmosphere for further cooling.

Example

Once bonding is complete, parts are unloaded and the channels of the core are formed. In one specific example, a building block core was manufactured as follows. A commercially available lamination machine was used for the roller press 50 in FIG. 5. The station 50 was set up to produce 4.0″×4.0″ sized constructs made from 0.025″ diameter 304SS round wire secured to 0.008″ thick 304SS sheet using Spunfab VI6010 adhesive. Along the sheet 40 fed from the roll 34 the round wire 30 wires are equally spaced at 0.150″ edge-to-edge distance from each other, except at the two edges of a laminate where flat rectangular ribbons of 0.025″×0.200″ cross-section are used instead of wires. The wider ribbons at the edges of a construct ensure complete sealing of a cross-flow core during bonding. The 4.0″×4.0″ construct has channels formed by twenty wires and two flat edge ribbons. The constructs were placed into a frame much like that shown in FIG. 7 and then bonded using LAB according to the above steps 1-8. The core structure bonded using LAB weighed 2.46 lb (1120 gm) with a 4.0″×4.0″ footprint, and 1.11″ high including top and bottom reinforcing plates each 0.057″ thick. The core is about 48% solid and 52% open. The assembly is well brazed with clear visual evidence of alloy flow and filleting. There is no indication of contamination-related effects from residual glue or carbonaceous material.

After the core is formed, side edges of the core were finished by EDM. This produces a flush, flat face on the core so that manifolds can be coupled to the channels with minimal flow or efficiency loss between the manifold conduit and channel entrances, especially for a core operated at high operating pressures.

The 4×4 core example given above used round wire 30 and flat edge ribbons of 0.2-inch width for extra edge sealing capability. Round wire is easier to handle and align on sheets than flat wires. However, flat wires can provide more direct contact with the sheet and thereby can yield much higher strength when bonded than round wires. Contact areas could be as much as five times higher and core burst strength would be similarly increased. Flat wires will add more supported area when the cores are bonded under loading in the furnace. While small misalignment in round wires can cause stack distortion, the flat wires provide sufficient support under load to minimize distortion. While flat wire will increase material usage, and require more face area to achieve an equivalent flow area, cost and volume increments are modest. To prove the use of flat wires, constructs 22 were also were fabricated using flat wire of 0.07-inches width, with similar edge ribbons and flat sheets. The same adhesive type and quantity was used as in the example using round wires. And a similar heating and load schedule was applied within the laminating machine.

Accordingly, the invention contemplates wires that are either flat or rounded as either approach provides advantages and can be successfully used as a heat exchanger in a wide variety of applications. Round wire assemblies can be made more compact than when flat wires are used, because there tends to be less channel blockage per face area. On the other hand, channels formed using flat wires can be stronger because there is greater contact area between plates. Based on these results, other types of wire (e.g. rectangular, trapezoidal, and oval) are also contemplated.

While stainless steel sheet and wires were used in the examples presented, aluminum, copper, oxidation resistant alloys, etc. can also be used. A copper based example is given below. Furthermore, mixed materials can be used, with plastic sheets and metal wires bonded with adhesives or by heat treating. Besides basic channel and wire dimension variations, channel heights and wire spacing can be varied from gas to liquid side, or with gas/gas or liquid/liquid. Wire spacing can also be varied across the sheet width as well. The size distribution of channels across the core can be varied. This will be particularly useful in single fluid heat sink applications, as described below.

Wire Mesh Example

In an alternative embodiment of forming constructs 22, a pre-sewn wire meshes, rather than individual wire threads, are used. The wire meshes are secured to sheets to form constructs in a similar manner as described in connection with FIG. 5.

The mesh wire is supplied to the rolling press. Cross-weave wire perpendicular to the long run that will form channels (the cross weave are slotted wires having the long run of wire passing through the slots). The cross weave maintains the desired spacing between the long run of wire that will form with the sheet material the channels (alternatives to slotted cross weave wire are within the scope of the disclosure. An alternative to slotted wire need only be capable of maintaining the spaced relationship as the long run of wire is dispensed and aligned with a strip of sheet material during the construct assembly process).

FIGS. 8A and 8C illustrate examples of a mesh 130 and 130′. Shown are two long wire portions 130a and 130b of the mesh 130, and long wire portions 130a′, 130b′ of mesh 130′ each having the same number of wire and interconnected to each other through a cross-weave wire 134 and 134′ (there are actually several of these wire portions in a mesh and the mesh 130 is carried on a spool for dispensing the meshes 130, 130′ during assembly using the station 140 illustrated in FIG. 8B). Each wire in the upper and lower long wire portions 130a, 130b and 130a′, 130b′ are spaced by the desired construct channel width W, e.g., 6 diameters of round wire, and are maintained in this spaced relationship by the cross weave 134. The spacing between the cross-weave 134 is greater than the length of a construct (for alignment purposes, below) and the width of each of the long wire portion 130a and 130b between the cross 134 is about the width of a construct. The meshes 130, 130′ in FIGS. 8A, 8C dispended from rolls from right to left. In the case of cross-weave 134′, tension in the long wires forming 130a′, 130b′ may be maintained by a crimping and wedge-shaped roller 143 that pulls the wires 30′ in tension. The portion of the mesh 130 illustrated in FIG. 8A shows wire for assembling four constructs 131a, 131b, 131c and 131d.

In general, wire weave material illustrated in FIGS. 8A and 8C is typically employed as a filter or screening device. These screen meshes are woven on equipment that is similar to what is used in the low cost manufacture of textiles. Given the high volume and low cost production of fine wire mesh weaves, and relatively low cost machines used to make these weaves, the cost of wire mesh like that illustrated in FIGS. 8A and 8C is low. This will help control construct fabrication cost.

The following describes an alternative assembly process to that described in connection with FIG. 5. Referring to FIG. 8B, there is illustrated a side-view of an assembly station 140 used to secure spooled wire mesh 130 to spooled sheet strip 135. The station 140 is preferably used to secure metal wire to a metal sheet strip. A similar assembly process and material supply can be employed when non-metal material is used. A four-layer stack of sheet strips and wire mesh is assembled using this station. This four layer stack, which will hereinafter be referred to as four layers of “wire-sheet strips” is cut to a size suitable for being placed in a furnace. The furnace is used to form channels from the wire disposed between the sheet strips. After forming channels, the blocks of constructs are formed, as explained in greater detail below.

There are eight spools illustrated in FIG. 8B. Four of these spools dispense the wire mesh 130, 1

30′ and four spools dispense the sheet strip 135. Spools 146a, 146b, 145a and 145b dispense the sheet strip 135. Each of these spools dispenses several separate sheet strips, each having a width of a construct 22 and being aligned with, or at a right angle to a corresponding one of the long wire portions in a mesh 130. A short section of the wire mesh 130 having the cross-weave 134 and sheet strip extends beyond the length of the constructs that will be later cut from the wire-sheet strips. The cross-weave 134 can be used to align the sheet strips with the mesh.

Spool 144b and 147a dispense the wire mesh 130. And spools 147b and 144a dispense the wire mesh 130′, which is the same mesh 130 but orientated at 90 degrees to wire mesh 130. Spools 144b and 145b form a first layer of wire-sheet strips. Spools 144a and 145a form a second layer of wire-sheet strips, placed below the first layer and with wire oriented at 90 degrees to the wire in the first layer. Spools 147a and 146a form a third layer of wire-sheet strips, placed below the second layer and with wire oriented at 90 degrees to the wire in the second layer (and parallel to the wire direction in the first layer). And spools 147b and 146b form a fourth layer of wire-sheet strips, placed below the third layer and with wire oriented at 90 degrees to the wire in the third layer (and parallel to the wire direction in the second layer). As will be appreciated, eight layers of wire-sheet strips can be formed using a similarly constructed station using 16 instead of 8 spools. The 4, 8 or more layers of constructs according to the above process are additional examples of constructs forming core precursors.

The wire mesh 130, 130′ is temporality secured to the sheet strips using an adhesive 149 sprayed to the sheets upstream of the press rollers 143 at the location corresponding to where the cross weave 134 contacts the sheet strip 135. As illustrated in FIG. 8B, each layer of wire-sheet strips are feed simultaneously between the press rollers of the moving press to adhere the wire to the sheet strips via the adhesive applied at the locations of the cross-weave 134. The wire-sheet strips 160 are cut at the locations of the cross-weave 134 to remove the adhesive and separate constructs. After forming the constructs in this manner the assembly process described earlier for the core may be used.

HEX Examples

“Building block” cores constructed according to the described processes can be stacked to form a stacked core 171, as illustrated in FIG. 9A, to achieve the needed capacity for the heat exchanger. To add strength to the blocks, bonding plates 175 can be added to the top and bottom of the core 170 or stacked cores 171, and right angle channels 176 are added to the corners to stiffen the overall structure and provide a base for brazing on manifolds. This is illustrated in FIG. 10. Manifolds 172 can then be attached to uniformly feed the core 171, as illustrated in FIG. 9B (four manifolds 172 are shown. The arrows indicate the flow direction through the heat exchanger). This manifolding of the core 171 can be based on accepted practice for conventional-type heat exchangers.

As noted above, low precision block edge cutting is used to control fabrication costs. In this case, some of the channels may be misaligned with the surface, resulting in some channels connecting to the side of the block. To avoid any impact of these misaligned channels, the manifolds 172 are constructed to be smaller than the face of the block, and brazing compound and/or stiffening corner channels 176 are used outside of the manifold 172, as illustrated in FIG. 10, to close off any misaligned channels. By brazing over these edge channels, about 10% of the total block flow area is lost. However, this is a small compromise for allowing the use of low precision and low cost cutting techniques. Also, with this brazing approach, the overall strength of the manifold and core structure is increased, which is particularly important for high pressure applications.

As the foregoing demonstrates, fabrication steps according to the disclosure can be carried out using low precision equipment. By first constructing building block cores from constructs 22 made from inexpensive wire and sheet material, the approach has the flexibility to address several different heat exchanger capacities and fluids. The configuration in FIG. 9B is for a cross flow type heat exchanger. To implement a counter flow, or co-flow, type heat exchanger, four or more cores 171 may be aligned as shown in FIG. 9C, with the cores 171 connected by “half pipe” type manifolds 173. While not a strict counter flow heat exchanger, by including several passes, a counter flow type condition can be approached. Therefore, besides the building block approach providing capacity and fluids flexibility, cores can also be used to create different classes (e.g. cross flow, counter flow, co-flow) of heat exchangers.

Performance

Microchannel heat exchangers constructed in accordance with the above methods have low channel Reynolds numbers and thereby operate with laminar flow, rather than turbulent flow. Under laminar flow conditions, heat transfer and pressure drop theory are well established. To confirm this, a core constructed in accordance with the disclosure was tested using hot and cold air flows. The 4-inch core described above was tested and compared to laminar flow theory. It was found that the tested core came with 10% of the theoretical result, given the channel size and dimensions, fluid properties, etc.

Cores constructed according to the disclosure have very high heat transfer coefficients that correspond with the expected laminar flow for a heat exchanger having microchannels. In these cases, the heat transfer coefficient and friction factor can be readily calculated using theoretical relationships.

The heat transfer coefficient varies inversely with the channel hydraulic diameter, and it rapidly increases as the hydraulic diameter is reduced, as given in expression 1, where h is the heat transfer coefficient, K is conductivity of the fluid, Nu is the Nusselt number and d is the channel hydraulic diameter.

$\begin{matrix} h = \frac{kNu}{d} & (1) \end{matrix}$

Likewise, the friction coefficient, as given in FIG. 10, also rapidly increases with reductions in channel hydraulic diameter, as given in expression 2, where C is a constant for the channel configuration and Re is the Reynolds number, based on channel hydraulic diameter, with V the velocity and u the kinematic viscosity.

$\begin{matrix} f = \frac{C}{R_{e}} = \frac{C υ}{Vd} & (2) \end{matrix}$

Therefore, as channel diameter is reduced, the heat transfer coefficient will increase, and the pressure drop will correspondingly increase. However, using expressions 1 and 2, it can be shown that the Coefficient of Performance (COP), or heat transfer per flow power (i.e. flow rate times pressure drop) remains constant for different channel hydraulic diameters. Therefore, from a COP perspective, there is no difference between a core manufactured according to the disclosure with different channel dimensions. However, by reducing channel size, the volume of the heat exchanger (HEX) can be substantially reduced.

To evaluate the advantages of a core manufactured according to the disclosure, the well proven Number of Transfer Units (NTU) approach of Kays and London [1] is utilized. In this approach, the HEX effectiveness is defined in expression 3:

$\begin{matrix} \in = \frac{q}{q_{ma x}} = \frac{c_{h} (T_{hi} - T_{ho})}{c_{m i n} (T_{hi} - T_{ci})} & (3) \end{matrix}$

Where q is the actual heat transferred, q_maxis the theoretical maximum, is the thermal capacitance (cp mdot) of the hot stream, c_minis the minimum thermal capacitance (either hot or cold stream), Th and Tc hot and cold temperatures, and i and o denote inlet and outlet, respectively.

The effectiveness ε compares the actual heat transfer (q) to the thermodynamically limited maximum possible heat transfer rate (qmax) as would be realized only in a counter flow heat exchanger of infinite heat transfer surface area. Therefore, one minus effectiveness is equal to the loss of heat transfer versus that theoretically possible. Data is available that provide cross flow HEX effectiveness as a function of Number of Thermal Transfer Units (NTU) and number of passes. As shown, for all cases the effectiveness increases as NTU increases.

Also, this case illustrates a core manufactured according to the disclosure where the cores are manifolded so that the cold fluid takes a serpentine path through aligned cores, with the number of passes equal to the number of cores arranged in sequence. For one pass, the HEX is a cross flow case, which has the lowest effectiveness. However, with multiple passes the HEX approaches a counter flow case (number of passes equal infinity). For this last case, the effectiveness is maximized. For any number of passes, the configurations effectiveness increases with NTU, which is defined in expression 4, where

$\begin{matrix} NTU = \frac{UA}{C_{m i n}} & (4) \end{matrix}$

U is the combined heat transfer coefficient, A is the heat transfer surface area and Cmin is the specific heat times the mass flow. The side of the HEX that has the minimum thermal capacitance is used in the formation of this NTU parameter. Since effectiveness defines the approach to the ideal heat transfer, as given by expression 3, the HEX volume analysis should use both a fixed effectiveness and heat transfer for a consistent comparison. This then requires fixing NTU to provide a direct comparison of the impact of channel dimension on HEX volume.

Ignoring the thermal resistance across the separating plates, the U parameter is equal to the heat transfer coefficient divided by two, to approximately account for the cold and hot side heat transfer coefficients. Of course, this assumes that both cold and hot sides have the same channel sizes, velocities and properties. This is a simplification, but illustrates the impact. For the same face velocity, the HEX face area is the same, and we assume the face width, W, and height, H, are equal between HEXs with different channel dimensions. By reducing the channel height, the number of channels, n, per fixed HEX height is increased. With increased number of channels, the heat transfer area, A, is then increased, unless the length, L, of the heat exchanger is reduced. Considering the channel hydraulic diameter reduction from 0.1-inch in the conventional case to 0.025-inch for core manufactured according to the disclosure, or a factor of four, the volume reduction can be calculated easily. Specifically, for equal NTU, expression 5 gives the relationship between h, A and Cmin, where 1 denotes the conventional case and 2 denotes the case for core manufactured according to the disclosure.

$\begin{matrix} \frac{h 1 n 1 W 1 L 1}{C \min 1} = \frac{h 2 n 2 W 2 L 2}{C \min 2} & (5) \end{matrix}$

where n1 and n2 are the number of channels for the 0.1-inch and 0.025-inch channel dimension cases, respectively. As noted above, W1=W2, Cmin1=Cmin2, and the length ratio of core manufactured according to the disclosure compared to a conventional HEX is then given by expression 4.

$\begin{matrix} \frac{L 2}{L 1} = \frac{n 1}{n 2} \frac{h 1}{h 2} & (6) \end{matrix}$

For the fixed HEX height of H, the number of plates are then given in expression 7

$\begin{matrix} n 1 = \frac{H}{d 1} n 2 = \frac{H}{d 2} & (7) \end{matrix}$

And (6) then becomes (8)

$\begin{matrix} \frac{L 2}{L 1} = \frac{d 2}{d 1} \frac{h 1}{h 2} & (8) \end{matrix}$

For the factor of 4 change in hydraulic diameter, the L2/L1, which ratios the hydraulic diameter for a core manufactured according to a conventional HEX length ratio, is then given in expression 9:

$\begin{matrix} \frac{L 2}{L 1} = \frac{1}{16} & (9) \end{matrix}$

This result shows that the length can be reduced by a factor of 16 relative to a conventional HEX, by reducing the channel size by a factor of four. Therefore, for a fixed effectiveness, the small channel cores according to the disclosure would reduce volume by a very substantial 84%. This ignores plate thickness impacts and assumes fixed velocity and face area, but the strong influence of channel diameter on HEX volume for same heat transfer is clear. Pressure drop for these cases is given by expression 10, where

$\begin{matrix} Δ p = f^{\frac{L}{d}} ρ^{\frac{V 2}{2}} & (10) \end{matrix}$

f is friction coefficient, V is the velocity, and ρ is the density, with L and d the heat exchanger length and channel hydraulic diameter, as defined previously. As noted in the case above, velocity, V, and density, ρ, are the same between cases, and taking expression 2 for f, the pressure drop ratio is then given by expression 11.

$\begin{matrix} \frac{Δ p_{2}}{Δ p_{1}} = \frac{L_{2}}{L_{1}} {(\frac{d_{1}}{d_{2}})}^{2} & (11) \end{matrix}$

As given by expression 9, L2/L1 is 1/16 and with d1/d2 equal 4, the pressure drop ratio is then given by expression 12.

$\begin{matrix} \frac{Δ p_{1}}{Δ p_{2}} = 1 & (12) \end{matrix}$

Therefore, while a core manufactured according to the disclosure is 84% lower in volume to achieve the same heat transfer as a conventional HEX, it has the same pressure drop. This is a very attractive outcome.

In shrinking the core size through the use of smaller channels, the heat transfer surface loss per heat transferred is reduced, as a result of the lower exposed relative surface area. The heat transfer produced by the core manufactured according to the disclosure, for a given temperature difference driver, is equal to the NTU parameter multiplied by the thermal capacitance and temperature difference. If both the thermal capacitance and temperature difference are both equal between this core and conventional heat exchanger, then the heat transfer for the equal NTU cases noted above will be the same. However, the HEX external surface area will be different. As given above, the face area of the small and larger channel dimension cases are the same, to achieve equal flow velocities. However, the length is reduced by a factor of 16 for the small diameter channel case. The surface area through which heat can be lost for the same internal heat transfer is then 84% lower for the small channel core case. Therefore, besides greatly reducing HEX volume, the small channel size for the core according to the disclosure also greatly reduces heat loss potential versus heat transferred.

By decreasing the differences between the hot and cold streams, or increasing HEX effectiveness, process losses can be decreased. For conventional heat exchangers, where h is lower than desired, expression 4 shows that surface area, A, has to be increased to increase NTU and thereby effectiveness, if thermal capacitance is fixed. While this is certainly possible, increasing A requires more volume. However, by going to smaller channel size, h rapidly increases and effectiveness is improved. Furthermore, as noted above, the number of separation plates and A are increased for a fixed heat exchanger height, H. This further increases effectiveness through increasing NTU, as given in expression 4.

To illustrate the impact of channel dimension on effectiveness, a fixed volume heat exchanger is considered for both conventional and for a core manufactured according to the disclosure. As channel dimension is reduced, h is increased, and number of separation plates is also correspondingly increased. For the same width, W, and length, L, plates that make up the heat exchanger, A is then correspondingly increased. For the same factor-of-four reduction in channel dimension, as used above, the NTU then increases. The NTU ratio is then given by expression 13, for fixed thermal capacitance.

$\begin{matrix} \frac{NTU 2}{NTU 1} = \frac{A 2}{A 1} \frac{h 2}{h 1} & (13) \end{matrix}$

The area ratio is equal to the separation plate number ratio, as given in expression 14.

$\begin{matrix} \frac{A 2}{A 1} = \frac{d 2}{d 1} & (14) \end{matrix}$

And the h ratio is given by expression 15:

$\begin{matrix} \frac{h 2}{h 1} = \frac{d 1}{d 2} & (15) \end{matrix}$

For the ratio of four smaller channel dimensions with WASHEX, the NTU ratio is then given by expression 16.

$\begin{matrix} \frac{NTU 2}{NTU 1} = 16 & (16) \end{matrix}$

Therefore, a core manufactured according to the disclosure has a much higher NTU. Using the simple counter flow HEX case in FIG. 40, the approach to the theoretical heat transfer limit can be calculated for conventional HEX and a core manufactured according to the disclosure for a factor of four smaller channel size. Results indicate that the conventional approach produces heat transfer results far from the theoretical maximum for a reasonable range of NTU1 conventional, whereas a core manufactured according to the disclosure comes much closer to the theoretical. This is because the equal volume core of the disclosure has a much higher NTU2 (16×NTU1 according to expression 16) than the conventional HEX.

To a first approximation, the ratio of conventional and core of the disclosure weight is equal to the plate area ratio, if both use the same plate thickness and material of construction. As given by expression 14, the area ratio, or weight ratio, is equal to the channel hydraulic diameter ratio. Therefore, for a factor of four lower channel diameter, the area and weight ratio is only 25%. This then results in a weight reduction of 75% for a core manufactured according to the disclosure. Furthermore, to a first approximation, if a core manufactured according to the disclosure and the conventional HEXs are constructed of the same material, then weight reductions of 75% will result in material cost reductions of 75%.

HEX Types

From the above, it will be appreciated that a core manufactured according to the disclosure has significant potential for various applications. Example applications are given below that further highlight specific benefits.

Through proper manifolding various heat transfer capabilities can be achieved for both liquid and gas or mixed liquid/gas applications. A single core may be manifolded to produce a cross flow HEX. For different capacities, more cores are simply stacked and manifolded. With this approach, HEX face velocity is kept relatively constant, along with pressure drop as capacity needs increase. For example, four cores of cross flow HEXs that are aligned in a row that are manifolded so that a serpentine multi-pass flow yields a counter flow overall arrangement. Only three passes are required to achieve a good representation of counter flow.

A multi-pass counter-flow heat exchanger can use four separate cores, or a single rectangular core can be produced that replaces the four. Specifically, the station 50 can be modified to produce constructs up to 50-inches wide. Also, different width flat wires can used. To create 4-inch wide multi-pass heat exchangers, 4-inch wide constructs can be cut at about 16-inch length to form the long core channels. These can then be combined with 16-inch wide constructs cut at approximately 4-inches. Together, these create a 16-inch long by 4-inch wide single core which substitutes for the four separate cores. Furthermore, to facilitate manifold attachment and sealing, wide ribbons can be inserted interior to the sheet at three locations corresponding to manifold attachment points along the length of the heat exchanger. By following this general approach, various size cores can be fabricated to address different applications of interest. This is an additional aspect of the disclosure.

Electronic cooling is one application of the cores of the disclosure. Besides demonstrating good performance for this application, examples have demonstrated that these type devices could be constructed from copper.

An application of evolving and increasing importance for HEXs is the direct cooling of high power density electronics, using Thermal Ground Planes (TGP) combined with mini channel or microchannel “cold plate” HEXs to remove the heat generated by the electronics. As electronics become more powerful and space is reduced to increase processing rates, or more components are packaged tightly in space limited applications (i.e. avionics), only MM HEXs with single or two phase liquid flows can meet heat dissipation needs. Many studies of boiling (evaporative) cooling in mini channel and microchannel HEXs have been discussed. These studies show the potential to reach >100 W/cm2, which is very high compared to conventional HEXs. Coolants included water and refrigerants (e.g. R134a). This work clearly shows the potential of mini and microchannel HEXs, as well as the beneficial use of refrigerants in these devices. Given that chillers utilize both water coolants and refrigerants, this work supports the use of WASHEX in chiller applications. A core manufactured according to the disclosure can be used to address these challenging applications, as an effective and low cost solution that can also yield more uniform temperatures in the electronics, which improves component life.

For this application, cold plate cores are made out of thin copper sheets (0.01″) and flat wires (0.04″×0.01″) in the similar process as described above. The constructs are made from braze alloy coated thin copper sheets and bare copper flat wires using the web adhesive. After a long construct is formed, it is cut to desired lengths by a water jet cutter. The cut constructs are 1.24″×0.94″. The cold plate core has a 1.24″×0.94″ footprint comprised of multiple layers of cross flow channels. Therefore, two types of constructs are required to fabricate a core, WO. 94″×L1.24″ and W1.24″×L0.94″. Initially the core was designed with only 4 layers of constructs, 2 layers of channels in longitudinal direction (long) and 2 layers of channels in lateral (short) direction. The core is assembled in a fixture and bonded by LAB.

To create the cold plate, the fabricated core is bonded inside a manifold block. The core block is placed in the manifold to allow coolant fluid to simultaneously enter two surfaces of the core and then exit into a collection chamber on the other sides of the core. The flow inlet and outlet fittings (Swagelok straight male weld) are also brazed during the same braze run.

A fabricated core serving as a based cold plate was tested and the results compared to existing microchannel electronics cooling systems at the same coolant flow rate. From these tests, it was shown that the core according to the disclosure had a thermal resistance of 0.055 C/w while the state-of-the-art (SOA) device had a 0.09 C/w at the same conditions. Also, the core weight and volume were 62% and 60% less than the SOA device.

While the local “cold plate” approach is viable for electronics cooling, a laser module, for example, includes multiple large holes and the placement of many laser diodes in locations on the plate dictated by laser coupling needs. In this case, a simple rectangular “cold plate” would not be optimal. Given this situation, a more flexible and integrated approach was considered.

Microchannel cores of small dimensions are fabricated, and these cores then aligned in a simple machined channel in the laser module that then acts as a manifold for the coolant flow. In this approach, called the Distributed Core Module Heat Exchanger (DCMHEX), the small cores can be sized to closely couple with the heat dissipation needs of components (e.g. laser diodes), with the module channels allowing the cores to be aligned as needed with any final electronics component arrangement. This approach has significant flexibility versus a simple rectangular cold plate approach, where mass has to be added to spread the concentrated heat loads to larger heat transfer modules within the plate. By a more concentrated and optimal alignment between cores and the heat load, system weight and volume can be minimized.

A core manufactured according to the disclosure offers high performance as a heat exchanger through the very high heat transfer coefficients produced by microchannels. By the proven analogy between heat and mass transfer, mass transfer will also be very high. If alternative channels in cores are wash-coated and/or coated with catalysts, then reactants in those channels could be rapidly converted to desired products, while the heat released (i.e. exothermic) or absorbed (i.e. endothermic) is transferred from gases or liquids in the alternative channels. This approach addresses a fuels reforming application of microchannel heat exchangers, where the reforming reactions require heat input from hot gases to maintain the proper reaction temperature. To optimize catalyst performance and reliability, the sheet materials can be selected to be more compatible with the wash-coat or catalyst. Within a range, wire and sheet materials can be different to optimize strength and catalytic reactivity when properly coated. Gas/gas, liquid/liquid and gas/liquid approaches can be considered.

Applications also include single-phase fluid cases. These also extend to multi-phase flow. For example, boiling heat transfer is very high for liquids, such as water or refrigerant type liquids. In these cases, small vapor bubbles form at the HEX surface that enhance agitation of the remaining liquid and promote more rapid heat transfer at the surface. This enhanced heat transfer is very useful for very high heat flux cases, such as special high powered electronics applications. However, once all of the liquid in a channel has been converted to vapor, the heat transfer abruptly decreases in this “dry out” zone. Surface temperature then jumps in this zone, which is detrimental to electronic component longevity. Cooling designs then need to be very conservative. However, with the core approach described herein, as illustrated by a cold plate example just described, there are multiple channels with coolant cross-flow that can yield more uniform heat transfer. Specifically, the multiple separation plates help to conduct heat from the fluid and spread that heat to the entire structure. Also, by cross flow, cold fluid enters on two faces of the core, rather than a single face as in typical heat exchangers. This also helps to even out heat transfer and temperatures. Therefore, multiple mechanisms can be implemented to control temperature uniformity and improve electronic component life. By proper design of channel dimensions and footprint, and number of channels, an optimal solution can be found for the heat management problem of interest. This same approach for a single fluid cold plate type application has benefits for a two fluid type case, where multi-phase flow is present. Both refrigeration evaporators and condensers can benefit from using a core manufactured according to the disclosure.

As noted earlier, applications include liquid and gas fluids flowing in each set of channels, as well as gas/gas and liquid/liquid. Given the much lower density of typical gases, the gas side channels have to be much larger than the liquid side channels to maintain a reasonable pressure drop at similar mass flow rates. The gas/liquid heat exchanger covers important applications, like radiators, condensers and evaporators that typically use liquid-in-tube and gas-in-fins style heat exchangers. To minimize volume and weight, these heat exchangers utilize high performance louvered type fins and extruded microchannel type flattened tubes, as illustrated in FIG. 3. These units are fabricated by bonding the fins, extruded tubes and manifolds together using standard furnaces. A core manufactured according to the disclosure, or stacked cores could be employed in a similar way to produce the microchannel tubes for these heat exchangers.

In this example, sheets, flat wires and edge ribbons are bonded together in long strips to create essentially an extruded coolant tube. Compared to conventional extruded tubes, these tubes will be lower in weight and volume and higher in performance than extruded channels. These manufactured tubes can then be inserted between corrugated fins to create a radiator type heat exchanger.

Phase Change Materials (PCM), such as paraffinic waxes have a high heat of fusion that can be used to manage transient heat loads produced, for example, by pulsed electronics applications. As the unit is pulsed, a very high heat spike will propagate through the cooling system, leading to the over-temperature of the electronic components, unless the thermal management system is sized for the heat spike. However, by sizing the system for the peak, the weight, volume and cost for the system will be excessive versus a system sized for the average heat load. By including a PCM material in the loop, the PCM can absorb substantial energy as it converts from a solid to a liquid at nearly a fixed temperature. This will shave the peak and allow an overall lower volume, weight and cost thermal management system.

While PCMs are very beneficial, those that are effective in the temperature range of interest are relatively poor conductors. In this case, a heat spike may not be absorbed in the time scale needed to prevent the over temperature of a component, due to the bottlenecking of heat transfer through the low conductivity PCM. To eliminate this bottleneck, a core manufactured according to the disclosure can be used, where one set of channels is filled with PCM and the other set contains the coolant flow. For this case the heat conduction path in the PCM is shortened to that on the order of the channel height. This greatly facilitates the thermal response of the PCM mass. In addition, this core structure helps to distribute the heat, through conduction in the metal structure. Both heat conductivity and heat capacity are balanced in PCM based thermal management systems. Lastly, while the beneficial case of a solid PCM is considered, a core manufactured according to the disclosure can also be used to optimize the impact of slurry type PCMs, where fluid heat capacity is enhanced by the addition of micro-encapsulated PCMs.

The above description of illustrated embodiments of the invention, including what is described in the Abstract, is not intended to be exhaustive or to limit the invention to the precise forms disclosed. While specific embodiments of, and examples for, the invention are described herein for illustrative purposes, various modifications are possible within the scope of the invention, as those skilled in the relevant art will recognize.

These modifications can be made to the invention in light of the above detailed description. The terms used in the claims should not be construed to limit the invention to the specific embodiments disclosed in the specification. Rather, the scope of the invention is to be determined entirely by the claims, which are to be construed in accordance with established doctrines of claim interpretation.

Microchannel Heat Exchangers and Reactors

Information

Publication Number

Date Filed

Date Published

Inventors

Original Assignees

CPC

US Classifications

International Classifications

Abstract

Description

Claims

PRIORITY CLAIM

Government Interests

Provisional Applications (1)