Method and system for measuring parameters related to the stability of a floating batch of materials

Description

1 / 22 “METHOD AND SYSTEM FOR MEASURING PARAMETERS RELATED TO THE STABILITY OF A FLUCTUATING BATCH OF MATERIALS” Technical field

[0001] The invention relates to computer-implemented methods and systems for monitoring the stability of a batch of material floating in a melting reservoir in an electric glass melting tank. Fundamentals of the technique

[0002] In conventional electric glass melting furnaces, a plurality of electrodes is immersed in a predetermined pattern within a pool of molten glass. An electric current is caused to flow through the molten glass between the electrodes to heat the glass by the Joule effect.

[0003] A batch of materials, which may comprise raw materials, glass shards, and other recycled materials, is continuously or intermittently supplied to the upper surface of the pool to provide both a source of materials and an insulating layer or crust. However, as it gradually melts to form additional molten glass, the thickness of the layer decreases, and the heat loss from the molten glass body in the furnace through the batch increases. Conversely, as additional discontinuous material is distributed over the upper surface of the molten glass, the heat loss through the thicker layer decreases.

[0004] The batch is usually supplied in a predetermined pattern by a feeder, conveyor, or mobile sprinkler to carefully control the quantity and maintain a minimum thickness over the top surface of the pool to reduce heat loss, protect the feeder, and prevent furnace overflow.

[0005] It is common practice to inspect the interior of glass melting furnaces using infrared optical systems placed within the furnace walls. These systems allow for human observation of the stability of the batch of materials.

[0006] JPS 5 339 204 A, NIPPON ELECTRON OPTICS LAB 11.04.1978, JPH 0 7216 422 A, NIPPON STEEL CORP 15.08.1995, JP 2010 002 150 A, TAKUMA CO LTD 07.01.2010 and US 2018 231875 A1, HER MAJESTY THE QUEEN IN RIGHT OF CANADA AS REPRESENTED BY THE MINISTRY OF NATURAL Petition 870250094514, dated 10 / 16 / 2025, p. 15 / 48 2 / 22 RESOURCES [CA] 16.08.2018 describes an inspection system comprising an infrared sensor positioned in front of a viewing window within the furnace wall.

[0007] More sophisticated systems are also provided in the technique for measuring parameters related to some characteristics of the batch of materials.

[0008] WO 8002833 A1, OWENS CORNING FIBERGLASS CORP [US], 24.12.1980 describes a system for controlling the level, i.e., the thickness, of a batch of materials in an electric glass melting furnace. The system comprises an infrared sensor that is mounted on or adjacent to a feeder to obtain a non-contact measurement of the external surface temperature of the batch. The measured temperature is compared to a setpoint temperature and the level, i.e., the thickness, of the batch is increased or decreased by adjusting the feed rate of the feeder according to an empirical relationship between the thickness and the external surface temperature of the batch.

[0009] US 4 194 077 A, OWENS CORNING FIBERGLASS CORP [US], The March 18, 1980 patent describes a system for controlling the level of a batch of materials in an electric glass melting furnace. The system comprises an ultrasonic sensor that is mounted on a feeder and moves with it along the batch. A non-contact measurement of the batch level is obtained, and the batch thickness can be calculated from a relationship between the densities of the batch and the molten glass and their respective levels.

[0010] US 4,409,012 A, OWENS ILLINOIS INC [US], 11.10.1983 describes a method and apparatus for monitoring the surface coverage of a batch of materials floating in a melt pool by processing a video recording of the top surface of the batch and the melt within a glass melting furnace. Bimodal distributions of the number of pixels as a function of their gray level are then extracted from the recorded images and, after a threshold operation to separate the two modes, the relative quantities of batch and melt are estimated by integrating the area covered by each mode for different regions within the furnace.

[0011] JPH 0 656 432 A, NIPPON ELECTRIC GLASS CO, 01.03.1994 describes a method and a system for measuring the batch level of materials floating in a melt pool by processing an image of a portion of the Petition 870250094514, dated 10 / 16 / 2025, page 16 / 48 3 / 22 upper surface of the lot that is irradiated by a lighting device. The image is binarized and the barycentric coordinates of the white pixels are calculated. It is assumed that a change in the barycentric coordinates thus calculated reflects a change in the lot level due to a variation in the amount of light reflected by the illuminated portion of the lot when its level is moving up or down.

[0012] WO 0248057 A1, SOFTWARE & TECH GLAS GMBH [DE], 20.06.2002 describes a method for measuring the linearized increase in batch coverage level in a melt pool by processing an image of the top surface of the batch and the melt. The ratio of dark pixels to the number of pixels within each line of the image in the direction transverse to the melt flow direction is calculated. The evolution of the ratio in the flow direction is assumed to reflect the batch coverage level in that direction.

[0013] EP 1 655 570 A1, FRANKE MATTHIAS [DE], 10.05.2006 discloses a computer-implemented method for monitoring the internal area of ​​a glass melting furnace by comparing the positions of selected objects between infrared images of the internal area acquired at successive points in time. The method can be used to monitor the movement of different moving objects, for example, the batch of materials floating in the melt pool.

[0014] WO 2018 / 104695 A1, LAND INSTRUMENTS INTERNATIONAL LTD [GB], 14.06.2018, discloses a control unit for identifying the batch of materials floating in a melt pool from a thermal imaging camera configured to acquire thermal images of the top surface of the batch and the melt. The batch's position is identified by the low-temperature regions of the thermal images. The batch's movement can be tracked and its speed calculated by processing thermal images acquired at successive points in time. The control unit can be configured to correct the perspective view of the thermal image, therefore the parameters derived from it are in real-world coordinates.

[0015] WO 2022 / 242843 A1, GLASS SERVICE AS [CZ], 24.11.2022 discloses an improvement to the method described in EP 1 655 570 A1, FRANKE MATTHIAS [DE], 10.05.2006. The method further includes a compensation step of Petition 870250094514, dated 10 / 16 / 2025, page 17 / 48 4 / 22 deviation to correct the deviation within infrared images that is due to variation in the infrared camera's view over time. Compared to the method described in EP 1 655 570 A1, FRANKE MATTHIAS [DE], As of May 10, 2006, data derived from image processing are more precise and allow for the determination of new parameters, such as batch temperature, batch melting rate, batch position, and batch migration patterns and speed. Summary of the invention Technical Problem

[0016] One key process parameter, among others, for the efficient operation of an electric glass or stone furnace is the thickness of the discontinuous materials, since it directly affects the amount of energy to be supplied for homogeneous melting.

[0017] However, current methods suffer from several disadvantages such as local heterogeneities within the batch, in particular what are called hot spots or volcanoes and correspond to holes within the batch that are forming after complete local melting and / or, due to the granular nature of the batch and its inherent instabilities as materials piled up in a moving molten pool, after sliding movements within the batch, are not well considered in the measurement or calculation based on it.

[0018] Methods based on measuring batch and melt levels and calculating their differences fail to accurately model the heterogeneous batch coverage over the melt pool in an electric or glassstone melting furnace. In fact, within the electric or glassstone melting furnace, batch portions can drift away from each other under convective effects occurring within the melt pool, similar to tectonic plates on Earth. Unlike glassstone or fuel-fired melting furnaces, this phenomenon can lead to strong batch thickness variation across the melt pool surface. For example, while in a fuel-fired melting furnace batch thickness is expected to decrease as it moves towards the distribution channels, in an electric melting furnace it may remain relatively stable in the distribution channels or decrease in directions other than the melt flow direction. Petition 870250094514, dated 10 / 16 / 2025, p. 18 / 48 5 / 22

[0019] On the other hand, for the same reasons, methods for estimating batch thickness that rely solely on a batch level measurement and assumptions about the relationship between batch and molten glass densities and their levels suffer from the same flaws. In fact, assuming that a heterogeneous batch should have the same density everywhere can lead to overestimation unless a correction function is applied to account for local variations. However, such a correction may not be easy, as prior modeling of the evolution of batch properties and dynamics over time may be necessary.

[0020] Furthermore, since a pressure drop can occur between the melting and refining sections in an electric glass or stone melting furnace, the level of molten glass within these two sections may not be equal. Using, as can currently be done in the art, the level in the refining section as a substitute for estimating the level of molten glass under the batch of materials in the melting section may then lead to an incorrect estimate of the same, and thus the thickness of the batch that can be determined from this estimate may be incorrect.

[0021] On the other hand, methods based on batch image processing and fusion should be able to provide results that are more representative of batch heterogeneities. Since the images are direct acquisitions of the batch along with its 'hot spots' or 'volcanoes', the actual surface coverage of the batch over the melt pool should be better measured. However, despite mentioning that batch thickness can be derived from image processing, current methods remain silent on how to proceed in concrete terms. In particular, they provide no details or examples of calculations or image transformations to perform to obtain an accurate value.

[0022] Finally, past experiences in optimizing the energy consumption of electric furnaces have shown that portions of the batch that have reached their local steady-state temperature, i.e., that are in thermal equilibrium, have a direct impact on the energy to be supplied for efficient melting. Thus, real-time monitoring of the amount of batch in thermal equilibrium should help overcome the problems of underheating and overheating that Petition 870250094514, dated 10 / 16 / 2025, page 19 / 48 6 / 22 can often be found by furnace operators when setting up the electric furnace. However, current methods are unable to provide such information about the batch status inside the furnace.

[0023] Thus, there is still a need for an efficient and reliable method to estimate the thickness of parts of a batch of materials floating in the weld pool that has reached its steady-state temperature. Ideally, the method should also be able to provide the variation over time and / or the spatial distribution of the thickness over the weld pool. In addition, it should also be versatile enough to provide values ​​for other key parameters related to the stability of the batch over time, for example, the spatial distribution of hot spots or volcanoes and their movement over time. Solution to the technical problem

[0024] In a first aspect of the invention, a computer-implemented method is provided for measuring the thickness of a batch of materials floating in a melt pool in a glass melting furnace as described in claim 1, the dependent claims being advantageous embodiments.

[0025] In a second aspect of the invention, a data processing device, a computer program and a computer-readable means are provided for implementing the method of the first aspect.

[0026] In a third aspect of the invention, a process is provided for measuring the thickness of a batch of materials floating in a molten pool.

[0027] In a fourth aspect of the invention, a system is provided for measuring the thickness of a batch of materials floating in a molten pool. Advantages of the invention

[0028] A first notable advantage of the invention is that the thickness of any region of the batch of materials that are in thermal equilibrium is calculated with precision and accuracy. Overestimation or underestimation of thickness is avoided because the invention does not depend on batch measurements and / or melting level, or relationships between their densities. Thickness variations along the batch from local heterogeneities or convective effects within the weld pool are also better modeled. Petition 870250094514, dated 10 / 16 / 2025, page 20 / 48 7 / 22

[0029] A second advantage of the invention is that it can provide the variation over time and / or the spatial distribution of the batch thickness on the melt pool. Thus, it can be advantageously implemented for real-time monitoring of the batch thickness throughout its life cycle within the electric furnace.

[0030] A third advantage, which is a consequence of the first, is that the invention provides more reliable and representative information about the batch state in terms of thickness, so that the energy to be supplied to the electric furnace electrodes can be better adjusted for efficient melting, saving energy consumption and costs. In this context, the information provided as output data by the invention can be advantageously fed as input data to a control circuit, for example, to a feedback controller, for real-time adjustment of the energy to be supplied to the furnace electrodes.

[0031] A fourth advantage is that, in certain embodiments, the invention can also provide key parameters related to batch stability over time, for example, the spatial distribution of hot spots or volcanoes, their movements and the batch velocity vector field. Brief description of the drawings

[0032] Fig. 1 is a schematic representation of an example of a fiberglass or stone manufacturing line.

[0033] Fig. 2 is a schematic cross-section of an example of an electric glass or stone melting furnace.

[0034] Fig. 3 is a data flow diagram of a computer-implemented method according to the first aspect of the invention.

[0035] Fig. 4 is a physical data flow diagram of a data processing system for implementing a method according to the first aspect of the invention. Detailed description of the modalities

[0036] With reference to fig. 1, a fiberglass or stone manufacturing line 1000, although the internal centrifugation method may generally comprise silos 1001 for storing raw materials 1001a, for example, Petition 870250094514, dated 10 / 16 / 2025, page 21 / 48 8 / 22 mineral compounds and / or glass shards, an electric glass or stone melting furnace 1002 for melting raw materials 1001a that are transported from silos 1001 by means of conveyor 1003 and one or more fiber forming tools 1005a, 1005b, 1005c that are fed with molten glass or stone 1006 through a distribution channel 1004. The distribution channel is generally designed as an open or closed channel provided with openings just above each fiber forming tool 1005a, 1005b, 1005c for feeding each of them with molten glass or stone 1006.

[0037] With reference to Fig. 2, adapted from US 4 194 077 A, OWENS CORNING FIBERGLASS CORP [US], 18.03.1980, an electric glass or stone melting furnace 1002 comprises refractory walls 2001 and a refractory floor 2002 forming a tank 2003 which is divided into a melting part 2003a and a refining part 2003b by a partition wall 2001a. The two parts 2003a, 2003b communicate with each other by an opening 2004 located in the lower part of the partition wall 2001a at the level of the floor 2002.

[0038] Raw materials 1001a are conveyed to the melting section 2003a, forming a batch 2005 that floats on the surface of the pool 2006 and is progressively melted by a series of immersed electrodes 2007. Alternatively, or complementarily, the electrode series may comprise plunging-type electrodes as described in FR 2 599 734 A1, SAINT GOBAIN RECH [FR] 11.12.1987. As convection occurs within the melting pool 2006, the melt flows through the mouth 2004 of the melting section 2003a to the refining section 2003b which is connected to the distribution channel or feeders. One or two immersed electrodes 2008 may be located at floor level 2002 on one or each side of the mouth 2004.

[0039] With reference to Fig. 2 and Fig. 3, in a first aspect of the invention, a computer-implemented method 3000 is provided for measuring the thickness, and, of a batch 2005 of materials floating in a melt pool 2006 in an electric glass or stone melting furnace 1002; wherein the aforementioned method takes, as input data, a set I3000 of timescale temperature maps, M[T] from batch 2005 of materials, a range, R[Ts] of expected values ​​for the Petition 870250094514, dated 10 / 16 / 2025, p. 22 / 48 9 / 22 steady-state temperature, Ts, of batch 2005, and a limit value, σ, for the temperature variation, ΔT, over time, Δ1; wherein the aforementioned method provides, as output data, the spatial distribution O3000 of thicknesses of batch 2005 on the surface 2005a of the same; whereby the aforementioned method 3000 comprises the following steps: (a) select 3001, within each temperature map, M[T], of the provided set I3000, the regions for which the temperature is within the range R[Ts] of expected values ​​for the stationary temperature, Ts, of batch 2005; (b) compute 3002 the temperature variation, ΔT, within the same regions detected between successive temperature maps in a given time range, Δ1; (c) select 3003 regions from step (b) in which the temperature variation ΔT over time, Δ1, is below the threshold value, σ, provided as input; (d) compute 3004, for each point on the temperature maps within each of the regions selected in step (c), the thickness, E, of batch 2005, wherein said thickness, E, is calculated by applying, at said point, a function, E(Ts), defining a relationship between the thickness and the steady-state temperature, Ts, and wherein said function, E(Ts), is based on a simulated or experimental heat flux transfer, or on an empirical model.

[0040] In the context of the invention, a temperature map should be interpreted as its common definition in the field of cartography, for example, as a representation or distribution of temperatures on a surface, to scale or out of scale. In practice, it can be interpreted as a collection of temperature data in which each temperature data point is linked to a position on the surface of a batch of material. These positions can be coordinated in pixels or in units of length. Petition 870250094514, dated 10 / 16 / 2025, page 23 / 48 10 / 22

[0041] In the context of the invention, a set of temperature maps should be understood as a collection of temperature maps whose number may be arbitrary or depend on the technical limitations or configurations of the means, for example, camera, to acquire them.

[0042] In the context of the invention, a set of timescale temperature maps is a collection of temperature maps acquired over a period of time, i.e., a timescale. The timescale can be arbitrarily fixed depending on the means, for example, the camera used to acquire the maps and / or the time during which the temperature maps are considered acquired. It can also be a continuous, i.e., endless, timescale, for example, in the case of monitoring. The frequency with which the temperature maps are acquired can be a matter of choice depending on the technical limitations of the means, for example, the camera, used to acquire them and the timescale during which the batch is known or expected to evolve. As a general rule, the frequency should be defined so that the evolution, for example, movements, of certain characteristics, for example, hot and / or cold spots, can be observed in successively acquired temperature maps.

[0043] In the context of the invention, the term steady-state temperature applied to the batch should be interpreted as the temperature corresponding to the thermal equilibrium of batch 2005, that is, when the heat flow from the melt pool under the batch is balanced by the heat flow from furnace 1002 within the internal area 2009 of furnace 1002 above batch 2005. In thermal equilibrium, that is, when the batch has reached its steady-state temperature, Ts, the temperature of batch 2005 should be stable over time or vary very slightly over time.

[0044] The batch heat equation can be expressed by the following equation (1): PCpe~dt=ê (Tme 1 ;~T)~ [h(T^)+ σεΤΤ^4)]= 0

[0045] Where, ρ is the density of the batch, cp is the specific heat capacity of the batch, e is the thickness of the batch, T is the temperature of batch 2005, t is the time, λ is the thermal conductivity of batch 2005, σ is the Stefan-Boltzmann constant, h is the convective heat transfer coefficient of batch 2005, ε is the coefficient of Petition 870250094514, dated 10 / 16 / 2025, page 24 / 48 11 / 22 thermal emissivity of batch 2005, Tf is the temperature inside the internal area 2009 of furnace 1002 above batch 2005, Tmelt is the temperature of the melting pool 2006 and T is the temperature of batch 2005.

[0046] In the furnace region 1002 where the fresh, unheated batch is supplied by the feeder, conveyor or sprayer 1003 into the melt pool 2006, the batch 2005 may be colder than in regions further away from furnace 1002. Over time, the batch is progressively heated and its temperature increases to reach a steady state.

[0047] Conversely, in the vicinity of hotspots or volcanoes, the temperature of batch 2005 may increase and become higher than its steady-state temperature due to the local imbalance between the heat flux from the melt pool 2006 under batch 2005 and the heat flux from the internal area of ​​furnace 1002 above batch 2005.

[0048] Between these extremes, for example, far from the feeding zone and hot spots or volcanoes, the 2005 plot is expected to be in thermal equilibrium and to have reached its steady-state temperature.

[0049] In addition to the problems mentioned above regarding the state of the art, the present invention provides an intelligent solution to this issue.

[0050] In step (a), the temperature map regions that are within the range, R[Ts], of the expected values ​​for the steady-state temperature provided as input are selected. This operation allows identifying the regions of batch 2005 that may be likely to be in thermal equilibrium, i.e., those regions for which the probability of having reached their steady-state temperature is the highest.

[0051] The range, R[Ts], of expected values ​​for the steady-state temperature can be determined empirically and / or through processing the temperature maps provided as input.

[0052] In advantageous embodiments, the range, R[Ts], of expected values ​​for the steady-state temperature, Ts, of batch 2005 may be a range of values ​​determined experimentally in similar batches of materials under similar melting conditions. For example, the values ​​may come from historical point measurements and / or monitored temperature data in certain regions of the batch surface. Petition 870250094514, dated 10 / 16 / 2025, page 25 / 48 12 / 22 acquired by operators and / or automatic sensors, for example, infrared sensors, from previous production tests or campaigns carried out in a given electric glass or stone melting furnace 1002. A notable advantage of this approach is that the range, R[Ts], of expected values ​​for the steady-state temperature can be more representative of the thermal behavior of the electric furnace and the composition of the batch under consideration.

[0053] Surprisingly, it has been found in the context of the invention that a range, R[Ts], of the expected values ​​for the stationary temperature between 20°C and 200°C, preferably between 70°C and 140°C, may be suitable for most cases of electric glass or stone melting furnaces.

[0054] Alternatively, or complementarily, the range, R[Ts], of expected values ​​for the steady-state temperature, Ts, of batch 2005 can be a defined range of values ​​between the two modes of a bimodal distribution calculated from one or a plurality of the temperature maps of the supplied set I3000. The calculation of the bimodal distribution of the temperature maps can allow discrimination between the two extremes which are the coldest regions of the batch, for example, the regions near the feeder, and the hottest regions, for example, the hotspots or volcanoes. Fixing a range of steady-state temperatures between these extremes can be interpreted as a kind of filter to select the regions with the highest probabilities of being in thermal equilibrium. Methods such as those described in US 4,409,012 A, OWENS ILLINOIS INC [US], 11.10.1983 can be adapted to temperature maps to calculate the bimodal distribution.

[0055] In step (c), the regions selected in step (b) are selected for which the temperature variation ΔT over time, Δ1, is below the limit value, σ, provided as input. If the regions of batch 2005 had reached their steady-state temperature, their temperature should remain stable over time or vary slightly, i.e., ΔT / Δ1 ≠ 0. Thus, this operation allows identifying the batch regions that are effectively in thermal equilibrium, i.e., those regions that have effectively reached their steady-state temperature, Ts. In other words, step (c) further refines the selection of regions in thermal equilibrium that was initiated in step (a). Petition 870250094514, dated 10 / 16 / 2025, page 26 / 48 13 / 22

[0056] Surprisingly, in the context of the invention, it has been found that the threshold value, σ, can be low for most cases of electric glass or stone melting furnaces. Consequently, in preferred embodiments, the limit value, σ, for the temperature change ΔT over time, Δ1, can be 5°C / min, preferably 2°C / min, more preferably 1°C / min.

[0057] Since the efficiency and accuracy of the invention for measuring the thickness, E, of batch 2005 depend on the effective and accurate selection of the batch regions in thermal equilibrium, the combination of steps (a) to (c) should be considered as one of the main features of method 3000 according to the invention and, consequently, as a notable difference from the practices described in the art.

[0058] In step (d), the thickness, E, of batch 2005 is calculated for the regions selected in step (c) according to a function that defines a relationship between the thickness, E, and the stationary temperature, Ts. There is no exclusive function. Any adapted function, empirical or theoretical, that may be able to link the batch thickness to its local stationary temperature may be used.

[0059] As already emphasized, an advantage of the invention is to allow an effective and precise selection of the 2005 batch regions in thermal equilibrium, i.e., regions that have locally reached their steady-state temperature. Thus, the batch thickness can be advantageously derived from the aforementioned thermal equilibrium equation (1), provided that the thermal properties of the batch can be determined.

[0060] Therefore, in certain embodiments, the method may also take, as input data, the temperature, Tf, within the internal area of ​​furnace 1002 above batch 2005, the temperature, Tmelt, of the melting pool 2006; where the function E(Ts) is a heat flux transfer function given by the following equation (2): / •ΎT) =------λ(ΤηιβΙt—T))-----1 Jh(Ts- Tf) + σε(Ts4- Tfi

[0061] where λ is the thermal conductivity of batch 2005, σ is the Stefan-Boltzmann constant, h is the convective heat transfer coefficient of the batch. Petition 870250094514, dated 10 / 16 / 2025, page 27 / 48 14 / 22 2005, ε is the thermal emissivity coefficient of batch 2005, Tf is the temperature inside the furnace inner area above batch 2005, Tmelt is the temperature of the molten pool 2006, and Ts is the steady-state temperature of batch 2005 for which the thickness, E, must be calculated.

[0062] The parameters λ, h and ε can be theoretically calculated knowing the convection effects within the internal area 2009 of furnace 1002, the chemical composition of batch 2005 and, as the batch is generally supplied as a granular material, its apparent density. However, as already emphasized, during the service life of batch 2005 in the furnace, the apparent density, the granular distribution and also the chemical composition of the batch can vary locally within its thickness and within its entire coverage, depending on the temperature gradients and convection movements that occur within the melt pool 2006 which locally determine the melting state and the melting rate of batch 2005. A direct consequence is that the theoretically calculated parameters may not be well representative of the local properties of batch 2005 and may induce inaccuracies in the determination of the local thickness within the batch.Despite these disadvantages, the theoretically calculated parameters can still be relevant as an approximate proxy for the batch whose thermal and physical properties should not vary excessively during its service life inside the furnace.

[0063] Alternatively, or complementarily, the parameters λ, h and ε can be determined experimentally by performing measurements on similar batches evolving under similar melting conditions. For example, experiments can be conducted in pilot electric furnaces to mimic the behavior within large-scale production furnaces, and the parameters λ, h and ε calculated from in situ and / or ex situ measurements of the thermal and physical properties of the batches. The parameters thus determined can then be more representative of the actual behavior of a batch in an electric furnace of a production line.

[0064] However, the timescale and rate of change in manufacturing lines may not always be compatible with those of experiments that can be conducted in laboratory or pilot furnaces. Furthermore, Petition 870250094514, dated 10 / 16 / 2025, p. 28 / 48 15 / 22 Scale effects can occur, so what is calculated and / or measured in laboratories or pilot plants may not work efficiently when applied to large-scale electric furnaces. Therefore, it may be more valuable to rely on empirical data to define a relationship between thickness and steady-state batch temperature.

[0065] In this context, in preferred embodiments, the function E(Ts) can be an empirical model given by the following equation (3): Ts-T\ ( (Ts) = e , —— ( e2-e ) ) '2 ' 1

[0066] where T1 and T2 are two experimentally measured stationary temperatures of batch 2005, e1 and e2 are two experimentally measured thicknesses of batch 2005 corresponding to temperatures T1 and T2 respectively, Ts is the stationary temperature of batch 2005 for which the thickness, E, must be calculated.

[0067] Surprisingly, equation (3) above has been found to be a simple, elegant and robust way to calculate the thickness of batch 2005 corresponding to the selected regions of the temperature maps with a high level of accuracy without relying on prior knowledge of any physical and / or thermal properties of the batch. Furthermore, the two temperatures, T1 and T2, and the two thicknesses, e1 and e2, can be measured experimentally on a laboratory scale or in pilot furnaces and / or industrial furnaces. In the latter case, results that are most representative of what is actually occurring in the industrial furnace can be obtained. As a further consequence of not relying on prior knowledge of any physical and / or thermal properties of batch 2005, equation (3) can be easily implemented industrially.

[0068] The method according to the invention can be adapted to provide other key parameters for batch stability over time, for example, the spatial distribution of hot spots or volcanoes and their movement over time.

[0069] In certain embodiments, the method may also include a region detection step (e), within each temperature map of the set. Petition 870250094514, dated 10 / 16 / 2025, p. 29 / 48 16 / 22 provided i3000, applying an object detection function to the aforementioned temperature map, wherein the object detection function is configured to process regions with temperature equal to or greater than a threshold value, θ; wherein the method also provides, as output data, the spatial distribution of the detected regions over time.

[0070] To detect hot spots or volcanoes, the threshold value, θ, can be fixed at the temperature at which said hot spots or volcanoes are expected or observed within the batch 2005 layer. This temperature generally varies depending on several parameters, such as the batch melting temperature, temperature gradients and convection movements within the melt pool 2006, the energy supplied to furnace 1002, the temperature within the furnace's internal area above batch 2005...

[0071] The limit value, θ, can be determined experimentally by internal inspection of furnace 1002 using temperature sensors, for example, thermocouples, infrared sensors. It can also be determined automatically by calculating the statistical distribution within the temperature maps and determining, from this distribution, the cutoff temperature above which the temperature corresponds to hot spots or volcanoes.

[0072] Since hot spots or volcanoes mainly correspond to holes within the 2005 plot through which the 2006 melt pool can rise or be seen, and the 2006 melt pool often shows the highest temperature, the limit value, θ, can be set so as to represent a relatively small difference from the highest temperature detected in the temperature maps.

[0073] The temperature map regions that are detected through the object detection function according to the above modalities can provide the spatial distribution of hotspots or volcanoes on the 2005 batch layer. Other indicators, such as their number, their density, i.e., number per surface area, their distribution in sizes, can also be derived.

[0074] The bubble detection function and a threshold value, θ, can be implemented through different image processing algorithms. Petition 870250094514, dated 10 / 16 / 2025, pp. 30 / 48 17 / 22 Therefore, in certain embodiments, the object detection function can be selected from among the Otsu threshold function, the Laplacian Gauss function, the Hessian determinant function, the Gauss difference method, and a watershed-based gray-level bubble detection function.

[0075] The technique provides different implementations of these algorithms, for example, the python scikit-image package.

[0076] On the surface of melt pool 2006, due to convection movements occurring within it, batch 2005 may not follow a strictly linear path from feeder 1003 to the partition wall 2001a that separates the melting section 2003a and the refining section 2003b of furnace 1002. Instead, part of batch 2005 may move away from the overall flow direction of melt pool 2006 to the refining section 2003b and may follow a more complex, curvilinear path. In this context, the displacement of batch 2005 can be best represented by a velocity vector field pointing in various directions, but on average towards the flow direction of melt pool 2006.

[0077] For better adjustment of furnace parameters, for example, batch feed rate, energy distribution between electrodes 2007, 2008, it may be valuable to monitor the batch velocity vector field 2005. In this scope, in complementary embodiments, method 3000 may further comprise a step (f) of calculating the velocity of the regions detected in step (e) by calculating the displacement of said regions over time on the timescale of the set of timescale temperature maps. For example, by comparing successive temperature maps in which the same regions are detected, the displacement of said detected regions can be calculated and their velocity calculated by dividing the calculated displacement by the time interval between successive temperature maps. This operation can be repeated for the entire set of timescale temperature maps.

[0078] In a second aspect of the disclosure, with reference to Fig. 4, a data processing system 4000 is provided comprising means 4001 for carrying out a method 4000 according to any of the embodiments of the first aspect of the invention. A computer program I4001 is also provided comprising instructions which, when executed by a Petition 870250094514, dated 10 / 16 / 2025, p. 31 / 48 18 / 22 computer, cause the computer to execute a method 4000 according to any of the embodiments of the first aspect of the invention.

[0079] The data processing system 4000 comprises means 4001 for carrying out a method according to any of the embodiments of the first aspect of the invention. An example of means 4001 could be a device that can be instructed to perform sequences of arithmetic or logical operations automatically to perform tasks or actions. Such a device, also called a computer, may comprise one or more Central Processing Units (CPUs) and at least one controller device that are adapted to perform these operations.

[0080] It may also include other electronic components, such as input / output interfaces 4003, non-volatile or volatile storage devices 4002 and buses which are communication systems for transferring data between components within a computer or between computers. One of the input / output devices may be the user interface for human-machine interaction, for example, a graphical user interface to display information understandable by humans.

[0081] Since computation can require a great deal of computational power to process substantial amounts of data, the data processing system may advantageously comprise one or more Graphics Processing Units (GPUs) whose parallel structure makes them more efficient than the CPU, particularly for image processing.

[0082] The I4001 computer program can be written using any type of programming language, compiled or interpreted, to implement the steps of the method according to any embodiments of the first aspect of the invention. The I4001 computer program can be part of a software solution, i.e., part of a collection of executable instructions, code, scripts or the like and / or databases.

[0083] In certain embodiments, storage or computer-readable media may also be provided 4002 comprising instructions which, when executed by a computer, cause the computer Petition 870250094514, dated 10 / 16 / 2025, pp. 32 / 48 19 / 22 execute the method according to any of the embodiments of the first aspect of the invention.

[0084] Computer-readable storage 4002 may preferably be non-transient non-volatile storage or memory, for example, a hard disk drive or a solid-state drive. Computer-readable storage may be removable storage media or non-removable storage media as part of a computer.

[0085] Alternatively, computer-readable storage can be volatile memory within a removable medium.

[0086] The 4002 computer-readable storage may be part of a computer used as a server from which executable instructions may be downloaded and, when executed by a computer, cause the computer to execute a method in accordance with any of the embodiments described in this document.

[0087] Alternatively, the I4001 program can be implemented in a distributed computing environment, for example, cloud computing. The instructions can be executed on the server to which client computers can connect and provide encoded data as inputs to the method. Once the data is processed, the output can be downloaded and decoded on the client computer or sent directly, for example, as instructions. This type of implementation can be advantageous as it can be performed in a distributed computing environment, such as a cloud computing solution.

[0088] In a third aspect of the invention, a process is provided for measuring the thickness of a batch 2005 of materials 1001a floating in a melt pool 2006 in an electric glass or stone melting furnace 1002, wherein said process comprises the following steps: - acquire a set of I3000 timescale temperature maps, M[T], of the surface of batch 2005 of materials 1001a within the internal area 2009 of the melting furnace 1002; - to implement, by means of a data processing device 4000, the method 3000 in accordance with any of the Petition 870250094514, dated 10 / 16 / 2025, pp. 33 / 48 20 / 22 embodiments of the second aspect of the invention, wherein the acquired set I3000 of time-scale temperature maps M[T] is provided as input data for said method 3000.

[0089] For an implementation of the process, in a fourth aspect of the invention, with reference to Fig. 2, a system is provided for measuring the thickness of a batch 2005 of materials 1001a floating in a melt pool 2006 in an electric glass or stone melting furnace 1002, wherein said system comprises: - an acquisition device 2010 configured to acquire an i3000 set of timescale temperature maps, M[T], of the surface of batch 2005 of materials 1001a within the internal area 2009 of furnace 1002; - a data processing device 4000 according to any of the embodiments of the second aspect of the invention, wherein said data processing device 4000 and the acquisition device 2010 are connected wired or wirelessly to each other to transfer data.

[0090] In preferred embodiments, the acquisition device 2010 may comprise an infrared camera configured to acquire a set of timescale infrared maps, M[T] and said infrared camera, or the data processing device 4000 comprises means for converting the set of timescale infrared maps into a set of timescale temperature maps.

[0091] Examples of acquisition devices comprising an infrared camera that are adapted to acquire infrared maps in a glass or stone melting furnace are described in the art, for example, JPS 5 339 204 A, NIPPON ELECTRON OPTICS LAB 11.04.1978, JPH 0 7216 422 A, NIPPON STEEL CORP 15.08.1995, JP 2010 002 150 A, TAKUMA CO LTD 07.01.2010 and US 2018 231875 A1, HER MAJESTY THE QUEEN IN RIGHT OF CANADA AS REPRESENTED BY THE MINISTRY OF NATURAL RESOURCES [CA] 16.08.2018. Petition 870250094514, dated 10 / 16 / 2025, pp. 34 / 48 21 / 22

[0092] An example of an infrared camera might be an OPTRIS® PI 400 infrared camera which is marketed by the company OPTRIS® Infrared Sensing. The camera has an optical resolution of 382x288 pixels and an FPA 25 pm x 25 pm IR sensor in the spectral range of 7.5 pm - 13 pm. The detected temperature range is 0 - 250°C or 150°C - 900°C.

[0093] The acquisition device 2010 may be provided with heat and dust shields in the form of protective windows and / or gas shields located in front of the objective. In preferred embodiments, the acquisition device may be an infrared camera in front of which is placed an enlarged corolla adapted to be placed in front of an opening in a wall 2001 of the electric glass or stone melting furnace 1002. The corolla may comprise on its periphery an injection nozzle for injecting inert gas, for example, nitrogen towards the objective, so as to purge any dust and evacuate heat from the furnace. Between the corolla and the camera lens, a non-removable infrared transparent window may also isolate the camera from the furnace atmosphere.

[0094] The means for converting infrared maps into temperature maps depend on the inference of blackbody or graybody radiation.They are extensively described in the technical specifications, for example, in the IEC 62492-1 TS Technical Specification: Industrial process control devices - Radiation thermometers - Part 1: Technical data for radiation thermometers, and in the ASTM-E1256 Standard Test Methods for Radiation Thermometers (Single Waveband Type).

[0095] Depending on the location of the acquisition device 2010 relative to the surface of the batch 2005 of materials 1001a, the temperature or infrared maps may suffer from perspective distortions. In this context, in certain embodiments, the data processing device 4000 may further comprise means for correcting the perspective view of the acquisition device by applying a perspective or homographic transformation function to each temperature map of the i3000 set of timescale temperature maps M[T].

[0096] Examples of perspective or homographic transformation functions are provided in the OPENCV python library. They can also be Petition 870250094514, dated 10 / 16 / 2025, pp. 35 / 48 22 / 22 implemented as part of the computer-implemented method 3000 according to the first aspect of the invention.

[0097] The invention, in all its aspects, can be applied to, but is not limited to, many processes for the manufacture of glass products, for example, glass wool, rock wool or stone wool, textile glass yarn, flat glass or hollow glass. List of Citations Patent literature

[0098] JPS 5 339 204 A, NIPPON ELECTRON OPTICS LAB 11.04.1978.

[0099] US 4 194 077 A, OWENS CORNING FIBERGLASS CORP [US], March 18, 1980.

[0100] WO 8002833 A1, OWENS CORNING FIBERGLASS CORP [US], December 24, 1980.

[0101] US 4 409 012 A, OWENS ILLINOIS INC [US], 10 / 11 / 1983.

[0102] JPH 0 656 432 A, NIPPON ELECTRIC GLASS CO, 01.03.1994.

[0103] JPH 0 7216 422 A, NIPPON STEEL CORP 15.08.1995.

[0104] WO 0248057 A1, SOFTWARE & TECH GLAS GMBH [DE], 20.06.2002.

[0105] EP 1 655 570 A1, FRANKE MATTHIAS [DE], 10.05.2006.

[0106] JP 2010 002 150 A, TAKUMA CO LTD 01.07.2010.

[0107] WO 2018 / 104695 A1, LAND INSTRUMENTS INTERNATIONAL LTD [GB], 14.06.2018.

[0108] US 2018 231875 A1, HER MAJESTY THE QUEEN IN RIGHT OF CANADA AS REPRESENTED BY THE MINI OF NATURAL RESOURCES [CA] 16.08.2018.

[0109] WO 2022 / 242843 A1, GLASS SERVICE A S [CZ], 24.11.2022. Literatura não patentária

[0110] Especificação Técnica IEC 62492-1 TS: Industrial process control devices - Radiation thermometers - Part 1: Technical data for radiation thermometers.

[0111] ASTM-E1256 Standard Test Methods for Radiation Thermometers (Single Waveband Type). Petition 870250094514, dated 10 / 16 / 2025, pp. 36 / 48

Claims

1 / 4 CLAIMS 1. Computer-implemented method (3000) for measuring the thickness, e, of a batch (2005) of materials floating in a melt pool (2006) in an electric glass or stone melting furnace (1002); characterized in that said method takes, as input data, a set (I3000) of time-scale temperature maps, M[T] of the batch (2005) of materials, a range, R[Ts] of expected values ​​for the steady-state temperature, Ts, of batch 2005, and a limit value, σ, for the temperature variation, ΔT, over time, Δ1; wherein said method provides, as output data, the spatial distribution (O3000) of thicknesses of the batch (2005) on the surface thereof;wherein the said method (3000) comprises the following steps: (a) selecting (3001), within each temperature map, M[T], of the given set I3000, the regions for which the temperature is within the range R[Ts] of expected values ​​for the stationary temperature, Ts, of the batch (2005); (b) computing (3002) the temperature variation, ΔT, within the same regions detected between successive temperature maps in a given time range, Δ1; (c) selecting (3003) the regions from step (b) in which the temperature variation ΔT over time, Δ1, is below the threshold value, σ, given as input;(d) compute (3004), for each point on the temperature maps within each of the regions selected in step (c), the thickness, E, of batch 2005, wherein said thickness, E, is calculated by applying, at said point, a function, E(Ts), defining a relationship between the thickness and the stationary temperature, Ts, and wherein said function, E(Ts), is based on a simulated or experimental heat flux transfer, or on an empirical model.; 2. Method (3000), according to claim 1, characterized in that said method also takes, as input data, the temperature, Tf, within the internal area of ​​furnace 1002 above batch (2005), the temperature, Tmelt, of the melt pool (2006); wherein the function E(Ts) is a transfer function Petition 870250094514, of 10 / 16 / 2025, p. 44 / 48 2 / 4 of heat flow provided by the following equation p — Tg)______ ( J h(Ts - Tf) + σε( Ts4 - Tfi where λ is the thermal conductivity of batch 2005, σ is the Stefan-Boltzmann constant, h is the convective heat transfer coefficient of batch 2005, ε is the thermal emissivity coefficient of batch 2005, T f is the temperature inside the furnace inner area above batch 2005, Tmelt is the temperature of the melt pool 2006 and Ts is the steady-state temperature of batch 2005 for which the thickness, E, must be calculated.

3. Method (3000), according to claim 1, characterized in that the function E(Ts) is an empirical model provided by the following equation: Ts -Ί\ E(TS)= + z?—zr (e2 — ex) '2 ' 1 where T1 and T2 are two stationary temperatures measured experimentally from batch 2005, e1 and e2 are two thicknesses measured experimentally from batch 2005 corresponding to temperatures T1 and T2 respectively, Ts is the stationary temperature of batch 2005 for which the thickness, E, is to be calculated.

4. Method (3000), according to any one of claims 1 to 2, characterized in that the limiting value, σ, for the temperature variation ΔT over time, Δ1, is 5°C / min, preferably 2°C / min, more preferably 1°C / min.

5. Method (3000), according to any one of claims 1 to 4, characterized in that the range, R[Ts], of expected values ​​for the stationary temperature, Ts, of batch 2005 is a range of values ​​determined experimentally in similar batches of materials under similar melting conditions or a defined range of values ​​between the two modes of a bimodal distribution calculated from one or a plurality of the temperature maps of the given set (i3000).

6. Method (3000), according to any one of claims 1 to 5, characterized in that the method further comprises a step (e) of region detection, within each temperature map of the set provided Petition 870250094514, 10 / 16 / 2025, p. 45 / 48 3 / 4 i3000, applying an object detection function to said temperature map, wherein the object detection function is configured to process regions with temperature equal to or greater than a threshold value, θ; wherein the method also provides, as output data, the spatial distribution of the detected regions over time.

7. Method (3000), according to claim 6, characterized in that the method comprises a step (f) of calculating the velocity of the regions detected in step (e) by calculating the displacement of said regions over time on the timescale of the set of timescale temperature maps.

8. Method (3000), according to any one of claims 6 to 7, characterized in that the object detection function is selected from the Otsu threshold function, the Gauss Laplacian function, the Hessian Determinant function, the Gauss Difference method and a watershed-based gray-level bubble detection function.

9. Data processing device (4000) characterized in that it comprises means for carrying out a method as defined in any one of claims 1 to 8.

10. Computer-readable medium (4002) characterized in that it comprises instructions which, when the program is executed by a computer, cause the computer to execute a method as defined in any one of claims 1 to 8.

11. Process for measuring the thickness of a batch (2005) of materials (1001a) floating in a melt pool (2006) in an electric glass or stone melting furnace (1002), characterized in that said process comprises the following steps: - acquiring a set (i3000) of time-scale temperature maps, M[T], of the surface of the batch (2005) of materials (1001a) within the internal area (2009) of the melting furnace (1002); - implementing, by means of a data processing device (4000), the method (3000) according to any of claims 1 to 8, wherein Petition 870250094514, dated 10 / 16 / 2025, p. 46 / 48 4 / 4 acquired set (i3000) of time scale temperature maps M[T] is provided as input data for the said method (3000).

12. System for measuring the thickness of a batch (2005) of materials (1001a) floating in a melt pool (2006) in an electric glass or stone melting furnace (1002), characterized in that said system comprises: - an acquisition device (2010) configured to acquire a set (i3000) of time-scale temperature maps, M[T], of the surface of the batch 2005 of materials (1001a) within the internal area (2009) of the furnace (1002); - a data processing device (4000) as defined in any one of claims 1 to 8, wherein said data processing device (4000) and the acquisition device (2010) are wired or wirelessly connected to each other to transfer data.

13. System according to claim 13, characterized in that the acquisition device (2010) comprises an infrared camera configured to acquire a set of timescale infrared maps, M[T] and said infrared camera, or the data processing device (4000) comprises means for converting the set of timescale infrared maps into a set i3000 of timescale temperature maps.

14. System, according to any one of claims 13 to 14, characterized in that the data processing device (4000) further comprises means for correcting the perspective view of the acquisition device by applying a perspective or homographic transformation function to each temperature map of the set (i3000) of timescale M[T] temperature maps. Petition 870250094514, dated 10 / 16 / 2025, pp. 47 / 48

Images