Laminates containing tempered glass

Abstract

In some embodiments, glass layers may be tempered in a furnace that uses air to support and / or convey the glass through it. As a result, there may be even pressure on a glass layer's major surfaces and reduced distortion will result. In some embodiments, a method comprises: receiving parameters associated with a trained machine learning model configured to receive inputs indicative of parameters associated with lite tempering and output one or more recipe and / or bedload changes for a subsequence batch of lites so to be tempered; obtaining parameters associated with a first batch of lites undergoing tempering; determining, by providing the obtained parameters to the trained machine learning model, one or more modifications to: a current recipe used to process the first batch of lites; and / or a bedload configuration for a second batch of lites, wherein the one or more modifications are used to process a second batch of lites.

Description

LAMINATES CONTAINING TEMPERED GLASS CROSS REFERENCE TO RELATED APPLICATIONS

[0001] A PCT Request Form is filed concurrently with this specification as part of the present application. Each application that the present application claims benefit of or priority to as identified in the concurrently filed PCT Request Form is incorporated by reference herein in its entirety and for all purposes. BACKGROUND

[0002] Laminated glass structures are used in many contexts. For example, laminated glass structures may be used in optically switchable windows such as electrochromic windows. Laminated glass structures include at least two glass sheets bonded at respective major surfaces of the sheets. One or more of the glass sheets may be tempered. An electrochromic device may be disposed on one or more major surfaces of the glass sheets. Unfortunately, glass sheets may have imperfections that produce optically undesirable effects, which may be magnified when the glass sheets are laminated. SUMMARY

[0003] Aspects of this disclosure pertain to laminated glass structures comprising: (a) a heat- treated first glass layer having no undulations with a peak-to-valley distance of about 0.2 mm or greater; and (b) a second glass layer adhesively bonded to the heat-treated glass layer. The laminated glass structure does not exhibit lensing caused by undulations. In some embodiments, the laminated glass structure further comprises an optically switchable device such as an electrochromic device.

[0004] In certain embodiments, the laminated glass structure further comprises an adhesive layer such as a resin (e.g., polyvinyl butyral (PVB), ionoplast (SentryGlas, SBP), or a combination thereof) between the heat-treated glass layer and the glass layer.

[0005] In certain embodiments, the laminated glass structure further comprises an electrochromic device disposed on a surface of the second glass layer opposite the heat-treated first glass layer. The second glass layer may be an annealed glass layer.

[0006] In certain embodiments, the laminated glass structure further comprises (i) an electrochromic device disposed on a surface of the second glass layer opposite the heat-treated first glass layer (the second glass layer may be an annealed glass layer); (ii) a heat-treated third glass layer; and (iii) a fourth glass layer adhesively bonded to the heat-treated third glass layer. In some cases, the electrochromic device is also disposed on a surface of the fourth glass layer.

[0007] In certain embodiments, the laminated glass structure further comprises an electrochromic device disposed on a surface of the second glass layer opposite the heat-treated first glass layer. The second glass layer may comprise heat-treated glass. In certain embodiments, the heat-treated first glass layer is a tempered glass layer.

[0008] Other aspects of this disclosure pertain to an integrated glass unit comprising any laminated glass structure as described above; a mate lite; and a sealed void region between the laminated glass structure and the mate lite.

[0009] Other aspects of the disclosure pertain to laminated glass structures comprising: (a) a heat-treated first glass layer having no bow or a bow of about 10 mm or less; and (b) a second glass layer adhesively bonded to the heat-treated glass layer. Such laminated glass structures may further comprise an optically switchable device and / or an adhesive layer between the heat-treated glass layer and the glass layer. The heat-treated first glass layer may be a tempered glass layer.

[0010] In certain embodiments, a laminated glass structure as described in these aspects further comprises an electrochromic device disposed on a surface of the second glass layer opposite the heat-treated first glass layer, where the second glass layer is an annealed glass layer.

[0011] In certain embodiments, a laminated glass structure as described in these aspects further comprises (i) an electrochromic device disposed on a surface of the second glass layer opposite the heat-treated first glass layer (where the second glass layer is an annealed glass layer); (ii) a heat-treated third glass layer; and (iii) a fourth glass layer adhesively bonded to the heat-treated third glass layer. The electrochromic device may be disposed on a surface of the fourth glass layer.

[0012] In certain embodiments, a laminated glass structure as described in these aspects further comprises an electrochromic device disposed on a surface of the second glass layer opposite the heat-treated first glass layer. The second glass layer may comprise heat-treated glass.

[0013] Other aspects of this disclosure pertain to laminated glass structures comprising: (a) a heat-treated first glass layer produced by heat-treating a pre-heat-treated glass layer in a furnace in which the pre-heat-treated glass layer is supported by a plurality of air jets; and (b) a second glass layer adhesively bonded to the heat-treated glass layer. In certain embodiments, these laminated glass structures further comprise an optically switchable device and / or an adhesive layer between the heat-treated glass layer and the glass layer. In various embodiments, the heat-treated first glass layer is produced by tempering the pre-heat-treated glass layer in the furnace.

[0014] A laminated glass structure of these aspects may further comprise an electrochromic device disposed on a surface of the second glass layer opposite the heat-treated first glass layer, where the second glass layer is an annealed glass layer.

[0015] A laminated glass structure of these aspects may further comprise: (i) an electrochromic device disposed on a surface of the second glass layer opposite the heat-treated first glass layer, wherein the second glass layer is an annealed glass layer; (ii) a heat-treated third glass layer; and (iii) a fourth glass layer adhesively bonded to the heat-treated third glass layer. In such structure, the electrochromic device may be disposed on a surface of the fourth glass layer.

[0016] A laminated glass structure of these aspects may further comprise: an electrochromic device disposed on a surface of the second glass layer opposite the heat-treated first glass layer. The second glass layer may comprise heat-treated glass.

[0017] Other aspects of the disclosure pertain to methods of fabricating a laminated glass structure, which methods may be characterized by the following operations: (a) heat-treating a pre- heat-treated glass layer by supporting the pre-heat-treated glass layer on a plurality of air jets while exposing the pre-heat-treated glass layer to heat-treating conditions to produce a heat-treated first glass layer; and (b) adhesively bonding the heat-treated first glass layer to s second glass layer.

[0018] In certain embodiments, during the heat-treating, the pre-heat-treated glass layer does not contact a solid surface. In certain embodiments, during the heat-treating, the pre-heat-treated glass layer does not contact a roller.

[0019] In certain embodiments, during the heat-treating, the pre-heat-treated glass layer floats substantially horizontally on air produced using the plurality of air jets. In certain embodiments, during the heat-treating, the pre-heat-treated glass layer floats at an inclined angle of not greater than about 10 degrees from the horizontal plane on air produced using the plurality of air jets. In certain embodiments, during the heat-treating, the pre-heat-treated glass layer floats substantially horizontal on air produced using the plurality of air jets. The air jets may produce air streams at an angle, or within a range of angles, that produces a force to push the pre-heat-treated glass layer along a pathway in a heat-treating furnace that applies the heat-treating conditions. The heat- treating furnace may include one or more guardrails that keep the pre-heat-treated glass layer within the pathway of the heat-treating furnace.

[0020] In certain embodiments, exposing the pre-heat-treated glass layer to heat-treating conditions comprises heating the pre-heat-treated glass with heat emitted by one or more heating plates while the pre-heat-treated glass is supported on the plurality of air jets. In some cases, the one or more heating plates include at least one heating plate disposed substantially under the pre- heat-treated glass.

[0021] In certain embodiments, the methods further comprise fabricating an electrochromic device on the heat-treated first glass layer and / or on the second glass layer. In certain embodiments, the methods further comprise forming an electrochromic device on a surface of the second glass layer opposite the heat-treated first glass layer, where the second glass layer is an annealed glass layer.

[0022] In certain embodiments, the methods further comprise: (a) forming a first portion of an electrochromic device on a surface of the second glass layer opposite the heat-treated first glass layer, wherein the second glass layer is an annealed glass layer; (b) forming a second portion of the electrochromic device on a surface of an annealed fourth glass layer opposite a heat-treated third glass layer; and (c) contacting the first portion of the electrochromic device to the second portion of the electrochromic device.

[0023] In certain embodiments, the methods further comprise forming an electrochromic device on a surface of the second glass layer opposite the heat-treated first glass layer, and wherein the second glass layer comprises heat-treated glass.

[0024] In certain embodiments, the heat-treating comprises quenching the pre-heat-treated glass layer by exposing different portions the pre-heat-treated glass layer to different cooling rates. In some embodiments, exposing the different portions the pre-heat-treated glass layer to different cooling rates is performed with one or more quenching air jets. In certain embodiments, the heat- treating comprises tempering the pre-heat-treated glass layer.

[0025] Other aspects of this disclosure pertain to methods of any of forming an integrated glass unit by assembling a laminated glass structure produced as described above, a mate lite, and a spacer in a manner that forms a sealed void region between the laminated glass structure and the mate lite.

[0026] Certain aspects of this disclosure pertain to apparatus comprising a furnace configured to accommodate a glass layer and comprising: (a) a heater; (b) an air float surface; (c) a plurality of apertures in the air float surface; and (d) a plenum or manifold configured to flow a gas flow through apertures to produce air jets that can impinge on a first surface of the glass layer and support the glass layer above the air float surface.

[0027] In some embodiments, the apparatus additionally includes a second plurality of apertures configured direct air onto a second surface of the glass layer, opposite the first layer. In certain embodiments, the apertures of the plurality of apertures are configured to direct the air jets at a non-normal angle with respect to the first surface of the glass layer.

[0028] In certain embodiments, the furnace is configured such that no edges of the glass sheet touch a surface of the furnace while the glass layer moves through the tempering furnace.

[0029] The apparatus may further comprise a quenching zone comprising a plurality of quenching apertures configured to provide quenching air jets that can support the glass sheet while in the quenching zone. In some cases, at least one of the plurality of quenching apertures comprises a nozzle configured to adjust the pressure of a quenching air jet emitted by the nozzle. These andother features of the disclosure will be presented in more detail below, with reference to the drawings.

[0030] In some embodiments, methods for determining lite tempering process parameters and / or bedload layouts are provided. In some embodiments, a method comprises: receiving parameters associated with a trained machine learning model configured to receive inputs indicative of parameters associated with lite tempering and output one or more recipe and / or bedload changes for a subsequence batch of lites so to be tempered; obtaining parameters associated with a first batch of lites undergoing tempering; and determining, by providing the obtained parameters to the trained machine learning model, one or more modifications to: a current recipe used to process the first batch of lites; and / or a bedload configuration for a second batch of lites, wherein the one or more modifications are used to process a second batch of lites.

[0031] In some embodiments, a method comprises: a) initializing a machine learning model configured to receive inputs indicative of parameters associated with lite tempering and output one or more recipe and / or bedload changes for a subsequent batch of lites so to be tempered; b) utilizing the machine learning model to generate an output comprising a recipe change and / or a bedload layout given an input of parameters associated with tempering of a batch of lites indicated in one or more training samples; c) updating weights of the machine learning model based at least in part on a difference between the generated output and a ground truth recipe changed and / or bedload layout indicated in the one or more training samples; and d) repeating operations (b) and (c) until a determination that the machine learning model has been trained is made.

[0032] In some embodiments, a method comprises: determining information associated with residual heat load of a set of rollers of a furnace used to perform a tempering process on a first batch of lites; and determining at least one of: 1) a modification to the tempering process, wherein the modified tempering process is to be used to process a second batch of lites; or 2) a bedload layout of the second batch of lites, wherein the modified tempering process and / or the bedload layout account for effects of the residual heat load of the set of rollers.BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 depicts an embodiment of a laminated glass structure that includes a tempered glass layer laminated to a second glass layer.

[0034] Figure 2A is an exploded view of components of a laminated glass structure.

[0035] Figure 2B presents an embodiment of a final laminated glass structure that may be fabricated from the components of Figure 2A.

[0036] Figures 3A and 3B illustrate a variation of the Figure 2A-B laminated glass embodiment.

[0037] Figure 4 is an example flow chart illustrating certain processes for fabricating a laminated glass structure.

[0038] Figure 5 illustrates a glass sheet in a heat-treatment process where the glass sheet is supported against gravity by air jets that provide pressure acting on a lower major surface of the glass sheet.

[0039] Figure 6 illustrates a glass sheet in a heat-treatment process, similar to the sheet in Figure 5, the pressure or air jets has a direction that is not substantially normal to the upper and lower major surfaces of the glass sheet.

[0040] Figure 7 illustrates an apparatus for implementing heat-treatments such as those of the embodiments of Figure 5.

[0041] Figures 8A-B illustrate an example of a tempering environment that allows a glass sheet to move back and forth within the environment while it is being heat-treated.

[0042] Figures 8C-D schematically illustrate an example quenching apparatus configured to effect quenching in a manner that applies differing degrees of quenching to different locations within the apparatus.

[0043] Figure 9 presents a schematic illustration of a cross-section of an electrochromic device in accordance with some embodiments.

[0044] Figure 10 depicts an insulated glass unit (IGU) with an electrochromic lite, which may comprise a laminated glass structure as disclosed herein.

[0045] Figure 11 is a flowchart of an example process for determining modifications to a current recipe and / or a bedload configuration using a trained machine learning model in accordance with some embodiments.

[0046] Figure 12 is a flowchart of an example process for training a machine learning model configured to determine modifications to a recipe and / or a bedload configuration in accordance with some embodiments.

[0047] Figure 13 is a flowchart of an example process for processing a second batch of lites using a modified layout and / or recipe determined based on a thermal load of a set of rollers previously used to process a first batch of lites in accordance with some embodiments.

[0048] Figure 14 is a diagram of an example system for determining recipe and / or layout modifications in accordance with some embodiments. DETAILED DESCRIPTION Introduction and Context

[0049] Existing processes for producing laminated glass structures often suffer from distortion, particularly when one or more glass layers in a laminate are tempered or otherwise heat-treated. The distortion may arise for the following reason. Tempered glass is typically tempered in a furnace that uses rollers to convey the glass. Since the glass is softened at these high temperatures, the glass gets “roller wave” during the tempering process. This results from the speed of the glass through the tempering furnace and the spacing of the rollers. There is some glass sag between rollers. Often the distortion is only slight—on the order of about 1mm or less—but the pattern is regular. When tempered glass is laminated to another piece of glass, even if the other glass is perfectly flat, these uneven surfaces, concavities, are filled with the lamination adhesive. This produces an optical effect that may be referred to as “lensing” where the light passing through is magnified and distorted.

[0050] In some embodiments, this issue is addressed by tempering glass layers in a furnace that uses air to support and / or convey the glass through it. As a result, there may be even pressure on the glass’ major surfaces and no roller wave or other distortion as a result.

[0051] In some embodiments, a laminated glass structure containing a tempered glass lite has little or no lensing or glass surface bow. Terminology

[0052] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art. The terms presented immediately below may be more fully understood by reference to the remainder of the specification. The following descriptions are presented to provide context and an introduction to the complex concepts described herein. These descriptions are not intended to limit the full scope of the disclosure.

[0053] Glass layer – A glass layer is a section or piece of glass that is generally flat and typically has two major surfaces that are significantly larger than all other surfaces. In various embodiments, the two major surfaces are parallel or substantially parallel. In embodiments disclosed herein, some glass layers have at least one generally planar surface that is susceptible to roller wave. A glass layer is sometimes referred to as a glass sheet. It may take the form of a glass lite, which may be used in a window such as an integrated glass unit (IGU). In various embodiments, a glass layer is provided as one layer of laminated glass structure.

[0054] Roller wave – Roller waves are repetitive, periodic waves observed in tempered glass. They can cause distortion of reflected images in the glass. Roller waves result from the tempering process as heated and softened glass moves along rollers. Roller wave may be characterized by a peak-to-valley distance and / or a wavelength.

[0055] Roller wave measurement may be made in accordance with ASTM C1651:

[0056] This test method is a procedure for determining the peak-to-valley depth and the wavelength of roll wave in flat glass and then calculating the optical distortion resulting from that roll wave. Peak-to-valley measurements provide a means of monitoring the roll wave distortion ina heat processed glass product. Roll wave is inherent in flat glass which has been heat-treated in a furnace in which rollers are used to convey the glass. Consult ASTM Specifications C1036 and C1048, which are incorporated herein by reference in their entireties, for additional glass characteristics and quality information.

[0057] Example gauges for measuring roller wave include the following ones available from Strainoptics, Inc. (North Wales, PA):

[0058] RWG-DF Digital Flat Roller Wave Gauges feature a flat, Delrin-coated bottom surface that is 16 inches long. The RWG-DF comes with a digital indicator for easy read-out of peak and valley measurements (W). The gauge is used with a measuring tape to determine the roller wavelength (L), or the distance between two consecutive peaks or troughs (see diagram on next page). The digital gauge can be set to either inches or metric readouts and has a sensitivity of 0.01mm (0.0005 inches).

[0059] The RWG-D Digital Roller-Wave Gauge features a non-wearing, three-point contact bottom surface and digital indicator readout with a metric resolution of 0.01 mm or an inch resolution of 0.0005 in. The contact point spacing is adjustable to the length of the peak-to-peak roller wave distance for more accurate, repeatable results. A high-sensitivity model (RWG-D-HS) is available for applications requiring high distortion sensitivity (1 mdpt).

[0060] Roller wave distortion – This is an optical effect caused by undulations, typically roller waves, on a window or other transparent structure. The optical effect may be manifest as lensing, which is a local variation in refraction of the transparent structure and gives a viewer the impression of magnification at periodic intervals on the structure surface. The lensing has a period corresponding to a period of the responsible roller wave. The lensing may be caused, at least in part, by a resin or other medium at the interface of a first transparent structure’s surface having roller waves and a second transparent structure bonded to the first transparent structure’s surface. In various embodiments, roller wave distortion occurs when the first transparent structure is a tempered glass sheet, and the second transparent structure is a second glass sheet laminated to thetempered glass sheet. However, the second transparent structure may or may not also have roller wave, which in the latter case can exacerbate the lensing.

[0061] Undulation (peak-to-valley distance for a given wavelength) – This is a term that embodies roller wave and other sorts for periodic variation on the face of a flat surface such as the face of glass sheet. While roller wave is caused by rollers on which a flat work piece rests or is transported during a heat-treatment such as tempering, an undulation may be caused by any process.

[0062] Bow – Bow refers to a deviation from flatness that occurs globally over the face of a glass sheet. In some cases, bow is measured at or near a center point of a glass sheet. In some cases, bow is measured from a three-point reference plane about the edge of the window. The value of bow is then calculated by measuring the location of the median surface at the center of the window and determining its distance from the reference plane.

[0063] In some embodiments, bow (or warp) is calculated using ASTM C1048 (section 11.6):

[0064] 11.6 Localized Warp and Overall Bow and Warp—Place sample glass in a freestanding vertical position, resting on blocks at the quarter points. With the glass in this position, place a straightedge across the concave surface, parallel to and within 25.4 mm (1 in.) of the edge, and measure the maximum deviation with a taper, or feeler gage, dial indicator or fine scale ruler. When the above procedure is impractical for larger sizes of 3-PP^^^»^-in.) thickness, place the glass on a flat surface, concave side down, and use a taper, feeler gage, dial indicator, or fine scale ruler, reading in 0.02-mm (0.001-in.) increments, to determine overall bow and warp. Overall bow and warp values shown on the second line of Table 2 [reproduced below] DSSO\^WR^^PP^^^»^^LQ^^^ when the alternative (horizontal) procedure is used.

[0065] While this specification generally describes “tempered” glass and methods and apparatus for producing tempered glass, the glass, methods, and apparatus described herein are understood to pertain to both heat strengthened and tempered glass unless otherwise clear from context. The term “heat-treated” glass is understood to include tempered glass, heat strengthened glass and other forms of glass subject to mechanical improvement from heat-treatments. Lamination

[0066] A laminated glass structure as disclosed herein includes at least two component glass layers, at least one of which may be a glass lite. In some cases, at least one of the component glass layers is a tempered glass layer, which is bonded to a different glass layer (sometimes referred to as “second glass layer”). The second glass layer may be another tempered glass layer, or it may be an untempered glass layer. Examples of untempered glass layers include annealed glass layers, heat strengthened glass (i.e., semi tempered or semi toughened glass), and the like.

[0067] The second glass layer may be bonded to the tempered glass layer by any of various types of bonding. In some examples, the bonding is adhesive bonding. For example, the laminated glass structure may have an adhesive layer between the tempered glass layer and the glass layer. The adhesive layer may comprise a resin. Examples of such resins include polyvinyl butyral (PVB) resins, polyurethane, ethylene vinyl acetate (EVA), ionoplast (e.g., available from SentryGlas, SBP), and combinations thereof.

[0068] Various techniques are used for lamination including interposing a solid sheet of resin between the glass layers and hot pressing, using a liquid resin, and so forth as known in the relevant art. Adhesive layers may be chosen appropriately according to the function, e.g., various structural and aesthetic properties are available depending upon how the laminate structure is to be deployed in a structure. Optical properties

[0069] As explained, laminated glass structures including at least one tempered glass sheet may exhibit lensing. In certain embodiments described herein, a laminated glass structure does not exhibit lensing caused by undulations on at least one surface of a glass sheet in the structure. This is because the bonded glass sheets do not contain significant roller wave or other undulations. Other optical properties and functions of lamination adhesives can include, e.g., tint color (static or dynamic electrochromism, UV blocking, antireflection, anti-iridescence, and the like. In some embodiments, the laminated glass structure includes an electrochromic device or other optically switchable device comprising one or more organic (e.g., polymeric) components (e.g., ion conducting layers, suspended particle layers, liquid crystals, etc.). In such embodiments, one or more such organic components may be included with the adhesive that bonds at least one tempered glass sheet. In this manner, the optically switchable device is formed when two glass sheets are laminated together. Such glass sheets may include transparent conducting layers such as transparent conducting oxides (e.g., indium tin oxide, fluorinated tin oxide, etc.) Undulations on Glass Surface

[0070] In certain embodiments, a laminated glass structure includes a glass sheet having at least one surface with no undulations with a peak-to-valley distance of about 0.5 mm or greater or no undulations of about 0.2 mm or greater. This peak-to-valley distance may be determined as described in, e.g., the discussion forum on ASTM C1048. In particular laminated glass structures, such surface may be bonded to a mating surface of an adjacent glass sheet in the laminated glass structure. In various embodiments, the glass sheet with no undulations having a peak-to-valley distance of about 0.5 mm or greater (or no undulations of about 0.2 mm or greater) is a tempered glass layer.Bow of Glass Surface

[0071] In certain embodiments, a laminated glass structure includes a glass sheet having at least one surface with no bow, or a bow of about 10 mm or less, or a bow of about 5 mm or less. This bow may be determined as described in, e.g., ASTM C1048. In particular laminated glass structures, such surface may be bonded to a mating surface of an adjacent glass sheet in the laminated glass structure. In various embodiments, the glass sheet with no bow (or a bow of about 10 mm or less or about 5 mm or less) is a tempered glass layer. Configurations of Laminated Glass

[0072] As mentioned, a laminated glass structure includes at least two component glass layers. In various embodiments, one or more of these component glass layers has an optically switchable glass layer provided on a surface. The optically switchable device may be an electrochromic device. A few example configurations of laminated glass structures are presented here.

[0073] Figure 1 depicts an embodiment of a laminated glass structure 101 that includes a tempered glass layer 105 laminated to a second glass layer 109. Second glass layer 109 may be a sheet of annealed glass, tempered glass, or any other substrate. If tempered, glass layer 109 typically has the same dimensional tolerances as tempered glass layer 105, i.e., at least in terms of roller-wave, undulation, and bow. Tempered glass layer 105 is bonded to second glass layer 109 by an interlayer adhesive 107 such as a resin (e.g., a PVB or an ionoplast). Second glass layer 109 may be inset from glass layer 105 about its perimeter so as to protect its edges and / or the optically switchable device thereon; although if glass layer 109 is tempered insetting may not be necessary. Glass layers 105 and 109 may be of the same or different thickness, each typically about 2mm to about 12 mm thick.

[0074] Laminated glass structure 101 includes an electrochromic device 111 disposed on a surface of second glass layer 109, which surface is opposite one that is bonded to tempered glass layer 105. Optionally, laminated glass structure 101 includes an anti-reflective layer that may be disposed on either surface of tempered glass layer 105 or second glass layer 109. In some cases,an anti-reflective layer is disposed on tempered glass layer 105 on the surface opposite adhesive 107.

[0075] If tempered glass layer 105 exhibited undulations such as roller waves, laminated glass structure 101 would exhibit lensing due to the adhesive 107 filling void regions formed in troughs of the undulations. In certain embodiments, tempered glass layer 105 is formed without roller waves or with roller waves having a peak-to-valley separation of no more than about 0.5 mm. In certain embodiments, tempered glass layer 105 does not have bow or warpage. In certain embodiments, tempered glass layer 105 does not have bow of greater than about 10 mm.

[0076] While not shown in Figure 1, a laminated structure having the elements of structure 101 may include one or more additional elements such as one or more additional glass layers (not shown). Structure 101 may be the outboard lite of an insulated glass unit.

[0077] In some embodiments, a laminated glass structure includes two or more tempered glass layers. In some cases, none of these tempered glass layers possesses a bow that will cause optical distortion (e.g., any bow is about 10 mm or less). In some cases, none of these tempered glass layers possesses undulations that will cause optical distortion (e.g., any undulations have a peak- to-valley distance of about 0.5 mm or less). Figures 2A and 2B present examples consistent with these requirements.

[0078] Figure 2A illustrates an exploded view of components 201 of a laminated glass structure. In some embodiments, components are separately fabricated in a process that forms a final laminated glass structure such as structure 231 of Figure 2B. Components 201 include a first laminated glass component 202 and a second laminated component 204. These may be separately fabricated and then joined to form the final laminated glass structure.

[0079] Component 202 includes a tempered first glass layer 205 bonded to a second glass layer 209. Bonding of layers 205 and 209 is accomplished by an adhesive interlayer 207. Interlayer 207 may include a resin such as a PVB or an ionoplast.

[0080] A partial electrochromic device 211 is disposed on a surface of second glass layer 209, opposite a surface bonded to tempered glass layer 205. Partial electrochromic device 211 may include one or more layers of an electrochromic device.

[0081] Component 204 may be a reflection of component 202 but with different components of an electrochromic device. In the depicted embodiment, component 204 includes a tempered third glass layer 215 bonded to a fourth glass layer 219. Bonding of layers 215 and 219 is accomplished by an adhesive interlayer 217. Interlayer 117 may include a resin such as a PVB or an ionoplast.

[0082] A second partial electrochromic device 221 is disposed on a surface of fourth glass layer 219, opposite a surface bonded to third tempered glass layer 215. Partial electrochromic device 221 may include one or more layers of an electrochromic device.

[0083] Partial electrochromic devices 211 and 221 include layers that form different and complementary portions of a complete electrochromic device. For example, partial device 211 may be an electrochromic layer (sometimes implemented using a cathodically coloring electrochromic material) and partial device 221 may be a counter electrode (optionally implemented using an anodically coloring electrochromic material). Partial devices may additionally contain layers used in electrochromic devices such as transparent conductive layers, ion conducting layers, etc.

[0084] Figure 2B illustrates a laminated glass structure 231 having a fully fabricated electrochromic device 230 formed from the partial electrochromic devices 211 and 221. A lamination adhesive 232 is used to bond 211 and 221 together; it serves as both an adhesive and an ion conductor (electrolyte) for device 230. Adhesive 232 may be applied as a polymeric sheet between the partial devices, as a liquid resin lamination, and the like. Adhesive 232 may have optical properties to complement the optical device’s optical properties.

[0085] Second glass layer 209 may be an annealed glass layer, a tempered glass layer, or any other form of glass layer. Independently, fourth glass layer 219 may be an annealed glass layer, a tempered glass layer, or any other form of glass layer. In certain embodiments, the combination of a fully fabricated electrochromic device 230 and the second and fourth glass layers 209 and 219 is relatively thin, e.g., on the order of about 1 to 5 mm.

[0086] In some examples, partial electrochromic device 211 includes an electrochromic cathode layer on glass layer 209 (e.g., a tungsten oxide on annealed glass). In some embodiments, partial electrochromic device 221 includes an electrochromic anode layer on fourth glass layer 219 (e.g.,a nickel tungsten oxide on annealed glass). In some embodiments, fully fabricated electrochromic device 230 includes an organic ion conducting layer between cathodic and anodic coloring electrochromic layers.

[0087] Note that laminated glass structure 231 illustrated in Figure 2B may be formed by methods other than that those employing components such as those represented in Figure 2A. For example, an electrochromic device laminate, 234, may be fabricated prior to lamination of the device laminate 234 to form laminated glass structure 231. Also, while not shown in Figure 2B, a laminated glass structure having the elements of structure 231 may include one or more additional elements such as one or more additional glass layers (not shown).

[0088] Figures 3A and 3B illustrate a variation of the Figure 2A-B laminated glass structure and fabrication technique. Figure 3A illustrates an exploded view of components 301 of a laminated glass structure. In some embodiments, components are separately fabricated in a process that forms a final laminated glass structure such as structure 331 of Figure 3B. Components 301 include a first laminated glass component 302 and a second laminated component 304. As with components 202 and 204 of Figure 2A, these may be separately fabricated and then joined to form the final laminated glass structure.

[0089] Component 302 includes a tempered first glass layer 305 and a partial electrochromic device 311 disposed on a surface of tempered glass layer 305. Partial electrochromic device 311 may include one or more layers of an electrochromic device.

[0090] Component 304 may be a reflection of component 302 but with different components of an electrochromic device. In the depicted embodiment, component 304 includes a tempered second glass layer 315 and a partial electrochromic device 321, which may include one or more layers of an electrochromic device.

[0091] Partial electrochromic devices 311 and 321 include layers that form different and complementary portions of a complete electrochromic device. For example, partial device 311 may be an electrochromic layer (sometimes implemented using a cathodically coloring electrochromic material) and partial device 321 may be a counter electrode (optionally implemented using an anodically coloring electrochromic material). Partial devices mayadditionally contain layers used in electrochromic devices such as transparent conductive layers, ion conducting layers, etc.

[0092] Figure 3B illustrates a laminated glass structure 331 having a fully fabricated electrochromic device 330 formed from the partial electrochromic devices 311 and 321. A lamination adhesive 332 is used to bond partial devices 311 and 321 together. Adhesive 332 may serve as both an adhesive and an ion conductor (electrolyte) for device 330. Adhesive 332 may be applied as a polymeric sheet between the partial devices, as a liquid resin lamination, and the like. Adhesive 332 may have optical properties to complement the optical device’s optical properties.

[0093] In some examples, partial electrochromic device 311 includes an electrochromic cathode layer on tempered first glass layer 305 (e.g., a tungsten oxide). In some embodiments, partial electrochromic device 321 includes an electrochromic anode layer on tempered second glass layer 315 (e.g., a nickel tungsten oxide). In some embodiments, fully fabricated electrochromic device 330 includes an organic ion conducting layer between cathodic and anodic coloring electrochromic layers.

[0094] Note that laminated glass structure 331 illustrated in Figure 3B may be formed by methods other than that those employing components such as those represented in Figure 3A. Also, while not shown in Figure 3B, a laminated glass structure having the elements of structure 331 may include one or more additional elements such as one or more additional glass layers (not shown). Further, in some embodiments, only one of glass layers 305 and 315 is tempered glass. The other may be, for example, annealed glass. Fabrication Methods

[0095] Any of various methods may be employed to fabricate laminated glass structures of the types disclosed herein. In some cases, the methods employ tempering a glass layer while supporting the glass layer, at least partially, on an air cushion such as air flow. In some embodiments, the tempering occurs without the glass layer touching a solid surface (or with only minimal touching). While supported, the glass layer is exposed to tempering conditions such as a thermal treatment.

[0096] In some embodiments, a laminated glass structure is fabricated by tempering a pre- tempered glass layer while supporting the pre-tempered glass layer with a gas, e.g., air, e.g., via a plurality of air jets. A pre-tempered glass layer is a glass layer that is not yet fully tempered but which may be subject to a tempering process and converted to a tempered glass layer. The tempering involves exposing the pre-tempered glass layer to tempering conditions that convert the pre-tempered glass layer to a tempered glass layer. The resulting tempered glass layer may have no or only minor surface undulations as described herein. The resulting tempered glass layer may have no or only minor surface bow as described herein.

[0097] To produce the laminated glass structure, the resulting tempered glass layer may be bonded to a second glass layer using a bonding process as described elsewhere herein.

[0098] In some embodiments, during tempering and while supported on the plurality of air jets, the pre-tempered glass layer does not contact a solid surface. In some embodiments, during tempering and while supported on the plurality of air jets, the pre-tempered glass layer does not contact a roller.

[0099] Generally, the glass layer is supported by gas, such that the gas pressure below the glass is higher than the gas pressure above the glass. In some embodiments, this is accomplished by applying gas pressure while the glass is in the horizontal or a substantially horizontal orientation, i.e., an applied pressure pushes against a major surface of the glass to lift the glass. The glass may be guided through the tempering environment by a gas pressure as well, i.e., with a net force resulting in a flow (e.g., a laminar flow) upon which the glass travels through the tempering environment. Gas pressures may also be used to guide the glass through the tempering environment, or not, e.g., a force with associated direction may be imparted to the glass prior to fully entering the tempering environment, which are sufficient to carry the glass through the tempering environment.

[0100] Temperatures within the tempering environment are consistent with conventional tempering conditions, i.e., a heating regime and a quenching regime, e.g., where the support gas and gas envelope around the glass is heated or not. For example, in a heating zone or zones of a furnace, the support gas may be heated as a result of heating elements in the zone and / or independently heated to aid thermal homogeneity of the heating regime. In conventionaltempering furnaces, the bottom surface of a glass sheet receives a greater heat flux than the top surface. This is due to the bottom surface receiving additional heat from rollers. The resulting temperature nonuniformity can introduce distortions within the tempered glass. When glass sheets are tempered while floating on air jets—as with various embodiments disclosed herein—this problem is eliminated or mitigated.

[0101] In the quenching regime, the support gas and gas envelope around the glass may be at ambient temperature, e.g., room temperature, and / or cooled to aid in quenching the hot glass to form the tension and compression zones of tempered glass.

[0102] As hot and cool gas have different densities, the pressure of the incoming gas to form the support gas and gas envelope, i.e., support and / or transport the glass through the tempering environment may be changed to accommodate density changes and uniformly support the glass through the tempering environment. This may require different manifolds and / or jetting but is well within the skill of ordinary artisans.

[0103] In some cases, during the tempering, the pre-tempered glass layer floats substantially horizontally on air produced using the plurality of air jets. In some implementations, during the tempering, the pre-tempered glass layer floats at an inclined angle of not greater than about 60 degrees (or not greater than about 20 degrees) from the horizontal plane on air produced using the plurality of air jets. In some embodiments, the pre-tempered glass layer floats at an inclined angle of not greater than about 10 degrees, or not greater than about 5 degrees, or not greater than about 2 degrees.

[0104] In some embodiments, the air jets produce air streams at an angle, or within a range of angles, that produces a force that may push the pre-tempered glass layer along a pathway in a tempering furnace that applies the tempering conditions.

[0105] In some embodiments, a tempering furnace includes one or more guardrails that keep the pre-tempered glass layer within a pathway of the tempering furnace. As examples, such guardrails may comprise rollers or immovable bars or rails, where only the edges of the glass are touched by a guardrail as the glass passes through the tempering environment.

[0106] In certain embodiments, exposing the pre-tempered glass layer to tempering conditions comprises heating the pre-tempered glass with heat emitted by one or more heating plates while the pre-tempered glass is supported on the plurality of air jets. In some implementations, the one or more heating plates include at least one heating plate disposed substantially under the pre- tempered glass.

[0107] Fabrication methods disclosed herein may include bonding a tempered glass layer having no or only minor undulations or bow to a second glass layer. The bonding may be performed by applying an adhesive material between the two glass layers and pressing those layers together.

[0108] Fabrication methods disclosed herein may include fabricating an electrochromic device on at least one glass surface such as one of the surfaces of the tempered first glass layer and / or one of the surfaces of the second glass layer.

[0109] In some examples, such fabrication methods include forming an electrochromic device on a surface of the second glass layer opposite the tempered first glass layer. The second glass layer may be tempered or untempered. For example, the second glass layer may be an annealed glass layer.

[0110] In some examples, such fabrication methods include forming a first portion of an electrochromic device on a surface of the second glass layer opposite the tempered first glass layer; forming a second portion of the electrochromic device on a surface of a fourth glass layer opposite a tempered third glass layer; and contacting the first portion of the electrochromic device to the second portion of the electrochromic device. Contacting the first and second portions of the electrochromic device provides a full electrochromic stack, including an electrochromic layer and a counter electrode layer. In some embodiments, the second glass layer is an annealed glass layer, the fourth glass layer is an annealed glass layer, or both the second and fourth glass layers are annealed glass layers.

[0111] A laminated glass structure as disclosed herein may be provided as a part of an insulated glass unit (IGU). In some embodiments, a fabrication method comprises assembling a laminated glass structure as disclosed herein, a mate lite, and a spacer in a manner that forms a sealed volumebetween the laminated glass structure and the mate lite. The sealed volume is typically filled with an inert gas to impart thermal insulative properties.

[0112] Figure 4 is an example flow chart illustrating certain processes 401 for fabricating a laminated glass structure. As illustrated, the process begins with an operation 403 in which a pre- tempered glass sheet is provided to a tempering apparatus. Next, in an operation 405, the pre- tempered glass sheet is tempered in the tempering apparatus. After that, the resulting tempered glass sheet is optionally processed at an operation 407 to apply one or more optical effects or features such as all or part of an optically switchable device, an anti-reflective layer, an anti- iridescence layer, a diffusion barrier, a passive tinting or coloring device, and the like. Next, the tempered glass sheet is assembled in an operation 409 with another glass sheet to form a laminated glass structure. In some embodiments, the resulting laminated glass structure is assembled into an insulated glass unit, a framed window, or other window structure. See operation 411.

[0113] In some embodiments, prior to providing a pre-tempered glass sheet to a tempering furnace, the sheet is cleaned to remove any particles that might be present on the surface of the sheet. Particles that are present before tempering will be baked into the sheet; they cannot be removed by post-tempering washing. The pre-tempering cleaning can be performed by any of various techniques such as washing with, e.g., a liquid comprising hot water, a surfactant (e.g., a detergent), and / or a basic solution (e.g., potassium hydroxide). Examples of other cleaning techniques include mechanical dry brush cleaning, plasma etching, high pressure air nozzle cleaning, and CO2cleaning.

[0114] In some embodiments, the tempering process is performed in an environment employing filtered air. One example of such environment is a clean room. Air reaching the glass sheet may be filtered regardless of origin, e.g., regardless of whether it is recirculated from a location within the tempering environment. Fabrication Apparatus

[0115] Apparatus used to fabricate a tempered glass sheet without significant roller wave and / or bow may take various forms. To reduce roller wave, an apparatus may include a furnace that exposes a pre-tempered glass sheet to tempering conditions without having the glass sheet contactrollers. In some embodiments, a furnace uses air to support and / or convey a pre-tempered glass sheet during tempering. As a result, there may be even pressure on the glass sheet’s major surfaces and little or no roller wave, bow, or other distortion.

[0116] Tempered glass is conventionally made by heating a pre-tempered glass sheet well above its transition temperature of about 564 °C (1,047 °F) to about 620 °C (1,148 °F), but typically not less than about 600 °C. The same is true for heat strengthened glass, but the cooling cycle(s) are different for tempering versus strengthening. Heat strengthened glass is generally 2 to 4 times as strong as annealed glass of the same thickness and configuration, while tempered glass is generally 2 to 3 times stronger than heat strengthened glass. As an example, heat strengthened glass generally achieves a residual surface compression of about 3,500 to 7,500 psi for 6 mm glass, according to ASTM C 1048. Tempered glass 6 mm thick must have either a minimum surface compression of 10,000 psi or an edge compression of not less than 9,700 psi, but for it to be considered safety glass, the surface compressive stress should exceed 15,000 psi.

[0117] In certain embodiments, “working temperature” for heat strengthening or tempering is about 600 to 640 °C or about 610-630 °C, or about 615-625 °C, or about 619- 621 °C as an example working temperature.

[0118] In some embodiments, a furnace employs an air float surface, which may take the form or be constructed similarly to an air float table, which uses a plenum or manifold to deliver air or other gas, e.g., inert gas, flow through apertures to produce air “jets,” in order to support the glass substrate above a surface, e.g., a table surface. An “air jet” or “air stream” may thus refer to the stream of air or other gas produced by a physical structure such as an aperture through which the air or other gas is passed. Thus a “plurality of air jets” may refer to such apertures or apertures, typically but not necessarily running through a surface such as a table, with air or other gas passing therethrough to support the substrate, or the corresponding air streams emanating from the physical apparatus.

[0119] In the case of a tempering furnace as described herein, the glass sheet’s major work surfaces (typically the two largest parallel planes of a glass sheet) are both exposed to such air jets while heating and quenching the glass substrate during the tempering process. In certain embodiments, the jets merely support the glass sheet, so that it does not touch the surface belowit, or possibly anything other than gas, during its traversal of the tempering furnace. For example, a roller system that supports the glass sheet at the inlet side of a tempering furnace may also transport the glass sheet into the furnace. The momentum imparted to the glass sheet from the roller system is sufficient to carry the glass sheet through the tempering furnace, while the glass sheet is supported by air jets. In some embodiments, the air jets are angled such that a flow is produced, sufficient to move the glass sheet through the tempering furnace. This may include air jets on one or both of the major work surfaces of the glass sheet. In some embodiments a combination of the glass sheet’s momentum entering the furnace and angled air jets is used to move the glass sheet through the tempering furnace.

[0120] In some embodiments, the tempering furnace is tilted from horizontal, from the end where the glass enters to the end where the glass sheet exits the furnace, in order to allow gravity to carry the glass sheet through the furnace. Combinations of the aforementioned movement means can be used to move the glass sheet through the furnace.

[0121] The edges of the glass sheet are not touched as the glass moves through the tempering furnace. In one embodiment, the glass sheet is introduced into the furnace with the appropriate momentum and alignment such that it passes through the furnace supported only by the plurality of air jets. In one embodiment, the glass sheet is introduced into the furnace where gravity pulls it through the furnace and the glass sheet is supported only by the plurality of air jets and aligned such that the glass sheet does not touch any solid surfaces. In certain embodiments, air jets are used to provide force acting upon the glass sheet edges in order to push or guide it through the furnace without allowing the glass sheet to touch any solid surfaces. In certain embodiments, air jets impinging on the edges of the glass sheet may be angled so as to provide not only a guide but also to impart a lateral force to move the glass through the furnace.

[0122] Figures 5-8B illustrate a few examples of ways in which a glass sheet may be supported and / or transported in a tempering environment such as a tempering furnace. A tempering environment may include both a heating zone and a cooling zone.

[0123] Figure 5 shows a glass sheet 503 supported against gravity by air jets that provide a first pressure 505 acting on a lower major surface of the glass sheet. By ambient conditions or otherwise, an upper major surface of glass sheet 503 experiences a second pressure 511. Duringtempering, first pressure 505 is greater than second pressure 511. In this manner, glass sheet 503 is supported on gas as it traverses the tempering environment. Note that pressures 505 and 511 have directions that are substantially normal to the upper and lower major surfaces of glass sheet 503. Glass sheet 503 traverses the tempering environment using momentum imparted prior to and / or during entry into the tempering environment. It need not touch anything but gas as it passes through. However, in some embodiments, there may be “guardrails” or “bumpers.” These need only lightly touch the edges to guide the glass sheet through.

[0124] Figure 6 shows a glass sheet 603 that, similar to the sheet in Figure 5, is supported against gravity by air jets that provide a first pressure 605 acting on a lower major surface of the glass sheet. And an upper major surface of glass sheet 603 experiences a second pressure 611. As in the case of Figure 5, during tempering, first pressure 605 is greater than second pressure 611, and, as a consequence, glass sheet 603 is supported on gas as it traverses the tempering environment. Unlike the case in Figure 5, pressure 605 has a direction that is not substantially normal to the upper and lower major surfaces of glass sheet 603. Rather, gas flow on the upper major surface of glass sheet 603 has a direction that may push the glass along by virtue of its force vector. In this example, the gas flow atop the glass pushes the glass along through the tempering environment. Such embodiments may establish laminar flow across the glass worksurface. In some embodiments, pressure 611 may also have a non-normal direction, and, as consequence, it too may contribute to the transport of glass sheet through the tempering environment. As in the case of Figure 5, glass sheet 603 may traverse the tempering environment without touching anything but gas as it passes through, or it may touch only “guardrails” or “bumpers.”

[0125] Figure 7 illustrates an implementation of the embodiments of Figure 5. As illustrated, a tempering system includes gas manifolds 713 and 715 that provide uniform gas pressure via appropriately spaced and sized apertures. Manifolds 713 and 715 may, as well, provide uniform heating. They may be made from metal, ceramic or other materials that can be evenly heated from behind (e.g., the manifold is between the glass sheet and a heater). The manifold and the gas temperature may be the same or substantially so. The manifolds are of sufficient thermal mass so as not be cooled by the incoming glass (glass sheet 703 in this example) to any significant measure (thus a stable heat source) and can handle high throughput of glass. Air jets from manifold 713provide a pressure 705 on an upper surface of sheet 703, and air jets from manifold 715 provide a pressure 711 on a lower surface of sheet 703.

[0126] In certain embodiments, a glass sheet is tempered in one pass through both a heating zone and quenching zone of a tempering furnace. For example, a pre-tempered glass sheet may enter a tempering furnace with momentum imparted prior to entering the furnace. The heating zone of the furnace is sufficiently long to bring the glass sheet up to a uniform temperature prior to entering a quenching zone. In some cases, a heating zone has different temperatures along a path followed by glass sheets. For example, a furnace may include relatively hot upstream zones configured to cause a glass sheet to quickly reach a desired temperature, and relatively cooler downstream zones configured to stabilize the glass sheet to a uniform final temperature suitable for tempering. Regardless of how many heating zones are employed, the glass sheet may be supported by a gas bed that keeps even pressure on the glass sheet. In many implementations, a heated glass sheet enters a quenching zone on the gas support. The quenching zone may also support the glass sheet with gas.

[0127] In some embodiments, a tempering environment employs a rastering mechanism such as depicted in Figures 8A-B. Such mechanism may be configured to push a floating glass sheet back and forth within a single, relatively short, heating zone until the sheet reaches a desired temperature. After reaching the desire temperature, a mechanism (such as one of the rastering mechanisms) provides a final push to float the glass downstream to enter one or more quench zone(s) which also support the glass sheet via gas.

[0128] Figures 8A-B illustrate an example of a tempering environment 801 that allows a glass sheet 803 to move back and forth within the environment. Such back and forth movement— sometimes referred to as rastering—is commonly employed in glass tempering furnaces. Rastering mechanisms may include rotating assemblies that push the glass upstream or downstream to help distribute heating of the glass sheet in the heating zone. As in the example depicted in Figures 8A-B, the mechanism may include arms or tines 821 and 823 that engage with the leading or trailing edge of a glass sheet to push it. Rotation may be timed to softly catch the glass, slow and stop its momentum, and then change its direction without imparting deforming forces to the glass sheet. The arms or tines may rotate on a shaft to enter and exit the plane of the glass to engagewith the glass sheet. A slot or channel in the heating manifold (e.g., heating manifold 815) may house the arms or tines 821 and 823 when not engaging with the glass sheet. Thus, the arms or tines 821 and 823 may be held at the same temperature as the heating environment by virtue of nesting in the channel or slot of the heating manifold 815. The arms or tines 821 and 823 may be made of metal for good heat transfer and may be very thin to minimize the physical footprint when touching the glass sheet. Since the glass sheet is floating on a bed of gas, minimal force may be needed to move the glass sheet or change its direction. There may be two or more arms or tines for directing the glass sheet in each direction, so that the collective force needed to stop or start the glass moving is spread across multiple arms or tines. The arms or tines may be springs or spring loaded to allow for gradual and / or less forceful stopping and starting the glass sheet moving.

[0129] The arms or tines may include an engagement of the mechanism to gradually engage a glass sheet in a manner that decreases its momentum gradually, then changes direction to push the glass in the opposite direction, in either case without sufficient force to warp or distort the shape of the glass.

[0130] Similar to other embodiments disclosed herein, the tempering environment may include manifolds 813 and 815 configured to generate gas jets that impart a pressure 811 on a lower major surface of glass sheet 803 and a pressure 805 on an upper major surface of glass sheet 803.

[0131] Note that the tempering environments presented in Figures 5-8B may be employed for the heating and / or quenching phases of tempering. Either or both the heating and quenching zones may employ a rastering mechanism such as illustrated in Figures 8A-B.

[0132] In certain embodiments, a tempering environment is configured to effect quenching in a manner that applies differing degrees of quenching to different locations within a quenching zone, e.g., a zone near the exit of a tempering furnace. This may be useful when a tempering process concurrently heats glass sheets of different sizes and / or different shapes (e.g., different aspect ratios). The collection of such glass sheets is sometimes called a mixed size load. When processing mixed size loads, larger glass sheets tend to have lower temperatures compared to smaller glass sheets in same load. Therefore, it may be desirable to more aggressively cool smaller glass sheets in a mixed load.

[0133] Applying differing degrees of quenching to different locations within a quenching zone may also be useful when the load includes relatively large glass sheets, e.g., sheets have a surface area of at least about 5000 square inches. Large glass sheets sometimes suffer having non-uniform temperatures, with their centers being cooler than their edges. This is because the sheet’s edges can get overheated due to radiant heat transfer. Selective cooling of the outer regions of large glass sheets can counteract non-uniform, center-to-edge temperature variations.

[0134] Selective cooling or bed load-dependent recipes may be employed to facilitate consistent and uniform cooling of glass undergoing quenching. In some implementations, a quenching zone is configured to produce one or more adjustable air jets, which may be issued via one or more quench nozzles near the exit of a tempering furnace. In some embodiments, a quench nozzle or other source of and air jet is translatable over a two-dimensional region in the quenching zone. This allows one or more air jets to be selectively directed or selectively activated in regions needing preferential cooling such as smaller glass sheets in a mixed load or the edge regions of glass sheets, regardless of load type.

[0135] In some embodiments, the quenching zone employs a plurality of adjustable air jets, which may be issued via a plurality of quench nozzles. Each adjustable air jet can have its air velocity, flow rate, pressure, and / or temperature controlled to permit differing levels of quenching at different locations. In some cases, the adjustment is merely turning the air jet on or off.

[0136] Because adjustable cooling may require rapid adjustments to account for moment-to- moment processing variations, a quenching system may employ feedback such as closed loop control. In some embodiments, closed loop control is implemented via imaging (mixed size load) and / or pyrometers (temperature non-uniformity).

[0137] Figure 8C schematically illustrates an example quenching apparatus 841 configured to effect quenching in a manner that applies differing degrees of quenching to different locations within the apparatus. Apparatus 841 includes a plurality of adjustable quench nozzles 843 from which air is blown on to glass in quench section. In Figure 8C, apparatus 841 is operating in a first mode in which each of a plurality of quench nozzles 843 has its air jet adjusted to tailor quenching conditions for different sized glass sheets that are concurrently undergoing quenching.

[0138] Figure 8C shows quench section with two mixed size lites in bed load that will have difference in temperature as when processing mixed size loads, larger glass sheets tend to have lower temperatures compared to smaller glass sheets in same load. Therefore, it may be desirable to more aggressively cool smaller glass sheets in a mixed load. In the figure, sheet 845 is a relatively larger glass sheet, and sheet 847 is a relatively smaller glass sheet.

[0139] In quenching mode depicted in Figure 8C, P1 and P2 represent air pressure intensity from respective nozzles, with P1 providing a higher pressure, and hence more aggressive cooling, to large glass sheet 845 and P2 providing a lower pressure, and hence less aggressive cooling, to small glass sheet 847. Larger glass sheets typically have lower exit temperatures entering into quench so these sheets will need to be more aggressively cooled to establish a compression layer on their outer surfaces and tension in their interiors before the glass temperature drops below its transition temperature to solid. Hence, the depicted embodiments, a higher pressure is used for large glass sheets and a lower pressure for small sheets.

[0140] Figure 8D schematically illustrates quenching apparatus 841 operating a different mode, which may be appropriate for quenching large glass sheets, such as large sheet 849 shown in the figure, that have non-uniform temperatures, with their centers being cooler than their edges. Selective cooling of the outer regions of large glass sheets can counteract non-uniform, center-to- edge temperature variations. In Figure 8D, some adjustable quench nozzles 843 produce air jets at a higher pressure P1 and other adjustable quench nozzles 843 produce air jets at a lower pressure P2. This variation in air jet pressure impinging on glass sheet 849 counteracts the non-uniform temperature that the glass sheet has when it enters quenching apparatus 841. Note that, upon entering a quenching apparatus, the center of a large glass sheet is typically at a lower temperature compared to its edges. Therefore, the center region of the sheet must be rapidly cooled to establish a compression layer on its outer surfaces and tension in its interior before center region reaches its transition temperature to solid. By contrast, the edge of sheet, which is at a higher temperature, can be cooled using less pressure. EC devices and IGUs

[0141] A schematic cross-section of an electrochromic device 900 in accordance with some embodiments is shown in Figure 9. Such a device may be formed on one or more glass sheets employed to form a laminated glass structure as described herein. The electrochromic device includes a substrate 902, a conductive layer (CL) 904, an electrochromic layer (EC) 906, an ion conducting layer (IC) 908, a counter electrode layer (CE) 910, and a conductive layer (CL) 914. Elements 904, 906, 908, 910, and 914 may be referred to collectively as an electrochromic stack 920. In some cases, the ion conductor layer 908 may be omitted. A voltage source 916 operable to apply an electric potential across the electrochromic stack 920 effects the transition of the electrochromic device from, e.g., a bleached state to a colored state. In other embodiments, the order of layers is reversed with respect to the substrate. That is, the layers are in the following order: substrate, conductive layer, counter electrode layer, ion conducting layer, electrochromic material layer, conductive layer.

[0142] Electrochromic layer 906 is cathodically tinting, while the counter electrode layer 910 may be anodically tinting or optically passive (sometimes referred to as an “ion storage layer” because ions reside there when the device is not tinted). In some embodiments, electrochromic layer 906 comprises a tungsten oxide. In some embodiments, counter electrode layer 910 comprises a nickel tungsten oxide.

[0143] It should be understood that the reference to a transition between a bleached state and tinted state is non-limiting and suggests only one example, among many, of an electrochromic transition that may be implemented. Unless otherwise specified herein, whenever reference is made to a bleached-tinted transition, the corresponding device or process encompasses other optical state transitions such non-reflective-reflective, transparent-opaque, colored-uncolored, etc. Further the term “bleached” refers to an optically neutral state, e.g., uncolored, transparent or translucent. Still further, unless specified otherwise herein, the “tint” or “color” of an electrochromic transition is not limited to any particular wavelength or range of wavelengths. As understood by those of skill in the art, the choice of appropriate electrochromic and counter electrode materials governs the relevant optical transition.

[0144] In certain embodiments, the electrochromic device reversibly cycles between a bleached state and a tinted state. In the bleached state, a potential is applied to the electrochromic stack 920such that available ions in the stack that can cause the electrochromic material 906 to be in the tinted state reside primarily in the counter electrode 910. When the potential on the electrochromic stack is reversed, the ions are transported across the ion conducting layer 908 to the electrochromic material 906 and cause the material to enter the tinted state.

[0145] In certain embodiments, all the materials making up electrochromic stack 920 are inorganic, solid (i.e., in the solid state), or both inorganic and solid. Because organic materials tend to degrade over time, inorganic materials offer the advantage of a reliable electrochromic stack that can function for extended periods of time. Materials in the solid state also offer the advantage of not having containment and leakage issues, as materials in the liquid state often do. It should be understood that any one or more of the layers in the stack may contain some amount of organic material, but in many implementations one or more of the layers contains little or no organic matter. The same can be said for liquids that may be present in one or more layers in small amounts. It should also be understood that solid state material may be deposited or otherwise formed by processes employing liquid components such as certain processes employing sol-gels or chemical vapor deposition.

[0146] Referring again to Figure 9, voltage source 916 is typically a low voltage electrical source (on the order of between about 1V and about 20V, depending upon the electrochromic device used) and may be configured to operate in conjunction with radiant and other environmental sensors. Voltage source 916 may also be configured to interface with an energy management system, such as a computer system that controls the electrochromic device according to factors such as the time of year, time of day, and measured environmental conditions. Such an energy management system, in conjunction with large area electrochromic devices (i.e., an electrochromic window), can dramatically lower the energy consumption of a building.

[0147] Figure 10 depicts an insulated glass unit (IGU) 1020 with an EC lite 1000, which may comprise a laminated glass structure as disclosed herein. EC lite 1000 includes an EC device and associated pair of bus bars 1005 which each energize the device via a transparent conductor. The pair of transparent conductors sandwich the EC materials between them so that a potential can be applied across the device materials. The IGU is fabricated by combining EC lite 1000 with a spacer 1010, and a mate lite 1015 along with the appropriate sealants and wiring (not shown) tothe bus bars. In some applications, a second set of spacer and mate lite may be added (e.g., a triple Pane IGU). The two mate lites may be of different types. As depicted on the bottom half of Figure 10, the IGU can be transparent (left), tinted to an intermediate state (middle) or fully tinted (right).

[0148] Aspects of this disclosure pertain to integrated glass units including at least two lites and a hermetically sealed void region between the two lites. At least one of the lites is a laminated glass structure having a tempered glass layer with minimal or no undulations or bow as described elsewhere herein. Other than possessing such tempered glass layer, the laminated glass structure may have any configuration.

[0149] Determining modifications to tempering recipes and / or bedload layout configurations

[0150] Many of the techniques described above relate to tools and hardware for tempering lites, e.g., such that a lite being tempered does not touch rollers and is instead moved through regions of a furnace using a set of air jets. Additionally disclosed herein are techniques for determining tempering recipe parameters and / or bedload layout configurations that cause improved lite tempering (e.g., to minimize lite bowing, minimize distortion, or to meet other lite tempering specifications) using existing tools and / or hardware, e.g., including those that utilize rollers to move lites. For example, disclosed herein are techniques for utilizing a machine learning model for determining tempering recipe parameters and / or bedload layout configurations for a batch of lites undergoing tempering based on parameters associated with a previously tempered batch of lites. It should be understood however that, in some embodiments, the techniques disclosed below for utilizing a machine learning model for determining tempering process parameters may be applied to a tool that utilizes air jets to move lites through regions of the furnace, for example, by determining changes to air temperature or pressure, timing of when air jets are on or off, or the like.

[0151] Conventionally, a batch of lites may be tempered, which may involve heating the lites to a high temperature (e.g., 600 or 700 degrees Celsius, or a similarly high temperature) and then quenching, or rapidly cooling the lites. The lites may be heated by moving the lites through a furnace (as described above), either using rollers or air jets (as described above). The lites may have a lite specification that is to be achieved after the tempering process. As used herein, a lite specification may involve bowing being below a target threshold, break weight being within atarget break weight criteria, gasp or compression being within a target range, and / or visual distortion being below a target threshold. It may be difficult to perform the tempering process such that the lites achieve the target lite specifications, as described above. Moreover, in some cases, tempering a first batch of lites may induce changes in aspects of the furnace that cause a second batch of lites, when processed using the same tempering process parameters, to not achieve the target lite specifications due to, e.g., residual thermal load on rollers of the furnace from processing a previous first batch of lites. The techniques described allow for modifications in tempering process parameters and / or a tempering recipe, and / or determination of a bedload configuration or layout that increase the likelihood that lites will achieve the target lite specifications.

[0152] In some embodiments, modifications to tempering recipe parameters and / or determination of a bedload layout configuration may be determined. Tempering recipe parameters and / or bedload layout configurations may be determined based on parameters associated with a first batch of lites that undergo tempering. For example, the modifications to the recipe parameters and / or the bedload layout configuration may be determined by providing the parameters associated with the first batch of lites undergoing tempering to a trained machine learning model, where the trained machine learning model is configured to provide, as output, information indicative of changes to the tempering recipe and / or a bedload layout for a subsequent batch of lites that undergo processing. For example, the parameters associated with the first batch of lites may include a layout of the first batch of lites (e.g., spacing of lites with respect to one another within the bedload), size and / or dimension information associated with each lite in the first batch of lites, bowing information associated with bowing of lites of the first batch of lites after undergoing tempering, distortion data, break weight, surface compression, or any combination thereof. The trained machine learning model may then generate output that indicates modifications to a tempering recipe and / or a bedload layout that cause the subsequent batch of lites that are processed using the modified recipe and / or the determined bedload layout to meet one or more lite specifications. The lite specifications may include criteria related to maximum bowing, break weight, distortion, gasp or compression, etc. Modifications to a tempering recipe may include changes to a heating and / or cooling profile used to perform the tempering. In some embodiments, the bedload layout may indicate a relative placement of lites. For example, the bedload layout may indicate where lites of particular dimensions are to be placed relative to lites of otherdimensions, e.g., that Lite A having a first size is to be placed next to Lite B having a second size, and so on.

[0153] FIG. 11 is a flowchart of an example process 1100 for determining tempering recipe parameters and / or bedload layout configurations in accordance with some embodiments. In some embodiments, blocks of process 1100 may be performed by one or more processors of one or more computing devices. For example, a computing device may be a desktop computer, a tablet computer, a laptop computer, or the like. The computing device(s) may be associated with lite tempering equipment. In some embodiments, the computing device(s) may be local to lite tempering equipment, or may be remote from the tempering equipment. For example, in some embodiments, the one or more computing devices may be cloud devices. In some embodiments, blocks of process 1100 may be executed in an order other than that shown in Figure 11. In some embodiments, two or more blocks of process 1100 may be performed substantially in parallel. In some embodiments, one or more blocks of process 1100 may be omitted.

[0154] At 1102, process 1100 can obtain parameters associated with a trained machine learning model configured to receive inputs indicative of parameters associated with lite tempering and configured to output one or more recipe and / or bedload changes for a subsequent batch of lites to be tempered. The parameters associated with the trained machine learning model may include weights associated with nodes of the machine learning model, where the weights were determined as a result of the training process, and / or an architecture of the machine learning model.

[0155] At 1104, process 100 can obtain parameters associated with a first batch of lites undergoing tempering. In some embodiments, the parameters associated with the first batch of lites may include a heating and / or cooling profile that specifies temperatures and timing information (e.g., that lites are to be heated in a zone of a furnace having a specified temperature and for a specified duration of time, that lites are to be cooled to a particular temperature and within a specified duration of time, etc.). The parameters associated with lite tempering may include characteristics of the first batch of lites after the tempering process was performed. For example, the parameters may include lite bowing information (which may be destructively measured on a subset of lites of the first batch of lites), distortion information (e.g., measured in units of milli- diopters) that indicates a magnitude of roller waves on the lites, gasp measurements which indicatesurface compression, break weight, or any combination thereof. The parameters associated with the first batch of lites may include a layout of the first batch of lites, e.g., within the furnace or within particular furnace zones, and / or with respect to rollers of the furnace. For example, the layout may indicate relative placements of lites with respect to one another, e.g., that Lite A having particular dimensions was at a particular location relative to Lite B having particular dimensions. In some embodiments, the parameters associated with the first batch of lites may include pyrometer data and / or data derived from pyrometer data. For example, various image processing algorithms may be applied (e.g., edge detection, object detection, etc.) to the pyrometer data to determine lite layout, mean and / or standard deviation of temperatures across the lites, or the like.

[0156] Process 1100 can determine, using the trained machine learning model, one or more modifications to a current recipe used to process the first batch of lites and / or a bedload configuration for a second batch of lites, wherein the one or more modifications are used to process the second batch of lites. The one or more modifications to the current recipe may include changes to heating and / or cooling profiles associated with the tempering recipe. For example, the one or more modifications may include changes to temperatures of one or more heating and / or cooling zones. As another example, the one or more modifications may include changes to durations of time lites spend in various heating and / or cooling zones.

[0157] In some embodiments, the bedload configuration for the second batch of lites may account for deficiencies of the first batch of lites with respect to characteristics of the first batch of lites after the tempering process, such as deficiencies in bowing, gasp, visual distortion, break weight, or the like. For example, the bedload configuration for the second batch of lites may be one that is determined by the trained machine learning model to be more likely to allow the second batch of lites to achieve lite tempering specifications. In some embodiments, the bedload configuration for the second batch of lites may account for thermal load or heat signature characteristics of rollers of the furnace due to processing of the first batch of lites. For example, in an instance in which the first batch of lites included a large lite, there may be cool spots on the rollers due to the processing of the large lite in the first batch of lites. Accordingly, the bedload configuration for the second batch of lites may indicate that particular sized lites should be placed in the cool zone, e.g., lites having an area less than a predetermined threshold.

[0158] The second batch of lites may then be processed using the modified tempering recipe and / or the bedload configuration determined using the output of the trained machine learning model. Due to the modified tempering recipe and / or the bedload configuration, the second batch of lites may be tempered in a manner such that lites of the second batch of lites are likely to meet tempering specifications, such as a lite bowing metric below a predetermined threshold, visual distortion metrics below a predetermined threshold, gasp metrics satisfying predetermined criteria, and / or break weight satisfying predetermined criteria.

[0159] It should be noted in instances in which a furnace without rollers is utilized, the machine learning model may determine tempering process modifications that include modifications to air flow of a plurality of air jets used to move a batch of lites through the regions of the furnace. For example, the modifications may include modifications to a temperature of air from one or more air jets, modifications to a pressure of air from one or more air jets, and / or whether one or more air jets are disabled or turned off.

[0160] As described above, in some embodiments, tempering recipe parameters and / or bedload layout configurations may be determined using a trained machine learning model. The machine learning model may be any suitable type of model. For example, the machine learning model may be a supervised machine learning model, such as a regression model (e.g., a linear regression model), a random forest model, a neural network, etc. For some types of models, feature selection may be performed prior to initializing and / or training the machine learning model to determine which parameters (e.g., lite bowing metrics, temperature measurements, gasp measurements, compression measurements, etc.) are to be used to train the model. Feature selection may be performed using principal component analysis (PCA), linear discriminant analysis (LDA), or the like. For some types of models, feature selection may not be explicitly performed. For example, using a collection of random forest models, each model may utilize a different combination of parameters with each model providing output predictions. Continuing with this example, aggregated outputs from the different random forest models of the collection of random forest models may be used to determine which parameter or combination of parameters are to be used by the trained machine learning model.

[0161] In some embodiments, in addition to or alternatively to using a supervised machine learning model, a reinforcement model may be used. When using a reinforcement model, the model may provide an output, such as a recipe change or a bedload configuration layout. The result of utilizing the output may be determined using a training dataset to determine whether to reward or punish the model based on the output (e.g., whether utilizing the output recipe parameters and / or bedload layout achieves target lite tempering specifications, the degree to which the output recipe parameters and / or bedload layout causes target lite tempering specifications to be missed, etc.), and the degree to which the model should be rewarded or punished. In this way, the reinforcement model may be trained to generate output tempering recipe parameters and / or bedload layouts that are generally rewarded, and, consequently, are more likely to achieve tempered lites that meet specifications. Through training the reinforcement learning model, an optimal (or approximately the optimal) solution may be determined based on the version of the model that receives the maximum reward.

[0162] Regardless of whether supervised training or reinforcement learning is used, a training set may be used to train the model. For supervised training, the training set may include lite parameters paired with a ground truth decision for which the model is to be trained to generate. By way of example, the lite parameters may include any combination of a bedload layout or a position of a particular lite, lite dimensions (e.g., length, width, area, thickness, etc.), temperature information, lite bowing information indicating lite bowing metrics after tempering, lite distortion metrics indicating visual distortion after tempering, gasp metrics indicating gasp after tempering, and / or break weight metrics indicating break weight after tempering. The ground truth paired with a given training sample may include one or more modifications to the tempering process and / or a bedload layout that is to be implemented to temper a subsequent batch of lites responsive to tempering a first batch of lites associated with the lite parameters of the training sample. In other words, each training sample may include parameters associated with a first batch of processed lites (which may include information indicative of the results of the tempering process), and a paired ground truth decision that indicates how the tempering process or bedload layout is to be modified for a subsequent batch of lites. The ground truth decision may include changes to the heating and / or cooling profiles, and / or a bedload layout for a second batch of lites.

[0163] In instances in which reinforcement learning is utilized, the dataset may indicate tempering process parameters and / or bedload layouts used to process a given batch of lites and lite tempering parameters that results for the batch of lites when tempered using the tempering process parameters and / or bedload layouts. The tempering parameters may include bowing information, gasp information, break weight information, visual distortion information, or any combination thereof. By updating the reinforcement model using metrics indicative of the outcome of the tempering process, the model may be trained to generate more optimal tempering process parameters and bedload layouts that are more likely to yield lites within a tempering specification.

[0164] In some embodiments, a machine learning model may additionally or alternatively take, as an input, temperature information indicating, e.g., temperature across a lite or at different portions of the lite. For example, the temperature information may indicate a mean temperature of a lite, a deviation across the lite (e.g., a standard deviation of temperature across the lite and / or a variance of temperature across the lite), or the like. In some embodiments, the temperature information may be derived from pyrometer data. The pyrometer data may be represented as image data. In some embodiments, image processing techniques may be applied to the pyrometer image data to derive temperature data. For example, the image processing techniques may involve performing edge detection to identify lite edges to determine pixels associated with a given lite. Based on the pyrometer heat map value and the determined pixel locations, temperature information associated with each lite may be determined.

[0165] Figure 12 is a flowchart of an example process 1200 for training a machine learning model in accordance with some embodiments. The machine learning model may be a supervised machine learning model that is trained to take, as input, parameters associated with a first batch of lites that underwent tempering, and generating, as an output, modifications to the tempering recipe and / or a bedload configuration to be used for a subsequent batch of lites. In some embodiments, blocks of process 1200 may be executed by one or more processors of one or more computing devices, such as a server device, a desktop computer, a laptop computer, etc. In some embodiments, the computing device may be a cloud device. In some embodiments, blocks of process 1200 may be executed in an order other than what is shown in Figure 12. In some embodiments, two or more blocks of process 1200 may be executed substantially in parallel. In some embodiments, one or more blocks of process 1200 may be omitted.

[0166] Process 1200 can begin at 1202 by initializing a machine learning model configured to receive inputs indicative of parameters associated with lite tempering and output one or more recipe and / or bedload changes for a subsequent batch of lites to be tempered. For example, initializing the machine learning model may comprise setting weights of the machine learning model to initial values. The machine learning model may be a linear regression model, a neural network model, a random forest model, a decision tree, etc.

[0167] It should be noted that, prior to initializing the machine learning model, various pre- training steps may be taken. For example, in some embodiments, data to be used as part of a training set may be cleaned to, e.g., remove outlier data. As another example, in some embodiments, feature selection may be performed using PCA, LDA, or the like, e.g., to select parameters to be used as inputs to the machine learning model.

[0168] At 1204, process 1200 can utilize the machine learning model to generate an output comprising a recipe change and / or a bedload layout given an input of parameters associated with tempering of a batch of lites. In other words, the machine learning model may take, as an input, parameters associated with tempering of the batch of lites, and generate, as an output, one or more recipe changes and / or a bedload layout to be used in conjunction with tempering a second subsequent batch of lites. As described above, the input parameters may include information regarding dimensions of the first batch of layouts, a layout of lites of the first batch of lites (e.g., relative placement of lites with respect to one another), metrics indicating outcomes of the tempering process on the first batch of lites (e.g., bowing metrics, visual distortion metrics, break weight metrics, gasp metrics, compression metrics, etc.), temperature information indicating temperatures at various locations of the lites or across a lite at different time points of the tempering process, or any combination thereof. The output may include one or more recipe changes, such as that a heating or cooling zone is to be modified to have a different temperature relative to the temperature used to process the first batch of lites, that lites are to remain in a particular heating or cooling zone for a different duration of time relative to the time used for the first batch of lites, or the like. As another example, the output may include a bedload layout for the second subsequent batch of lites, such as that a particular sized lite is to be placed in a particular region of a set of rollers used to move the lites through the furnace, that two lites of particular relative sizes (e.g., a big lite and a small lite) are to be placed next to one another, or any other placement information.

[0169] At 1206, process 1200 can update weights of the machine learning model based on a difference between a predicted recipe and / or bedload change and the ground truth recipe and / or bedload change indicated in the training data. For example, in an instance in which the ground truth data indicates that the temperature of a heating zone is to be raised by 100 degrees and in which the output of the model indicates that the heating zone is to be raised by 50 degrees, the weights may be updated based on the error between the model output and the ground truth output, e.g., based on the difference of 50 degrees. Error in bedload layout may be determined based on how close lites of different sizes are placed relative to the ground truth layout, a distance between a center of a given lite in the predicted layout and the center of the lite in the ground truth layout, etc. In some embodiments, the error may be provided to a cost function, and weights of the model may be updated based on the cost function. For example, the weights may be updated based on a minima of the cost function (e.g., where the error is minimized). Updating of the weights may be based on moving in the direction of steepest descent toward the minima, e.g., using gradient descent. In some cases, backpropagation may be used to calculate the gradients used to update the weights in the direction of steepest descent. Any suitable learning rate may be utilized in updating the weights, and the learning rate may be tune using hyperparameter tuning.

[0170] At 1208, process 1200 can determine whether to continue training the model. For example, process 1200 can determine whether more than a predetermined number of training iterations have been performed, or whether the weights were updated by less than a predetermined threshold relative to the previous value of the weights, and can determined, in response, that training is finished.

[0171] If, at 1208, process 1200 determines that the model should continue to be trained (“yes” at 1208), process 1200 can loop back to block 1204 and can continue training the model. Conversely, if, at 1208, process 1200 determines that training has been completed (“no” at 1208), process 1200 can end.

[0172] Parameters (e.g., weights) of the trained model may then be used to determine recipe changes and / or bedload layouts, e.g., as shown in and described above in connection with Figure 11.

[0173] It should be noted that although Figure 12 depicts a machine learning model being trained based on ground truth data that indicates a target tempering process change and / or target bedload layout that the model is trained to predict, in some embodiments, the model may instead be trained to predict lite characteristics of a tempered lite undergoing the tempering process with particular process parameters and / or or particular bedload layout. For example, the model may be trained to predict that a lite tempered with a particular tempering process may have a predicted bow, a predicted break weight, a predicted gasp or compression, and / or a predicted distortion. Continuing with this example, modifications to the tempering process and / or a bedload layout may be determined based on the predicted lite characteristics, for example, to determine modifications and / or a bedload layout more likely to bring the predicted lite characteristics within a target specification. As a more particular example, in an instance in which the model predicts that lites will have a bow that exceeds a bow threshold, modifications to the tempering process and / or the bedload layout may be determined that are likely to cause lites to have a bow that is below the threshold. In some embodiments, modifications to the tempering process and / or the bedload layout may be determined in an iterative process. For example, a machine learning model may predict, for tempering process parameters A (e.g., for a particular layout and / or heating / cooling profile), lite characteristics B (e.g., a predicted bow, a predicted break weight, a predicted distortion, a predicted gasp or compression, etc.) resulting from process parameters A. Continuing with this example, responsive to determining that lite characteristics B cause the lites to be out of specification, process parameters A’ may be determined, and the model may be used to determined lite characteristics B’ resulting from process parameters A’. Lite characteristics B’ may be closer to target specification(s) than lite characteristics B, however, responsive to determining that lite characteristics B’ are also out of target specifications, process parameters A’’ may be determined. This process may be repeated iteratively until the predicted lite characteristics are determined to be within specification.

[0174] In some embodiments, information associated with a thermal load of a set of rollers used to temper a first batch of lites may be used to determine tempering recipe changes and / or a bedload layout to be used to process a subsequent second batch of lites. For example, in some embodiments, the information associated with the thermal load of a set of rollers used to temper a first batch of lites may indicate that a particular region of the rollers is cooler due to a relatively large lite of the first batch of lites being placed on it, or conversely that a different region of therollers is warmer due to a relatively smaller lite of the first batch of lites being placed on it during the tempering process. Continuing with this example, in some embodiments, the bedload layout for the second subsequent batch of lites may include having a relatively smaller lite placed in a cool zone of the rollers that previously had a larger lite of the first batch of lites, and / or may include having a relatively larger lite placed in a warmer zone of the rollers that previously had a smaller lite of the first batch of lites. Additionally or alternatively, in some embodiments, the recipe changes may include changes to a heating or cooling profile (e.g., as described above.

[0175] Figure 13 is a flowchart of an example process 1300 for modifying a tempering process and / or a bedload layout based on a thermal load of rollers of a furnace in accordance with some embodiments. In some embodiments, blocks of process 1300 may be performed by one or more processors of one or more computing devices. The one or more computing devices may be local to the furnace or remote (e.g., a cloud device). In some embodiments, blocks of process 1300 may be performed in an order other than what is shown in Figure 13. In some embodiments, two or more blocks of process 1300 may be executed substantially in parallel. In some embodiments, one or more blocks of process 1300 may be omitted.

[0176] Process 1300 can begin at 1302 by obtaining information associated with a thermal load of a set of rollers used to process a first batch of lites. The thermal load information may include heating information and / or temperature information associated with rollers of the set of rollers. The thermal load information may be derived based on thermocouple data. For example, the thermocouple data may be used to determine that a given region of the set of rollers is cooler or warmer than another region of the set of rollers.

[0177] At 1304, process 1300 may determine one or more of a layout of a second batch of lites to be processed using a set of rollers or a modification to a recipe to be used to process the second batch of lites based on the thermal load. As described above, the layout may indicate that a lite of a given size is to be positioned in a particular region due to the thermal load. For example, a smaller lite may be placed in a cooler zone of the set of rollers, and a larger lite may be placed in a warmer zone of the set of rollers. In some embodiments, the layout may indicate relative positions of lites of the second batch of lites with respect to one another. In some embodiments, the layout may indicate spacing between two lites of the batch of lites, e.g., that the two lites areto be a given distance apart from one another. The layout information may be determined based on known dimensional information of lites in the second batch of lites, such as known lengths and / or widths of the lites, known aspect ratios, etc. In some embodiments, modifications to the recipe may include changes to a heating and / or a cooling profile, as described above.

[0178] At 1306, process 1300 may process the second batch of lites using the determined layout and / or the modified recipe.

[0179] Figure 14 depicts a schematic diagram of an example system for modifying tempering processes and / or determining bedload layouts in accordance with some embodiments. As illustrated, a lite tempering control system 1402 may be configured to determine recipe modifications and / or bedload layouts. For example, lite tempering control system 1402 may be configured to utilize data associated with tempering of a first batch of lites to determine recipe modifications and / or bedload layout configurations to be utilized for a second batch of lites, for example, to increase the likelihood that the second batch of lites, after being tempered, meet certain lite specifications (e.g., with respect to bowing, visual distortion, break weight, gasp, or the like). As described above in connection with Figures 11 and 12, in some embodiments, lite tempering control system 1402 may determine the recipe modifications and / or the bedload layouts using a trained machine learning model. Additionally or alternatively, as shown in and described above in connection with Figure 13, in some embodiments, lite tempering control system may determine the recipe modifications and / or the bedload layouts based on a residual thermal load of rollers of the furnace due to processing of the first batch of lites.

[0180] As shown in Figure 14, lite tempering control system 1402 may receive various data. For example, lite tempering controls system 1402 may receive furnace load data 1404. Furnace load data 1404 may indicate characteristics of lites of a first batch of lites that undergo a tempering process, such as sizing or dimension information of the first batch of lites, thickness information, a layout of the lites of the first batch of lites within the furnace and / or on the rollers, or the like.

[0181] Lite tempering control system 1402 may receive heating and / or temperature data 1406. In some embodiments, the heating and / or temperature data 1406 may include pyrometer data, which may be image data. The pyrometer data may be used to determine temperature information at various locations on lites of the first batch of lites while in the furnace, a mean temperature of agiven lite, a standard deviation or variance of temperature across a lite, or the like. In some embodiments, the heating and / or temperature data 1406 may include thermocouple data indicating temperature information of various locations within the furnace or associated with regions of a set of rollers.

[0182] Lite tempering control system 1402 may receive quench data 1408. Quench data 1408 may be obtained after the first batch of lites are moved to a quench zone where the lites undergo rapid cooling after being heated in one or more zones of a furnace. In some embodiments, quench data 1408 may include data obtained after the quenching process is completed. For example, quench data 1408 may include visual distortion measurements, e.g., that indicate peak to valley metrics for waves or undulations in the glass layer. The visual distortion measurements may be Osprey measurements.

[0183] Lite tempering control system 1402 may receive furnace unload data 1410. Furnace unload data 1410 may include various measurements such as break weight, gasp, and / or lite bowing metrics. Note that, in some embodiments, some data of furnace unload data 1410 may be destructive and may accordingly be measured on only a subset of the first batch of lites.

[0184] Data 1404-1410 may be provided to lite tempering control system 1402 during and after processing of the first batch of lites. Based on data 1404-1410, lite tempering control system 1402 may determine a manner in which the tempering process is to be adjusted when implemented with a subsequent second batch of lites. The adjustments may account for deficiencies observed in the first batch of lites due to the tempering process used for the first batch of lites (e.g., deficiencies of excessive bow, deficiencies in break weight and / or gasp, visual distortion exceeding a particular distortion threshold, etc.), and / or to compensate for thermal load due to lites of the first batch of lites.

[0185] It should be noted in instances in which a furnace without rollers is utilized, lite tempering control system 1402 may determine tempering process modifications that include modifications to air flow of a plurality of air jets used to move a batch of lites through the regions of the furnace. For example, the modifications may include modifications to a temperature of air from one or more air jets, modifications to a pressure of air from one or more air jets, and / or whether one or more air jets are disabled or turned off.Other Embodiments and Conclusion

[0186] Although omitted for conciseness, embodiments of the system and / or method can include every combination and permutation of the various system components and the various method processes, wherein one or more instances of the method and / or processes described herein can be performed asynchronously (e.g., sequentially), concurrently (e.g., in parallel), or in any other suitable order by and / or using one or more instances of the systems, elements, and / or entities described herein.

[0187] As a person skilled in the art will recognize from the previous detailed description and from the figures and claims, modifications and changes can be made to the preferred embodiments of the invention without departing from the scope of this invention defined in the following claims.

[0188] Example Embodiments: Embodiment 1. A laminated glass structure comprising: a heat-treated first glass layer having no undulations with a peak-to-valley distance of about 0.2 mm or greater; and a second glass layer adhesively bonded to the heat-treated glass layer, wherein the laminated glass structure does not exhibit lensing caused by undulations. Embodiment 2. The laminated glass structure of embodiment 1, further comprising an optically switchable device. Embodiment 3. The laminated glass structure of embodiment 2, wherein the optically switchable device is an electrochromic device. Embodiment 4. The laminated glass structure of any of the foregoing embodiments, further comprising an adhesive layer between the heat-treated glass layer and the glass layer. Embodiment 5. The laminated glass structure of embodiment 4, wherein the adhesive layer comprises a resin.Embodiment 6. The laminated glass structure of embodiment 5, wherein the resin comprises polyvinyl butyral (PVB), ionoplast (SentryGlas, SBP), or a combination thereof. Embodiment 7. The laminated glass structure of any of the foregoing embodiments, further comprising an electrochromic device disposed on a surface of the second glass layer opposite the heat-treated first glass layer, and wherein the second glass layer is an annealed glass layer. Embodiment 8. The laminated glass structure of any of embodiments 1-6, further comprising an electrochromic device disposed on a surface of the second glass layer opposite the heat-treated first glass layer, wherein the second glass layer is an annealed glass layer; a heat-treated third glass layer; and a fourth glass layer adhesively bonded to the heat-treated third glass layer. Embodiment 9. The laminated glass structure of embodiment 8, wherein the electrochromic device is also disposed on a surface of the fourth glass layer. Embodiment 10. The laminated glass structure of any of embodiments 1-6, further comprising an electrochromic device disposed on a surface of the second glass layer opposite the heat- treated first glass layer, and wherein the second glass layer comprises heat-treated glass. Embodiment 11. The laminated glass structure of any of the foregoing embodiments wherein the heat-treated first glass layer is a tempered glass layer. Embodiment 12. An integrated glass unit comprising the laminated glass structure of any of embodiments 1-11; a mate lite; and a sealed void region between the laminated glass structure and the mate lite. Embodiment 13. A laminated glass structure comprising: a heat-treated first glass layer having no bow or a bow of about 10 mm or less; and a second glass layer adhesively bonded to the heat- treated glass layer.Embodiment 14. The laminated glass structure of embodiment 13, further comprising an optically switchable device. Embodiment 15. The laminated glass structure of embodiment 13 or 14, further comprising an adhesive layer between the heat-treated glass layer and the glass layer. Embodiment 16. The laminated glass structure of any of embodiments 13-15, further comprising an electrochromic device disposed on a surface of the second glass layer opposite the heat- treated first glass layer, and wherein the second glass layer is an annealed glass layer. Embodiment 17. The laminated glass structure of any of embodiments 13-15, further comprising an electrochromic device disposed on a surface of the second glass layer opposite the heat- treated first glass layer, wherein the second glass layer is an annealed glass layer; a heat-treated third glass layer; and a fourth glass layer adhesively bonded to the heat-treated third glass layer. Embodiment 18. The laminated glass structure of embodiment 17, wherein the electrochromic device is also disposed on a surface of the fourth glass layer. Embodiment 19. The laminated glass structure of any of embodiments 13-15, further comprising an electrochromic device disposed on a surface of the second glass layer opposite the heat- treated first glass layer, and wherein the second glass layer comprises heat-treated glass. Embodiment 20. The laminated glass structure of any of embodiments 13-19, wherein the heat- treated first glass layer is a tempered glass layer. Embodiment 21. A laminated glass structure comprising: a heat-treated first glass layer produced by heat-treating a pre-heat-treated glass layer in a furnace in which the pre-heat-treated glass layer is supported by a plurality of air jets; and a second glass layer adhesively bonded to the heat-treated glass layer.Embodiment 22. The laminated glass structure of embodiment 21, further comprising an optically switchable device. Embodiment 23. The laminated glass structure of embodiment 21 or 22, further comprising an adhesive layer between the heat-treated glass layer and the glass layer. Embodiment 24. The laminated glass structure of any of embodiments 21-23, further comprising an electrochromic device disposed on a surface of the second glass layer opposite the heat- treated first glass layer, and wherein the second glass layer is an annealed glass layer. Embodiment 25. The laminated glass structure of any of embodiments 21-23, further comprising an electrochromic device disposed on a surface of the second glass layer opposite the heat- treated first glass layer, wherein the second glass layer is an annealed glass layer; a heat-treated third glass layer; and a fourth glass layer adhesively bonded to the heat-treated third glass layer. Embodiment 26. The laminated glass structure of embodiment 25, wherein the electrochromic device is also disposed on a surface of the fourth glass layer. Embodiment 27. The laminated glass structure of any of embodiments 21-23, further comprising an electrochromic device disposed on a surface of the second glass layer opposite the heat- treated first glass layer, and wherein the second glass layer comprises heat-treated glass. Embodiment 28. The laminated glass structure of any of embodiments 21-27, wherein the heat- treated first glass layer is produced by tempering the pre-heat-treated glass layer in the furnace. Embodiment 29. A method of fabricating a laminated glass structure, the method comprising: heat-treating a pre-heat-treated glass layer by supporting the pre-heat-treated glass layer on a plurality of air jets while exposing the pre-heat-treated glass layer to heat-treating conditions, wherein heat-treating the pre-heat-treated glass layer produces a heat-treated first glass layer; and adhesively bonding the heat-treated first glass layer to s second glass layer.Embodiment 30. The method of embodiment 29, wherein during the heat-treating, the pre-heat- treated glass layer does not contact a solid surface. Embodiment 31. The method of embodiment 29, wherein during the heat-treating, the pre-heat- treated glass layer does not contact a roller. Embodiment 32. The method of embodiment 29, 30, or 31, wherein during the heat-treating, the pre-heat-treated glass layer floats substantially horizontally on air produced using the plurality of air jets. Embodiment 33. The method of embodiment 29, 30, or 31, wherein during the heat-treating, the pre-heat-treated glass layer floats at an inclined angle of not greater than about 10 degrees from the horizontal plane on air produced using the plurality of air jets. Embodiment 34. The method of embodiment 33, wherein during the heat-treating, the pre-heat- treated glass layer floats substantially horizontal on air produced using the plurality of air jets. Embodiment 35. The method of any of embodiment 32, 33, or 34, wherein the air jets produce air streams at an angle, or within a range of angles, that produces a force to push the pre-heat- treated glass layer along a pathway in a heat-treating furnace that applies the heat-treating conditions. Embodiment 36. The method of embodiment 35, wherein the heat-treating furnace includes one or more guardrails that keep the pre-heat-treated glass layer within the pathway of the heat- treating furnace. Embodiment 37. The method of any of embodiments 29-36, wherein exposing the pre-heat- treated glass layer to heat-treating conditions comprises heating the pre-heat-treated glass with heat emitted by one or more heating plates while the pre-heat-treated glass is supported on the plurality of air jets.Embodiment 38. The method of embodiment 37, wherein the one or more heating plates include at least one heating plate disposed substantially under the pre-heat-treated glass. Embodiment 39. The method of any of embodiments 29-38, further comprising fabricating an electrochromic device on the heat-treated first glass layer and / or on the second glass layer. Embodiment 40. The method of any of embodiments 29-39, further comprising forming an electrochromic device on a surface of the second glass layer opposite the heat-treated first glass layer, and wherein the second glass layer is an annealed glass layer. Embodiment 41. The method of any of embodiments 29-38, further comprising: forming a first portion of an electrochromic device on a surface of the second glass layer opposite the heat- treated first glass layer, wherein the second glass layer is an annealed glass layer; forming a second portion of the electrochromic device on a surface of an annealed fourth glass layer opposite a heat-treated third glass layer; and contacting the first portion of the electrochromic device to the second portion of the electrochromic device. Embodiment 42. The method of any of embodiments 29-38, further comprising: forming an electrochromic device on a surface of the second glass layer opposite the heat-treated first glass layer, and wherein the second glass layer comprises heat-treated glass. Embodiment 43. The method of any of embodiments 29-42, wherein the heat-treating comprising quenching the pre-heat-treated glass layer by exposing different portions the pre- heat-treated glass layer to different cooling rates. Embodiment 44. The method of any of embodiments 43, wherein exposing the different portions the pre-heat-treated glass layer to different cooling rates is performed with one or more quenching air jets. Embodiment 45. The method of any of embodiments 29-44, wherein the heat-treating comprises tempering the pre-heat-treated glass layer.Embodiment 46. A method of forming an integrated glass unit, the method comprising: assembling the laminated glass structure produced by a method of any of embodiments 29-45, a mate lite, and a spacer in a manner that forms a sealed void region between the laminated glass structure and the mate lite. Embodiment 47. An apparatus comprising: a furnace configured to accommodate a glass layer and comprising: a heater; an air float surface; a plurality of apertures in the air float surface; and a plenum or manifold configured to flow a gas flow through apertures to produce air jets that can impinge on a first surface of the glass layer and support the glass layer above the air float surface. Embodiment 48. The apparatus of embodiment 47, further comprising a second plurality of apertures configured direct air onto a second surface of the glass layer, opposite the first layer. Embodiment 49. The apparatus of embodiment 47 or 48, wherein the apertures of the plurality of apertures are configured to direct the air jets at a non-normal angle with respect to the first surface of the glass layer. Embodiment 50. The apparatus of any of the embodiments 47-49, wherein the furnace is configured such that no edges of the glass sheet touch a surface of the furnace while the glass layer moves through the tempering furnace. Embodiment 51. The apparatus of any of the embodiments 47-50, further comprising a quenching zone comprising a plurality of quenching apertures configured to provide quenching air jets that can support the glass sheet while in the quenching zone. Embodiment 52. The apparatus of embodiment 51, wherein at least one of the plurality of quenching apertures comprises a nozzle configured to adjust the pressure of a quenching air jet emitted by the nozzle.Embodiment 53. A method, comprising: receiving parameters associated with a trained machine learning model configured to receive inputs indicative of parameters associated with lite tempering and output one or more recipe and / or bedload changes for a subsequence batch of lites so to be tempered; obtaining parameters associated with a first batch of lites undergoing tempering; determining, by providing the obtained parameters to the trained machine learning model, one or more modifications to: a current recipe used to process the first batch of lites; and / or a bedload configuration for a second batch of lites, wherein the one or more modifications are used to process a second batch of lites. Embodiment 54. The method of embodiment 53, wherein the parameters associated with the first batch of lites comprise layout of a plurality of lites in the first batch of lites within a furnace used to perform the tempering. Embodiment 55. The method of embodiment 54, wherein the layout is determined based at least on pyrometer data. Embodiment 56. The method of embodiment 55, wherein the layout is determined by applying image processing techniques to the pyrometer data. Embodiment 57. The method of embodiment 56, wherein the image processing techniques are used to determine mean and standard deviations of temperatures across the first batch of lites. Embodiment 58. The method of any one of embodiments 53-57, wherein the parameters associated with the first batch of lites comprises bowing information associated with bowing of lites in the first batch of lites. Embodiment 59. The method of any one of embodiments 53-58, wherein the parameters associated with the first batch of lites comprises distortion data. Embodiment 60. The method of any one of embodiments 53-59, wherein the parameters associated with the first batch of lites comprises break weight of the first batch of lites.Embodiment 61. The method of any one of embodiments 53-60, wherein the parameters associated with the first batch of lites comprises surface compression of the first batch of lites. Embodiment 62. The method of any one of embodiments 53-61, wherein the machine learning model was trained by updating weights associated with the machine learning model based at least in part on a comparison of a ground truth recipe modification and / or predicted bedload configuration for a training sample to a predicted recipe modification and / or predicted bedload configuration for the training sample. Embodiment 63. The method of any one of embodiments 53-62, wherein the one or more modifications cause the second batch of lites to meet a lite specification, wherein the lite specification comprises at least one of: a target lite bowing metric, a target break weight, a target distortion metric, or a target compression metric. Embodiment 64. The method of any one of embodiments 53-63, wherein the one or more modifications comprise one or more modifications to the current recipe, and wherein the one or more modifications to the current recipe comprise a modification to a heating and / or cooling profile used to perform the tempering. Embodiment 65. The method of any one of embodiments 53-64, wherein the one or more modifications are to the bedload configuration for the second batch of lites, and wherein the modifications to the bedload configuration comprise placement of lites of the second batch of lites based on size. Embodiment 66. The method of embodiment 65, wherein the placement of lites comprises a relative placement of lites with respect to one another in the second batch of lites. Embodiment 67. The method of any one of embodiments 53-66, wherein the one or more modifications comprises modifications to air flow provided by one or more jets within a furnace.Embodiment 68. The method of embodiment 67, wherein the modifications to the air flow comprise a modification to a pressure provided by the jet. Embodiment 69. A method, comprising: a) initializing a machine learning model configured to receive inputs indicative of parameters associated with lite tempering and output one or more recipe and / or bedload changes for a subsequent batch of lites so to be tempered; b) utilizing the machine learning model to generate an output comprising a recipe change and / or a bedload layout given an input of parameters associated with tempering of a batch of lites indicated in one or more training samples; c) updating weights of the machine learning model based at least in part on a difference between the generated output and a ground truth recipe changed and / or bedload layout indicated in the one or more training samples; d) repeating operations (b) and (c) until a determination that the machine learning model has been trained is made. Embodiment 70. The method of embodiment 69, wherein the parameters associated with the batch of lites comprise a layout of a plurality of lites in the first batch of lites within a furnace used to perform the tempering. Embodiment 71. The method of embodiment 70, wherein the layout is determined based at least on pyrometer data. Embodiment 72. The method of embodiment 71, wherein the layout is determined by applying image processing techniques to the pyrometer data. Embodiment 73. The method of any one of embodiments 69-72, wherein the parameters associated with tempering of the batch of lites comprises at least one of: distortion data, bowing information, break weight, and / or surface compression. Embodiment 74. The method of any one of embodiments 69-73, wherein the machine learning model comprises a regression model.Embodiment 75. The method of any one of embodiments 69-74, further comprising performing feature selection prior to initializing the machine learning model in operation (a). Embodiment 76. The method of embodiment 75, wherein the feature selection comprises use of principal components analysis (PCA) or linear discriminant analysis (LDA). Embodiment 77. The method of embodiments 75 or 76, further comprising performing data cleaning to remove a subset of data associated with the first batch of lites prior to initializing the machine learning model in operation (a). Embodiment 78. The method of any one of embodiments 69-77, wherein the machine learning model comprises a random forest architecture. Embodiment 79. The method of any one of embodiments 69-77, wherein the machine learning model utilizes a reinforcement learning architecture. Embodiment 80. A method comprising: determining information associated with residual heat load of a set of rollers of a furnace used to perform a tempering process on a first batch of lites; and determining at least one of: 1) a modification to the tempering process, wherein the modified tempering process is to be used to process a second batch of lites; or 2) a bedload layout of the second batch of lites, wherein the modified tempering process and / or the bedload layout account for effects of the residual heat load of the set of rollers. Embodiment 81. The method of embodiment 80, wherein the modified tempering process comprises modifications to a heating and / or cooling profile to be applied. Embodiment 82. The method of any one of embodiments 80 or 81, wherein the bedload layout of the second batch of lites comprises locations for lites of the second batch of lites based at least in part on lite size.Embodiment 83. A system comprising: one or more processors; and a non-transitory computer- readable medium storing instructions that, upon execution by the one or more processors, cause the one or more processors to perform operations of any one of embodiments 53-82. Embodiment 84. A non-transitory computer-readable medium storing instructions that, upon execution by one or more processors, cause the one or more processors to perform operations of any one of embodiments 53-82.

Claims

CLAIMS what is claimed is:

1. A laminated glass structure comprising: a heat-treated first glass layer having no undulations with a peak-to-valley distance of about 0.2 mm or greater; and a second glass layer adhesively bonded to the heat-treated glass layer, wherein the laminated glass structure does not exhibit lensing caused by undulations.

2. The laminated glass structure of claim 1, further comprising an optically switchable device.

3. The laminated glass structure of claim 2, wherein the optically switchable device is an electrochromic device.

4. The laminated glass structure of any of the foregoing claims, further comprising an adhesive layer between the heat-treated glass layer and the glass layer.

5. The laminated glass structure of claim 4, wherein the adhesive layer comprises a resin.

6. The laminated glass structure of claim 5, wherein the resin comprises polyvinyl butyral (PVB), ionoplast (SentryGlas, SBP), or a combination thereof.

7. The laminated glass structure of any of the foregoing claims, further comprising an electrochromic device disposed on a surface of the second glass layer opposite the heat-treated first glass layer, and wherein the second glass layer is an annealed glass layer.

8. The laminated glass structure of any of claims 1-6, further comprising an electrochromic device disposed on a surface of the second glass layer opposite the heat-treated first glass layer, wherein the second glass layer is an annealed glass layer; a heat-treated third glass layer; and a fourth glass layer adhesively bonded to the heat-treated third glass layer.

9. The laminated glass structure of claim 8, wherein the electrochromic device is also disposed on a surface of the fourth glass layer.

10. The laminated glass structure of any of claims 1-6, further comprising an electrochromic device disposed on a surface of the second glass layer opposite the heat-treated first glass layer, and wherein the second glass layer comprises heat-treated glass.

11. The laminated glass structure of any of the foregoing claims wherein the heat-treated first glass layer is a tempered glass layer.

12. An integrated glass unit comprising the laminated glass structure of any of claims 1-11; a mate lite; and a sealed void region between the laminated glass structure and the mate lite.

13. A laminated glass structure comprising: a heat-treated first glass layer having no bow or a bow of about 10 mm or less; and a second glass layer adhesively bonded to the heat-treated glass layer.

14. A laminated glass structure comprising: a heat-treated first glass layer produced by heat-treating a pre-heat-treated glass layer in a furnace in which the pre-heat-treated glass layer is supported by a plurality of air jets; and a second glass layer adhesively bonded to the heat-treated glass layer.

15. The laminated glass structure of claim 14, further comprising an optically switchable device.

16. A method of fabricating a laminated glass structure, the method comprising: heat-treating a pre-heat-treated glass layer by supporting the pre-heat-treated glass layer on a plurality of air jets while exposing the pre-heat-treated glass layer to heat-treating conditions, wherein heat-treating the pre-heat-treated glass layer produces a heat-treated first glass layer; and adhesively bonding the heat-treated first glass layer to s second glass layer.

17. The method of claim 16, wherein during the heat-treating, the pre-heat-treated glass layer does not contact a solid surface.

18. The method of claim 16, wherein during the heat-treating, the pre-heat-treated glass layer does not contact a roller.

19. The method of claim 16, wherein during the heat-treating, the pre-heat-treated glass layer floats substantially horizontally on air produced using the plurality of air jets.

20. An apparatus comprising: a furnace configured to accommodate a glass layer and comprising: a heater; an air float surface; a plurality of apertures in the air float surface; and a plenum or manifold configured to flow a gas flow through apertures to produce air jets that can impinge on a first surface of the glass layer and support the glass layer above the air float surface.

21. The apparatus of claim 20, further comprising a second plurality of apertures configured direct air onto a second surface of the glass layer, opposite the first layer.

22. A method, comprising: receiving parameters associated with a trained machine learning model configured to receive inputs indicative of parameters associated with lite tempering and output one or more recipe and / or bedload changes for a subsequence batch of lites so to be tempered; obtaining parameters associated with a first batch of lites undergoing tempering; and determining, by providing the obtained parameters to the trained machine learning model, one or more modifications to: a current recipe used to process the first batch of lites; and / or a bedload configuration for a second batch of lites, wherein the one or more modifications are used to process a second batch of lites.

23. The method of claim 22, wherein the parameters associated with the first batch of lites comprise layout of a plurality of lites in the first batch of lites within a furnace used to perform the tempering.

24. The method of claim 23, wherein the layout is determined based at least on pyrometer data.

25. The method of claim 24, wherein the layout is determined by applying image processing techniques to the pyrometer data.

26. The method of claim 25, wherein the image processing techniques are used to determine mean and standard deviations of temperatures across the first batch of lites.

27. The method of any one of claims 22-26, wherein the parameters associated with the first batch of lites comprises bowing information associated with bowing of lites in the first batch of lites.

28. The method of any one of claims 22-26, wherein the parameters associated with the first batch of lites comprises distortion data.

29. The method of any one of claims 22-26, wherein the parameters associated with the first batch of lites comprises break weight of the first batch of lites.

30. The method of any one of claims 22-26, wherein the parameters associated with the first batch of lites comprises surface compression of the first batch of lites.

31. The method of any one of claims 22-26, wherein the machine learning model was trained by updating weights associated with the machine learning model based at least in part on a comparison of a ground truth recipe modification and / or predicted bedloadconfiguration for a training sample to a predicted recipe modification and / or predicted bedload configuration for the training sample.

32. The method of any one of claims 22-26, wherein the one or more modifications cause the second batch of lites to meet a lite specification, wherein the lite specification comprises at least one of: a target lite bowing metric, a target break weight, a target distortion metric, or a target compression metric.

33. The method of any one of claims 22-26, wherein the one or more modifications comprise one or more modifications to the current recipe, and wherein the one or more modifications to the current recipe comprise a modification to a heating and / or cooling profile used to perform the tempering.

34. The method of any one of claims 22-26, wherein the one or more modifications are to the bedload configuration for the second batch of lites, and wherein the modifications to the bedload configuration comprise placement of lites of the second batch of lites based on size.

35. The method of claim 34, wherein the placement of lites comprises a relative placement of lites with respect to one another in the second batch of lites.

36. The method of any one of claims 22-26, wherein the one or more modifications comprises modifications to air flow provided by one or more jets within a furnace.

37. The method of claim 36, wherein the modifications to the air flow comprise a modification to a pressure provided by the jet.

38. A method, comprising: a) initializing a machine learning model configured to receive inputs indicative of parameters associated with lite tempering and output one or more recipe and / or bedload changes for a subsequent batch of lites so to be tempered;b) utilizing the machine learning model to generate an output comprising a recipe change and / or a bedload layout given an input of parameters associated with tempering of a batch of lites indicated in one or more training samples; c) updating weights of the machine learning model based at least in part on a difference between the generated output and a ground truth recipe changed and / or bedload layout indicated in the one or more training samples; and d) repeating operations (b) and (c) until a determination that the machine learning model has been trained is made.

39. The method of claim 38, wherein the machine learning model comprises a regression model.

40. The method of any one of claims 38 or 39, further comprising performing feature selection prior to initializing the machine learning model in operation (a).

41. The method of claim 40, wherein the feature selection comprises use of principal components analysis (PCA) or linear discriminant analysis (LDA).

42. The method of claim 40, further comprising performing data cleaning to remove a subset of data associated with the first batch of lites prior to initializing the machine learning model in operation (a).

43. The method of any one of claims 38 or 39, wherein the machine learning model comprises a random forest architecture.

44. The method of any one of claims 38 or 39, wherein the machine learning model utilizes a reinforcement learning architecture.

45. A method comprising: determining information associated with residual heat load of a set of rollers of a furnace used to perform a tempering process on a first batch of lites; anddetermining at least one of: 1) a modification to the tempering process, wherein the modified tempering process is to be used to process a second batch of lites; or 2) a bedload layout of the second batch of lites, wherein the modified tempering process and / or the bedload layout account for effects of the residual heat load of the set of rollers.

Images