Batch-type electric arc furnace system

Abstract

Provided are methods and systems for determining respective masses associated with respective portions of respective layers of scrap metal material deposited in a feed bucket associated with a batch-feed electric arc furnace (EAF), the methods and systems determining respective masses associated with respective portions of respective layers of scrap metal material based on (a) respective volumes of the respective portions of the respective layers of scrap metal material and (b) respective assigned densities assigned by a machine learning classification model based on digital images of the respective portions of the respective layers of scrap metal material.

Description

[Technical field]

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS This application claims priority under 35 USC § 119 from U.S. patent application Ser. No. 63 / 318,505, filed March 10, 2022, the disclosure of which is expressly incorporated by reference in its entirety herein.

[0002] Inventive embodiments of the present disclosure relate generally to methods and systems for determining respective masses associated with respective portions of respective layers of scrap metal material deposited in a feed bucket associated with a batch-feed electric arc furnace (EAF), as well as methods and systems related to operating a batch-feed EAF system. [Background technology]

[0003] The traditional method of steel production is based on the blast furnace (BOF) route, where coke is used to reduce and melt the iron ore in a blast furnace, followed by smelting in a BOF converter. The coke is burned to generate energy and the iron ore is reduced by reacting with carbon and carbon monoxide. In contrast, in the last few years, the electric arc furnace (EAF) steelmaking process has been gaining increasing interest worldwide over BOF technology.

[0004] The benefits of the EAF process are numerous, but one of the leading arguments for EAF, especially next to lower emissions, is flexibility. The process is much more flexible than the BOF route, which cannot be turned on and off at will depending on current market conditions. The trend towards more advanced, more flexible and cleaner processes calls for technologies that can be rapidly implemented and adapted to each steelmaker's individual needs.

[0005] Batch-feed EAF is currently the most common method for recycling steel from scrap metal (e.g. scrap iron). Scrap iron is very diverse in both composition and form. By melting the scrap iron in a furnace using electrodes and electric current, it is possible to produce new, functioning steel from old products. Instead of investing in raw material resources, basic steel elements and useful alloys can be reused, which is beneficial from both an economic and environmental perspective.

[0006] Depending on the grade of steel being produced, the scrap mix composition (e.g., a mix of different types or classes of scrap iron fed into the EAF) must meet a specific range or target of chemical composition. Typically, each production shift is scheduled in advance, with scrap yard operators preparing the scrap iron feed buckets according to the needs of the production program. Thus, preparing the feed buckets is a critical operation not only to ensure proper melt chemistry, but also to ensure good melting conditions in the EAF. The scrap metal materials must be stacked in layers in the feed bucket according to their size and density to promote rapid formation of a liquid pool of steel in the EAF hearth and to protect the hearth sidewalls and roof from electrical arc radiation. Another consideration is to minimize the collapse of the scrap metal materials in the EAF, which could destroy the electrodes. This ensures that large and heavy pieces of scrap metal are not placed directly in front of the burner ports, which would cause the flames to blow back onto the water-cooled panels. In some cases, the batched scrap metal material may also contain lime and carbon, or may be injected into the EAF during heats. At this point, the EAF operates as a batch melting process that produces batches of steel, called "heats."

[0007] Thus, there remains a need in the art of batch-feed EAFs for reliable methods and / or systems for identifying batch profiles (e.g., mass, volume, chemical composition, etc.) directed to a batch-feed EAF that can facilitate safe and efficient operation of the batch-feed EAF. Summary of the Invention [Problem to be solved by the invention]

[0008] The present invention has been made to solve the problems in the conventional techniques described above. [Means for solving the problem]

[0009] Some embodiments according to the present invention provide a method for identifying a batch profile for a batch-feed electric arc furnace (EAF) including one or more input buckets. According to some embodiments of the present invention, the method includes the steps of: (i) acquiring respective red-green-blue (RGB) digital images of respective portions of respective layers (e.g., one or more individual layers stacked on top of one another) of scrap metal material deposited in a first input bucket; (ii) generating, in a computer processor, respective stereo camera depth images of the respective portions of the respective layers of scrap metal material, the computer processor correlating, in the computer processor, one or more pixels of the respective depth images with respective 3D coordinates of the respective portions of the respective layers of scrap metal material; and (iii) generating, in the computer processor, respective 3D stereo camera depth images of the respective portions of the respective layers of scrap metal material. (v) identifying, by a machine learning model in the computer processor, a classification of each of the portions of each of the layers of scrap metal from among a group of known scrap metal materials based on at least a portion of the respective RGB digital images and assigning a respective density to each of the portions of each of the layers of scrap metal; and (vi) identifying a respective mass associated with each of the portions of each of the layers of scrap metal based on (a) the respective volumes of the respective portions of the respective layers of scrap metal, and (b) the respective assigned densities.

[0010] In another aspect, some embodiments according to the present invention provide a method of operating a batch-feed electric arc furnace (EAF) including one or more input buckets. The method may include identifying an actual batch profile defined by scrap metal material contained in the one or more input buckets. Identifying the actual batch profile may include (i) identifying respective masses associated with respective portions of respective layers of scrap metal material based on (a) respective volumes of respective portions of respective layers of scrap metal material and (b) respective assigned densities as described and disclosed herein, and / or (ii) identifying a total mass of the scrap metal material in the one or more input buckets as described and disclosed herein, and / or a total volume of the scrap metal material in the one or more input buckets as described and disclosed herein, and / or (iii) identifying a total chemical composition of the scrap metal material in the one or more input buckets as described and disclosed herein. The method may further include, in a computer processor, comparing the actual batch profile to operating conditions and / or operating set points of the EAF and modifying the operating conditions and / or operating set points based on the actual batch profile.

[0011] In another aspect, some embodiments according to the present invention provide a system for identifying a batch profile for a batch-feed electric arc furnace (EAF) including one or more feed buckets. The system may include a sensor fusion module, which includes two monochrome cameras configured to provide stereoscopic vision, an RGB camera, an infrared spectrum projector, and a first computer processor in operative communication with the two monochrome cameras, the RGB camera, and the infrared spectrum projector. The first computer processor may be configured to (a) generate respective stereoscopic depth images and correlate one or more pixels of the respective depth images with respective 3D coordinates of respective portions of respective layers of scrap metal material, and (c) identify respective volumes of respective portions of respective layers of scrap metal material based at least in part on the respective 3D coordinates of respective portions of respective layers of scrap metal material. The system may also include a scrap classification module, which is in operative communication with the sensor fusion module and includes a machine learning model and a second computer processor. The second computer processor may be configured to: (a) identify a classification of each portion of each layer of the scrap metal material from among a group of known scrap metal materials based on at least a respective portion of the respective RGB digital images received from the sensor fusion module; (b) assign a respective density to each portion of each layer of the scrap metal material; and (c) determine a respective mass associated with each portion of each layer of the scrap metal material based on (a) a respective volume of the each portion of the respective layer of the scrap metal material and (b) the respective assigned density.

[0012] In yet another aspect, some embodiments according to the present invention provide an electric arc furnace (EAF) system. The system may include an EAF including a furnace body defining an interior portion and a set of electrodes extending into the interior portion. The system may also include one or more input buckets. According to some embodiments of the present invention, the system may include a sensor fusion module mounted above and proximate to the one or more input buckets, the sensor fusion module including two monochrome cameras configured to provide stereoscopic vision, an RGB camera, an infrared spectrum projector, and a first computer processor in operative communication with the two monochrome cameras, the RGB camera, and the infrared spectrum projector. The first computer processor may be configured to (a) generate respective stereoscopic depth images and correlate one or more pixels of the respective depth images with respective 3D coordinates of respective portions of respective layers of scrap metal material, and (c) determine respective volumes of respective portions of respective layers of scrap metal material based at least in part on the respective 3D coordinates of respective portions of respective layers of scrap metal material. The system may also include a scrap classification module in operative communication with the sensor fusion module and including a machine learning model and a second computer processor configured to: (a) identify a classification of each portion of each layer of scrap metal material from among a group of known scrap metal materials based on at least a respective portion of each RGB digital image received from the sensor fusion module; (b) assign a respective density to each portion of each layer of scrap metal material; and (c) determine a respective mass associated with each portion of each layer of scrap metal material based on (a) a respective volume of the each portion of the each layer of scrap metal material and (b) the respective assigned density.

[0013] The present invention will now be described more fully hereinafter with reference to the accompanying drawings, in which some, but not all, embodiments of the invention are shown. Indeed, the invention may be embodied in many different forms and should not be construed as limited to the embodiments set forth herein; rather, these embodiments are provided so that this disclosure will satisfy applicable legal requirements. Like reference characters refer to like elements throughout. [Brief description of the drawings]

[0014] [Figure 1] 1 illustrates a scrap bay layout for an EAF system, according to some embodiments of the present invention. [Figure 2A] ~ [Figure 2D] 1 is an image of a scrap bin containing various types or classes of scrap metal material, according to some embodiments of the present invention. [Diagram 3] 1 illustrates an input bucket containing multiple layers of various types or classes of scrap metal material, according to some embodiments of the present invention. [Figure 4] 1 shows an input bucket containing multiple layers of various types or classes of scrap metal material, including oversized pieces of scrap material that can be detected and cause an alarm to be triggered according to some embodiments of the present invention. [Diagram 5] 1 illustrates an input schedule for three buckets that make up a batch input into an EAF according to some embodiments of the present invention. [Figure 6] 1 illustrates a three-dimensional (3D) or depth camera that may be used in accordance with some embodiments of the present invention. [Figure 7] The stereoscopically generated depth image is shown on the left and the corresponding RGB digital image on the right. [Figure 8] 2 illustrates a system functionality flow chart according to some embodiments of the present invention. [Figure 9] 1 illustrates a schematic of a system for operating an EAF system according to some embodiments of the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0015] In the following, embodiments of the present invention will be described in more detail with reference to the accompanying drawings, in which some, but not all, embodiments of the present invention are shown. Indeed, the present invention may be embodied in many different forms and should not be construed as being limited to the embodiments set forth herein, but rather, these embodiments are provided so that this disclosure will satisfy applicable legal requirements. As used in this specification and the appended claims, the singular forms "a," "an," and "the" include plural referents unless the context clearly dictates otherwise. As used herein, the term "and / or" includes any and all combinations of one or more of the associated listed items.

[0016] Embodiments of the present invention generally relate to methods and systems for measuring and / or determining the mass, volume, and / or chemical composition of scrap metal material contained in one or more input buckets, which may include one or more layers of different types or classes of scrap metal material. In this regard, a single input bucket or each input bucket may include a layered mixture of multiple types or classes of scrap metal material based on a proposed input schedule. In this regard, methods and systems according to some embodiments of the present invention are capable of measuring and / or determining the mass, volume, and / or chemical composition of each and every layer of scrap metal material deposited in one or more input buckets. Thus, according to some embodiments of the present invention, the mass, volume, and / or chemical composition of the aggregate (e.g., scrap mixture) for each input bucket may be determined. Additionally or alternatively, the methods and systems may determine an actual batch profile defined by the scrap metal material contained in the one or more input buckets. That is, by identifying and using the mass, volume, and / or chemical composition of the aggregates (e.g., scrap mixture) for each input bucket, it is possible to identify an overall batch profile (e.g., mass, volume, and / or average chemical composition across all input buckets) in preparation for operation of the EAF.

[0017] The actual batch profile determined according to some of the methods and systems described and disclosed herein may be integrated into existing programmable logic controllers (PLCs) associated with the operation of the EAF (whether with only one or multiple input buckets) to provide accurate tap weight (e.g., weight of molten steel exiting the EAF) estimates and optimized energy usage (e.g., current consumed by the electrodes of the EAF). According to some embodiments of the present invention, the methods and systems can determine the real-time volume of scrap metal material in one or more input buckets, for example, by generating a depth image using stereoscopic vision (e.g., comparing relative pixel positions between two digital images taken from two vantage points and extracting three-dimensional information by examining the relative positions of objects or pixels in the two digital images) and calculating the volume of the scrap metal material (e.g., of each layer) in a computer processor based on the depth image. For example, the method and system may also include obtaining an RGB digital image of the scrap metal material and processing the RGB digital image using a machine learning model configured to identify and / or classify the nature of each individual layer of scrap metal material placed in the one or more input buckets, which may have been previously trained and validated by machine learning operations. That is, the scrap metal material placed in the one or more input buckets may include different types or classes of scrap metal material (e.g., different chemical compositions, different shapes, etc.). For example, Table 1 provides a non-exhaustive list of some common classes of scrap metal material along with their respective chemical compositions, bulk densities, and morphologies.

[0018] [Table 1]

[0019] The machine learning model may, for example, identify in real time a particular class or classes of scrap metal material for each layer based at least in part on the RGB digital image of each layer. For example, RGB digital images (e.g., frames) of each individual layer after being deposited in the input bucket may be acquired by an RGB camera and processed in real time by the machine learning model to identify when a first portion of scrap metal material including a first type or class of scrap metal material is newly deposited on a previously deposited layer of a second portion of scrap metal material including a second type or class of scrap metal material. Based on the identification of the type or class of scrap metal material, a scrap density may be assigned to each component or piece of scrap metal material for each layer of scrap metal material placed in the input bin. According to some embodiments of the present invention, the scrap densities assigned to each of the scrap metal materials (whether in one or more layers of the same type or class of scrap metal material or different types or classes of scrap metal material) as well as respective volume values ​​may be utilized in real time to identify an actual batch profile to be input to the EAF.

[0020] According to some embodiments of the invention, the methods and systems described and disclosed herein may issue one or more alarms or warnings (e.g., in real-time). For example, the machine learning model may be further configured to identify undesirable and / or hazardous materials (e.g., undesirable materials in an operating EAF) as they pass through (e.g., containers that contain or may contain oil or undesirable liquids (e.g., one or more oils, etc.)) based on the RGB digital images. In this regard, the identification of the undesirable and / or hazardous materials may be passed to a computer processor, which may be configured to trigger an alarm or warning to an operator that the undesirable and / or hazardous materials are present in the scrap metal material, thereby providing the operator with an opportunity to remove the undesirable and / or hazardous materials before they are undesirably introduced into the EAF. Additionally or alternatively, the method and system may include a computer processor configured to trigger an alarm or warning if a piece of scrap metal is identified as having a volume exceeding a size threshold and / or an orientation extending outside a predetermined work zone, where such identification may be based on the volume identification results and / or the depth image. Should such an oversized piece of scrap metal be introduced into the EAF, the act of introducing the oversized piece of scrap metal may cause the piece to strike or otherwise damage an electrode extending into the hearth section of the EAF. Identifying the presence of such an oversized piece of scrap metal and triggering an alarm may allow an operator to remove the oversized piece or to at least partially retract the electrode from the hearth of the EAF to prevent damage to the electrode.

[0021] Additionally or alternatively, the methods and systems may include a computer processor configured to trigger an alarm or warning when it is determined that the volume of scrap metal material present in one or more input buckets exceeds the available volume of the EAF. In this regard, identifying such a volumetric situation and triggering an alarm prevents an undesirable situation in which the volume of scrap metal material input to the EAF exceeds the available volume of the EAF (e.g., to the extent that the top of the EAF cannot be closed and / or sealed). Identifying such a situation and / or triggering an alarm may enable an operator to proactively manipulate or control the input methodology for the EAF (e.g., change the order in which input buckets are input, only partially input certain input buckets, etc.).

[0022] FIG. 1 illustrates an example scrap yard layout for a batch-feed EAF. As shown in FIG. 1, the scrap yard may include a number of input buckets 10, storage bins 15 (or rail cars), and a crane 20. In this regard, the storage bins 15 may each contain scrap metal material, as shown in FIGS. 2A-D. For example, as shown in FIGS. 2A-D, the scrap metal material 47 may include at least a first type or class of scrap metal material 47a and a second type or class of scrap metal material 47b. In this regard, the crane 20 may aggregate input batches to be fed into the batch-feed EAF by moving portions of the widely varying types or classes of scrap metal material from one or more storage bins 15 to one or more input buckets 20. FIG. 1 also illustrates a 3D or depth camera mounted proximate above at least one input bucket 10. In this regard, the 3D or depth camera 110 is oriented to capture a digital image of the interior of the input bucket 10. Thus, digital images may be acquired after each layer of scrap metal material of the same type or class or different types or classes is deposited in the input bucket, allowing for layer-by-layer analysis of the scrap metal material deposited in the input bucket. The 3D or depth camera may include, for example, two stereoscopic monochrome cameras, a red-green-blue (RGB) camera, and an infrared spectrum projector. The 3D or depth camera may also include an inertial measurement unit (IMU) including a gyroscope and an accelerometer, which may help counteract unwanted vibrations of the camera during use. An example of a 3D or depth camera is the Intel® RealSense™ Depth Camera D435. The 3D or depth camera is configured, for example, to acquire digital images of the scrap metal material, for example, layer-by-layer as each layer of scrap metal material is deposited.

[0023] For example, Figure 3 illustrates an input bucket 10 containing scrap metal material 47 defined by multiple layers of various types or classes of scrap metal material 47a, 47b, 47c, 47d, according to some embodiments of the present invention. As shown in Figure 3, methods and systems according to some embodiments of the present invention can identify the volume of each layer of the scrap metal material by using (for example) stereoscopic vision to define a depth image, where one or more pixels of the depth image are mapped or otherwise correlated with 3D coordinates of the scrap metal material, and can be used (for example, in a computer processor) to identify the volume of the layer of scrap metal being analyzed. Methods and systems according to some embodiments of the present invention can also identify and / or classify (for example, by a computer processor) a type or class of each layer of the scrap metal material by layer based at least in part on at least one RGB digital image of each layer immediately after (or after) deposition of the respective layer of the scrap metal material, and assign to each identified layer of the scrap metal a density and / or chemical composition associated with the identified type or class of the scrap metal material. In this regard, a combination of the respective volume of each layer of scrap metal material and the respective assigned density of each layer of scrap metal material can be used to determine the respective mass of each layer of scrap metal material (as shown in FIG. 3). According to some embodiments of the present invention, the respective mass and / or respective chemical composition of each layer of scrap metal material can be used in a computer processor to determine a total mass, chemical composition, and volume of the scrap metal material loaded into the input bucket. In this regard, an actual batch profile of the contents of the input bucket can be determined in a computer processor, which may be integrated (e.g., directly wired or wirelessly connected) with another programmable logic controller (PLC) that may be associated with monitoring and / or controlling the operation of the EAF (e.g., positioning the electrodes, controlling current to the electrodes, etc.).As described throughout, methods and systems according to some embodiments of the present invention are capable of identifying and classifying a first type or class 47a and a second type or class 47b of scrap metal material using a machine learning model based at least in part on an RGB digital image captured by a 3D camera in use, the operation of which may be implemented in a computer processor or PLC associated with monitoring and / or controlling the operation of the EAF.

[0024] 4 illustrates an input bucket holding multiple layers of various types or classes of scrap metal material, including oversized pieces of scrap material 49 that may be detected and cause an alarm to be triggered according to some embodiments of the present invention. In this regard, the oversized pieces of scrap material 49 may be identified in a computer processor based on depth images acquired with a 3D camera or depth camera as described and disclosed herein. For example, as shown in FIG. 4, the oversized pieces of scrap metal material may be associated with individual volumes or geometric shapes that exceed a volumetric or size threshold, which may be fixed or may vary with the available volume of the EAF depending on the operational state of the EAF.

[0025] According to some embodiments of the present invention, one or more controllable illumination sources may be mounted proximate to the 3D or depth camera and configured to provide constant illumination in real time. Consistent illumination during operation may be helpful in classifying scrap metal material because the machine learning model may include deep learning with one or more convolutional layers that consider and / or weigh color scales at one or more locations of the scrap metal material. In this regard, consistent or constant illumination reduces the likelihood of errors due to image quality of the RGB digital images used to identify and / or classify the scrap metal material.

[0026] As discussed above, a scheduled batch to be input to the EAF may include multiple input buckets. For example, FIG. 5 illustrates an input schedule for three buckets that constitute a batch to be input to the EAF, according to some embodiments of the present invention. As generally illustrated in FIG. 5, each of the input buckets 10a, 10b, 10c may uniquely contain different volumes and different types or classes of scrap metal material, which may be selected by an operator based on a planned timeline from one or more of the storage bins 15. According to some embodiments of the present invention, a computer processor may determine an overall batch profile based on the scrap metal material loaded into any number of input buckets containing the scrap metal material. In this regard, layer-by-layer characteristics of each input bucket may be determined, as described and disclosed herein. Digital images of each layer of scrap metal material may be acquired with a 3D camera (e.g., as shown in FIG. 6) or a depth camera, as described above. FIG. 6 shows an exemplary 3D or depth camera 110, which includes a pair of monochrome cameras 112 and an RGB camera 113 configured for stereoscopic viewing.

[0027] In one aspect, some embodiments of the invention provide a method for identifying a batch profile for a batch-feed EAF including one or more input buckets. According to some embodiments of the invention, the method includes the steps of: (i) acquiring respective RGB digital images of respective portions of respective layers (e.g., one or more individual layers stacked on top of one another) of scrap metal material deposited in a first input bucket; (ii) generating, in a computer processor, respective stereo camera depth images of the respective portions of the respective layers of scrap metal material, the computer processor correlating, in the computer processor, one or more pixels of the respective depth images with respective 3D coordinates of the respective portions of the respective layers of scrap metal material; and (iii) correlating, in the computer processor, the respective 3D coordinates of the respective portions of the respective layers of scrap metal material. (v) identifying, by a machine learning model in the computer processor, a classification of each of the portions of each of the layers of scrap metal material from among a group of known scrap metal materials based at least in part on each of the RGB digital images and assigning a respective density to each of the portions of each of the layers of scrap metal material, and (vi) identifying a respective mass associated with each of the portions of each of the layers of scrap metal material based on (a) the respective volumes of each of the portions of each of the layers of scrap metal material and (b) the respective assigned densities. For example, FIG. 7 shows a stereoscopically generated depth image on the left and a corresponding RGB digital image on the right.

[0028] As discussed above, the method may include assigning, in a computer processor, by a machine learning model, respective chemical compositions to respective portions of respective layers of scrap metal material based at least in part on the respective RGB digital images. Additionally or alternatively, the method may include determining, in a computer processor, a total mass of the scrap metal material in the first input bucket, and / or a total volume of the scrap metal material in the first input bucket, and / or a total chemical composition of the scrap metal material in the first input bucket. Additionally or alternatively, the method may further include determining, in a computer processor, a total mass of the scrap metal material in the second input bucket, and / or a total volume of the scrap metal material in the second input bucket, and / or a total chemical composition of the scrap metal material in the second input bucket. Additionally or alternatively, the method may further include determining a total mass of the scrap metal material in the third input bucket, and / or a total volume of the scrap metal material in the third input bucket, and / or a total chemical composition of the scrap metal material in the third input bucket. According to some embodiments of the invention, the method may include, in a computer processor, determining a multi-bucket total mass of the scrap metal material in the two or more input buckets, and / or a multi-bucket total volume of the scrap metal material in the two or more input buckets, and / or a multi-bucket total chemical composition of the scrap metal material in the two or more input buckets. In this regard, a batch profile (whether single or multiple input buckets of scrap metal material) may be determined and integrated with other control schemes associated with the operation of the EAF.

[0029] According to some embodiments of the invention, for example, where each portion of each layer of scrap metal material deposited in a first input bucket includes a first portion of the scrap metal material deposited in the first input bucket as a first layer, the method may include the steps of: (i) acquiring an RGB digital image of the first portion of the scrap metal material deposited in the first input bucket as a first layer; and (ii) generating, in a computer processor, a first stereo camera depth image of the first portion of the scrap metal material, wherein the computer processor correlates one or more pixels of the first depth image with 3D coordinates of the first portion of the scrap metal material. (iii) identifying, in the computer processor, a first volume of the first portion of the scrap metal material based at least in part on the 3D coordinates of the first portion of the scrap metal material, (v) identifying, in the computer processor, with a machine learning model, a first classification of the first portion of the scrap metal material from among a group of known scrap metal materials based at least in part on the first RGB digital image and assigning a first density to the first portion of the scrap metal material, and (vi) identifying a mass associated with the first portion of the scrap metal material based on (a) the first volume of the first portion of the scrap metal material and (b) the assigned first density. The method may, according to some embodiments of the present invention, include assigning, in the computer processor, with a machine learning model, a first chemical composition of the first portion of the scrap metal material based at least in part on the first RGB digital image.Additionally, each portion of each layer of scrap metal material deposited in the first input bucket includes a second portion of scrap metal material deposited in the first input bucket on the first layer as a second layer, the second portion of scrap metal material being different from the first portion of scrap metal material, the method further comprising: (i) acquiring a second RGB digital image of the second portion of scrap metal material deposited in the first input bucket on the first layer as a second layer, the second portion of scrap metal material being different from the first portion of scrap metal material; and (ii) generating, in the computer processor, a second stereoscopic depth image of the second portion of scrap metal material, the second depth image being different from the first portion of scrap metal material, the computer processor generating, in the computer processor, a stereoscopic depth image of the second portion of scrap metal material. The generating step may include correlating the pixels with 3D coordinates of the second portion of the scrap metal material, (iii) identifying, in the computer processor, a second volume of the second portion of the scrap metal material based at least in part on the 3D coordinates of the second portion of the scrap metal material, (v) identifying, in the computer processor, with a machine learning model, a second classification of the second portion of the scrap metal material from among a group of known scrap metal materials based at least in part on the second RGB digital image and assigning a second density to the second portion of the scrap metal material, and (vi) identifying a mass associated with the second portion of the scrap metal material based on (a) the second volume of the second portion of the scrap metal material and (b) the assigned second density. According to some embodiments of the invention, the method may include assigning, in the computer processor, with a machine learning model, a second chemical composition of the second portion of the scrap metal material based at least in part on the second RGB digital image.

[0030] According to some embodiments of the invention, the method may optionally include merging, in a computer processor, the respective depth images and the respective RGB digital images to generate respective 3D composite images of the respective portions of the respective layers of the scrap metal material, and optionally displaying the respective 3D composite images on a user display. Additionally or alternatively, the method may include performing a volumetric calibration operation, the volumetric calibration operation including mapping a plurality of pixels associated with a plurality of known locations of a calibration tool (e.g., an inanimate object of known dimensions disposed in an empty input bucket) to a 3D coordinate system, and generating, in the computer processor, a volumetric computational model based on the mapped pixels, the volumetric computational model configured, in the computer processor, to identify respective volumes of the respective portions of the respective layers of the scrap metal material based at least in part on the respective 3D coordinates of the respective portions of the respective layers of the scrap metal material.

[0031] According to some embodiments of the invention, the first RGB image may be sent to a machine learning model (for example) to cause the machine learning model to converge on a solution, e.g., by a neural network in a computer processor, to identify and / or classify the imaged scrap metal (e.g., a first portion of the scrap metal) from among a group of known scrap metals based on at least a portion of the first RGB image. Once the machine learning model has converged on a solution for identifying the imaged scrap metal (e.g., NO2 Heavy), the computer processor may assign (for example) a first density and / or a first chemical composition to the first portion of the scrap metal (e.g., NO2 Heavy has a density of 0.75 kg / m 3 ), a mass and / or chemical composition associated with the first portion of the scrap metal material may be determined based on (a) the first volume of the first portion of the scrap metal material and (b) the assigned first density and / or the assigned first chemical composition.

[0032] According to some embodiments of the present invention, the machine learning models disclosed and described herein may include one or more types of machine learning, such as, for example, neural networks (e.g., convolutional neural networks, recurrent neural networks, deep learning neural networks, natural language processing, etc.), transformers, support vector machines (SVMs), and clusters (e.g., clustered machine learning). In this regard, for example, use of the term "machine learning model" throughout this disclosure may encompass (i) a single type of machine learning model according to some embodiments of the present invention, and (ii) different types of machine learning models according to other embodiments of the present invention. By way of example only, a first type of machine learning model (e.g., a neural network) may be used to identify and classify scrap metal material from among one or more types of classifications, while a second type of machine learning model may be used to identify and / or classify undesirable or potentially undesirable materials (e.g., oil or other liquids that are undesirable in the EAF, sealed containers that may contain oil or other liquids that are undesirable in the EAF, etc.) that are mixed with the scrap metal material intended for input into the EAF. The first type of machine learning model may include, for example, a deep learning solution having multiple convolutional layers and / or multiple pooling layers and at least one fully connected layer configured for the identification and / or classification of one or more types or classes of scrap metal material. In this regard, for example, at least one of the convolutional layers may capture the highest scoring scrap type selected as the classification result for the input scrap image. According to some embodiments of the present invention, the first convolutional layer may consider and / or weigh the identified straight edges, arc edges, colors, and / or color gradients across the surface. According to some embodiments of the present invention, different layers may have different node layouts, where each node may have a respective weight. The respective weights may be calculated during the training phase, and the model output may be customized according to different customer wrap types utilized in the EAF.

[0033] According to some embodiments, the scrap classification model is developed using machine learning techniques. For example, an initial model for training may be selected, which may be done, for example, by backpropagation, which serves to feed the classified data into the model and then measures the performance of the model. The error rate is measured using a loss function. In this regard, for example, backpropagation serves to measure the rate of change of the loss function with respect to the weighting of each connection, and a gradient descent step is used to reduce the error rate of each connection as close to zero as possible to ensure that the network converges to a solution. According to some embodiments of the present invention, training of an initial algorithm or initial model configured for general image identification may be performed by selecting one or more of the following: a percentage of images to be used for training and validation (e.g., 70% for training and 30% for validation), applying various preprocessing to the input images, setting training and validation thresholds, defining the number of iterations per epoch, selecting a gradient thresholding method, selecting a learning rate, and validating a confusion matrix. Once the trained model meets all user specifications, the model may be exported to a computer processor for use in image identification and / or image classification according to some embodiments of the present invention. For example, the algorithm or model may be exported as an ONNX (Open Neural Network Exchange) model. ONNX (Open Neural Network Exchange) is an open format for expressing artificial intelligence models that is widely supported and found in many frameworks, tools, and hardware. According to some embodiments, the algorithm or model may be retrained if the number of non-solutions reaches an unacceptable number during operation of some methods described and disclosed herein.

[0034] In another aspect, some embodiments according to the present invention provide a method of operating a batch-feed electric arc furnace (EAF) including one or more input buckets. The method may include identifying an actual batch profile defined by scrap metal material contained in the one or more input buckets. Identifying the actual batch profile may include (i) identifying respective masses associated with respective portions of respective layers of scrap metal material based on (a) respective volumes of respective portions of respective layers of scrap metal material and (b) respective assigned densities as described and disclosed herein, and / or (ii) identifying a total mass of the scrap metal material in the one or more input buckets as described and disclosed herein, and / or a total volume of the scrap metal material in the one or more input buckets as described and disclosed herein, and / or (iii) identifying a total chemical composition of the scrap metal material in the one or more input buckets as described and disclosed herein. The method may further include, in a computer processor, comparing the actual batch profile to operating conditions and / or operating set points of the EAF and modifying the operating conditions and / or operating set points based on the actual batch profile.

[0035] According to some embodiments of the invention, the method may include modifying one or more operating conditions and / or one or more operating set points, such as adjusting the depth of the electrode extending into the EAF to prevent scrap from falling into the EAF, which could destroy the electrode. Additionally or alternatively, the method may include modifying one or more operating conditions and / or one or more operating set points, such as adjusting the power profile of the EAF to provide a more conservative inner diameter of the approach path, and / or adjusting the electrode water cooling profile. Additionally or alternatively, the computer processor may be configured to identify and / or trigger a first alarm when a piece of scrap metal material is identified as having a volume or three-dimensional size that exceeds a size threshold and / or an orientation that extends outside of a predetermined working zone defined by a fixed volume of the EAF. In this regard, the size threshold may be based on a volume or size that may undesirably strike an electrode in the EAF during operation. In such a situation, the operator has the opportunity to remove the oversized piece of scrap metal material or to modify the operating state of the EAF based on information that an oversized piece of scrap metal material is present. Additionally or alternatively, the machine learning model in the computer processor may be configured to trigger a second alarm upon identifying an undesirable substance (e.g., oil or other liquid that is undesirable in an EAF) mixed in with the scrap metal material.

[0036] In another aspect, some embodiments according to the present invention provide a system for identifying a batch profile for a batch-feed electric arc furnace (EAF) including one or more feed buckets. The system may include a sensor fusion module, which includes two monochrome cameras configured to provide stereoscopic vision, an RGB camera, an infrared spectrum projector, and a first computer processor in operative communication with the two monochrome cameras, the RGB camera, and the infrared spectrum projector. The first computer processor may be configured to (a) generate respective stereoscopic depth images and correlate one or more pixels of the respective depth images with respective 3D coordinates of respective portions of respective layers of scrap metal material, and (c) identify respective volumes of respective portions of respective layers of scrap metal material based at least in part on the respective 3D coordinates of respective portions of respective layers of scrap metal material. The system may also include a scrap classification module, which is in operative communication with the sensor fusion module and includes a machine learning model and a second computer processor. The second computer processor may be configured to: (a) identify a classification of each portion of each layer of the scrap metal material from among a group of known scrap metal materials based on at least a respective portion of the respective RGB digital images received from the sensor fusion module; (b) assign a respective density to each portion of each layer of the scrap metal material; and (c) determine a respective mass associated with each portion of each layer of the scrap metal material based on (a) a respective volume of the each portion of the respective layer of the scrap metal material and (b) the respective assigned density.

[0037] In yet another aspect, some embodiments according to the present invention provide an electric arc furnace (EAF) system. The system may include an EAF including a furnace body defining an interior portion and a set of electrodes extending into the interior portion. The system may also include one or more input buckets. According to some embodiments of the present invention, the system may include a sensor fusion module mounted above and proximate to the one or more input buckets, the sensor fusion module including two monochrome cameras configured to provide stereoscopic vision, an RGB camera, an infrared spectrum projector, and a first computer processor in operative communication with the two monochrome cameras, the RGB camera, and the infrared spectrum projector. The first computer processor may be configured to (a) generate respective stereoscopic depth images and correlate one or more pixels of the respective depth images with respective 3D coordinates of respective portions of respective layers of scrap metal material, and (c) determine respective volumes of respective portions of respective layers of scrap metal material based at least in part on the respective 3D coordinates of respective portions of respective layers of scrap metal material. The system may also include a scrap classification module in operative communication with the sensor fusion module and including a machine learning model and a second computer processor configured to: (a) identify a classification of each portion of each layer of scrap metal material from among a group of known scrap metal materials based on at least a respective portion of each RGB digital image received from the sensor fusion module; (b) assign a respective density to each portion of each layer of scrap metal material; and (c) determine a respective mass associated with each portion of each layer of scrap metal material based on (a) a respective volume of the each portion of the each layer of scrap metal material and (b) the respective assigned density.

[0038] Figure 8 illustrates a system functionality flow chart according to some embodiments of the present invention. As shown generally in Figure 8, the 3D camera 110 (part of the sensor fusion module) captures digital images that may be processed or analyzed in (i) the volume measurement module 180 and (ii) the scrap classifier 210, where (i) the volume measurement module 180 identifies the volume of the scrap metal material based on the depth image in a computer processor, and (ii) the scrap classifier 210 includes a machine learning model adapted to identify and / or classify the scrap metal material based on the RGB digital image of the scrap metal material as described and disclosed herein. Once the scrap metal material is classified, an associated density and / or chemical composition is assigned to the scrap metal material. The identified volume and assigned density may be utilized in the input bucket integration module, which identifies the layer-by-layer and / or total mass of the scrap metal material in the input bucket, thereby defining the EAF batch input 400. FIG. 8 also illustrates that some embodiments may also use data received from a load cell 300 located below the input bucket (i.e., shown as truck weight in FIG. 8). In this regard, the mass and volume determined by the volume measurement module 180 and scrap classifier 180 may be compared in a computer processor to the weight / mass measured by the load cell 180. In this regard, if there is a discrepancy between the determined mass and the measured mass from the load cell, an alarm may be triggered that there may be more than just scrap metal material present in the input bucket, such as lime or carbon, which may work their way down to the bottom of the input bucket and increase the weight and / or mass held in the input bucket, but do not significantly affect the volume determined based on the respective depth images. According to some embodiments of the present invention, operating conditions may be modified based on the identification of lime or carbon present in the batch being input to the EAF.Additionally, from the difference between the weight / mass measured by the load cell and the determined mass of the batch, the mass of lime and / or carbon present in the batch can be estimated, which may be used to modify the proposed operation of the EAF.Also as shown in FIG. 8, an alarm 350 may be triggered if the determined volume of scrap metal material exceeds a threshold, as described above.

[0039] FIG. 9 illustrates a schematic of a system for operating an EAF system according to some embodiments of the present invention, including a sensor fusion module 100 in which a computer processor 120 is in operative communication with a scrap metal volume measurement module 180. The sensor fusion module 100 may transmit real-time volume values ​​of the imaged scrap metal material to a PLC 200 (or other computer processor) associated with controlling the operation of the EAF. The sensor fusion module 100 also transmits an RGB digital image 107 of the scrap metal material to a scrap classification module 210. The scrap classification module 210 uses machine learning models to identify a type or class of the scrap metal material and assign a density and / or chemical composition to the scrap metal material. The assigned density and / or assigned chemical composition of the scrap metal material, as well as the determined volume, may be input to a bucket layer module 220, which outputs a layer-by-layer mass and / or a total mass of the scrap metal material. The PLC 200 may also receive various scheduled batch targets 500 for the batch input. The PLC may compare the scheduled batch targets 500 for the batch input with the specified profile of the scrap metal material of the batch and adjust the operating conditions or parameters of the EAF for processing the batch.

[0040] These and other modifications and variations to the embodiments of the present invention may be made by those skilled in the art without departing from the spirit and scope of the present invention, which is more particularly set forth in the appended claims. Moreover, it will be understood that aspects of the various embodiments may be interchanged in whole or in part. Moreover, those skilled in the art will appreciate that the above description is merely illustrative and is not intended to limit the invention, as further set forth in such appended claims. Thus, the spirit and scope of the appended claims should not be limited to the exemplary description of the forms contained herein.

Claims

1. 1. A method for identifying a batch profile for a batch-feed electric arc furnace (EAF) including one or more feed buckets, comprising: (i) acquiring respective red-green-blue (RGB) digital images of respective portions of respective layers of scrap metal material deposited in a first input bucket; (ii) generating, in a computer processor, respective stereo camera depth images of the respective portions of the respective layers of scrap metal material, wherein the computer processor correlates one or more pixels of the respective depth images with respective 3D coordinates of the respective portions of the respective layers of scrap metal material; (iii) in the computer processor, determining a volume of each of the respective portions of the respective layers of scrap metal material based at least in part on the 3D coordinates of the respective portions of the respective layers of scrap metal material; (v) identifying, with a machine learning model in the computer processor, a classification for each of the respective portions of the respective layers of scrap metal from among a group of known scrap metal materials based on at least each of the respective portions of the respective RGB digital images and assigning a respective density to each of the respective portions of the respective layers of scrap metal; (vi) determining respective masses associated with the respective portions of the respective layers of scrap metal material based on (a) the respective volumes of the respective portions of the respective layers of scrap metal material, and (b) the respective assigned densities; The method includes:

2. 2. The method of claim 1, further comprising: assigning, by the machine learning model in the computer processor, respective chemical compositions to the respective portions of the respective layers of scrap metal material based at least in part on the respective RGB digital images.

3. 3. The method of claim 1, further comprising determining, in the computer processor, a total mass of scrap metal material in the first input bucket, and / or a total volume of scrap metal material in the first input bucket, and / or a total chemical composition of scrap metal material in the first input bucket.

4. 4. The method of claim 3, further comprising determining in the computer processor a total mass of scrap metal material in a second input bucket, and / or a total volume of scrap metal material in the second input bucket, and / or a total chemical composition of the scrap metal material in the second input bucket.

5. 5. The method of claim 3, further comprising determining a total mass of scrap metal material in a third input bucket, and / or a total volume of scrap metal material in the third input bucket, and / or a total chemical composition of the scrap metal material in the third input bucket.

6. 6. The method of claim 3, further comprising determining in the computer processor a multi-bucket total mass of the scrap metal material in two or more input buckets, and / or a multi-bucket total volume of the scrap metal material in two or more input buckets, and / or a multi-bucket total chemical composition of the scrap metal material in two or more input buckets.

7. Each portion of each layer of scrap metal material deposited in the first input bucket includes a first portion of the scrap metal material deposited in the first input bucket as a first layer, and the method further comprises: (i) acquiring an RGB digital image of a first portion of scrap metal material deposited in a first input bucket as a first layer; (ii) generating, in a computer processor, a first stereo camera depth image of the first portion of the scrap metal material, the generating step correlating, in the computer processor, one or more pixels of the first depth image with 3D coordinates of the first portion of the scrap metal material; (iii) determining, in the computer processor, a first volume of the first portion of the scrap metal material based at least in part on the 3D coordinates of the first portion of the scrap metal material; (v) identifying, with a machine learning model in the computer processor, a first classification of the first portion of scrap metal material from among a group of known scrap metal material based at least in part on the first RGB digital image and assigning a first density to the first portion of scrap metal material; (vi) determining a mass associated with the first portion of the scrap metal material based on (a) the first volume of the first portion of the scrap metal material, and (b) the assigned first density; The method according to any one of claims 1 to 6, comprising:

8. 8. The method of claim 7, further comprising: assigning, by the machine learning model in the computer processor, a first chemical composition of the first portion of the scrap metal material based at least in part on the first RGB digital image.

9. each portion of each layer of scrap metal material deposited in a first input bucket includes a second portion of scrap metal material deposited in the first input bucket on the first layer as a second layer, the second portion of scrap metal material being different from the first portion of scrap metal material, the method further comprising: (i) acquiring a second RGB digital image of a second portion of scrap metal material deposited on the first layer in the first input bucket as a second layer, the second portion of scrap metal material being different from the first portion of scrap metal material; (ii) generating, in the computer processor, a second stereoscopic depth image of the second portion of the scrap metal material, the generating step correlating, in the computer processor, one or more pixels of the second depth image with 3D coordinates of the second portion of the scrap metal material; (iii) determining, in the computer processor, a second volume of the second portion of the scrap metal material based at least in part on the 3D coordinates of the second portion of the scrap metal material; (v) identifying, with the machine learning model in the computer processor, a second classification of the second portion of scrap metal material from among a group of known scrap metal material based at least in part on the second RGB digital image and assigning a second density to the second portion of scrap metal material; (vi) determining a mass associated with the second portion of the scrap metal material based on (a) the second volume of the second portion of the scrap metal material, and (b) the assigned second density; The method according to any one of claims 1 to 8, comprising:

10. 10. The method of claim 9, further comprising: assigning, by the machine learning model in the computer processor, a second chemical composition of the second portion of the scrap metal material based at least in part on the second RGB digital image.

11. 10. The method of claim 1, further comprising the step of performing a volumetric calibration operation, the volumetric calibration operation comprising mapping a plurality of pixels associated with a plurality of known locations of a calibration tool positioned in an empty input bucket to a 3D coordinate system, and generating in the computer processor a volumetric calculation model based on the mapped pixels, the volumetric calculation model being configured in the computer processor to determine the respective volumes of the respective portions of the respective layers of scrap metal material based at least in part on the respective 3D coordinates of the respective portions of the respective layers of scrap metal material.

12. 12. The method of any one of claims 1 to 11, further comprising merging, in the computer processor, the respective depth images with the respective RGB digital images to generate respective 3D composite images of the respective portions of the respective layers of scrap metal material, and optionally displaying the respective 3D composite images on a user display.

13. The method according to any one of claims 1 to 12, wherein the machine learning model includes deep learning having multiple convolutional layers and / or multiple pooling layers and at least one fully connected layer, and at least one of the convolutional layers captures the scrap type with the highest score selected as a classification result for the input scrap image.

14. The method according to any one of claims 1 to 13, wherein the classification model implements an algorithm developed by machine learning.

15. 1. A method of operating a batch-feed electric arc furnace (EAF) including one or more feed buckets, comprising: (i) determining an actual batch profile defined by scrap metal material contained in one or more input buckets, the determining comprising: (a) determining the respective volumes of the respective portions of the respective layers of scrap metal material, and (b) determining respective masses associated with the respective portions of the respective layers of scrap metal material based on the respective assigned densities as claimed in any one of claims 1 to 14; and / or determining a total mass of scrap metal material in one or more input buckets; and / or a total volume of scrap metal material in the one or more input buckets as claimed in any one of claims 3 to 14; and / or determining a total chemical composition of the scrap metal material in the one or more input buckets as claimed in any one of claims 3 to 14; (ii) in a computer processor, comparing the actual batch profile to operating conditions and / or operating set points of the EAF; (iii) modifying said operating conditions and / or said operating set points; The method includes:

16. The method of claim 15 , wherein modifying the operating conditions and / or the operating set points comprises adjusting a depth of an electrode extending into the EAF.

17. 17. The method of any one of claims 15 to 16, wherein the computer processor is configured to identify and / or trigger a first alarm when the piece of scrap metal material is identified as having a volume or three-dimensional size that exceeds a size threshold and / or an orientation that extends outside a predetermined working zone defined by a fixed volume of the EAF.

18. 18. The method of any one of claims 15 to 17, wherein the machine learning model in the computer processor is configured to trigger a second alarm upon identifying an undesirable substance mixed into the scrap metal material.

19. 1. A system for identifying a batch profile for a batch-feed electric arc furnace (EAF) including one or more feed buckets, comprising: (i) a sensor fusion module including two monochrome cameras configured to provide stereoscopic vision, an RGB camera, an infrared spectrum projector, and a first computer processor in operative communication with the two monochrome cameras, the RGB camera, and the infrared spectrum projector, the first computer processor configured to: (a) generate stereoscopic depth images and correlate one or more pixels of the respective depth images with respective 3D coordinates of respective portions of the respective layers of scrap metal material; and (c) determine respective volumes of the respective portions of the respective layers of scrap metal material based at least in part on the respective 3D coordinates of the respective portions of the respective layers of scrap metal material. (ii) a scrap classification module in operative communication with the sensor fusion module, the scrap classification module including a machine learning model and a second computer processor configured to: (a) identify a classification of each of the portions of each of the layers of scrap metal material from among a group of known scrap metal materials based on at least a respective portion of each of the RGB digital images received from the sensor fusion module; (b) assign a respective density to the each portion of each of the layers of scrap metal material; and (c) determine a respective volume of the each portion of the each of the layers of scrap metal material, and (b) a respective mass associated with the each portion of the each of the layers of scrap metal material based on the respective assigned density. A system including:

20. 1. An electric arc furnace (EAF) system comprising: (i) an EAF including a furnace body defining an interior portion and a set of electrodes extending into the interior portion; (ii) one or more input buckets; (iii) a sensor fusion module mounted above and proximate to the one or more input buckets, the sensor fusion module including two monochrome cameras configured to provide stereo vision, an RGB camera, an infrared spectrum projector, and a first computer processor in operative communication with the two monochrome cameras, the RGB camera, and the infrared spectrum projector, the first computer processor configured to: (a) generate respective stereo depth images and correlate one or more pixels of the respective depth images with respective 3D coordinates of respective portions of respective layers of scrap metal material; and (c) determine respective volumes of the respective portions of the respective layers of scrap metal material based at least in part on the respective 3D coordinates of the respective portions of the respective layers of scrap metal material; (ii) a scrap classification module in operative communication with the sensor fusion module, the scrap classification module including a machine learning model and a second computer processor configured to: (a) identify a classification of each of the portions of each of the layers of scrap metal material from among a group of known scrap metal materials based on at least a respective portion of each of the RGB digital images received from the sensor fusion module; (b) assign a respective density to the each portion of each of the layers of scrap metal material; and (c) determine a respective volume of the each portion of the each of the layers of scrap metal material, and (b) a respective mass associated with the each portion of the each of the layers of scrap metal material based on the respective assigned density. A system including:

Images