Furnace controller and method of operating a furnace

Abstract

The control scheme of the furnace can use real-time and historical data to model performance and determine the relationship between different data and performance parameters for real-time correction of suboptimal performance of the furnace. The operating parameters can be recorded throughout the cycle for all cycles over a period of time to establish a baseline. This data can then be used to calculate the performance of the process. Regression analysis can be performed to determine which parameters affect different aspects of performance. These relationships can then be used to predict performance during a single cycle in real time and provide closed-loop or open-loop feedback to control the furnace operation to improve performance.

Description

Technical Field

[0001] The present invention relates to controllers configured to facilitate controlling the operation of a furnace used to melt metalliferous materials and methods of making and using the same. Background Art

[0002] Secondary metal recovery and reprocessing processes may involve the melting of metal scraps from different sources and upstream processes. The scrap being recovered often varies greatly in shape, size, and composition. This high degree of inconsistency in the input scrap material must be accounted for in the melting process. As a result, the energy requirements of a furnace used for metal recovery or reprocessing may often vary frequently from melt to melt. This may result in a high degree of variability in the operating efficiency and productivity of the furnace. When an operator operates multiple furnaces simultaneously, the losses caused by process variability may be magnified.

[0003] Examples of secondary metal recovery processes can be found in U.S. Patent No. 4,484,947 and U.S. Patent Application Publication No. 2019 / 0360067. U.S. Patent No. 4,484,947 relates to controlling the aluminum melting process in a closed furnace chamber by measuring a specific temperature in the furnace and adjusting the firing rate of the burner to quickly melt the aluminum. U.S. Patent Application Publication No. 2019 / 0360067 discloses a system for controlling a melting process in a furnace. Summary of the invention

[0004] We have determined that there can be considerable variability in the secondary remelting process or recovery process for metals (e.g., secondary aluminum remelting / recovery process, secondary copper remelting / recovery process, secondary steel remelting or recovery process, secondary lead remelting or recovery process, secondary iron remelting or recovery process, etc.). This can include dross material, recovered metal (e.g., aluminum), and the metal used for remelting (e.g., aluminum, copper, iron, lead, or steel used for remelting, etc.). We have determined that there is significant variability from charge material to furnace conditions to operation. These sources of variability can introduce a significant amount of uncertainty into the process, which can have a significant impact on performance.

[0005] For example, the charge material often includes different types of scrap, dross, and metal, all of which are different in size, shape, and composition. Flux (salt) is also often added to the charge material, which can vary in composition and amount. The composition of certain types of charge materials (such as dross) may vary from batch to batch (or cycle) in unknown amounts due to the production method. Therefore, a single type of dross material from a specific source may still contain different amounts of metal (e.g., 30 to 50 weight % aluminum), thereby causing further unknown or unidentified variability in the material input.

[0006] We have determined that furnace condition can be used to describe various furnace parameters that can significantly affect performance, such as furnace type and condition. Furnace type can include furnace design (such as rotary, reverb, sidewell, dual chamber, etc.) and furnace configuration, such as single pass or dual pass, burner fuel type, oxygen percentage in oxidant, flue configuration, burner design, burner mounting details (including angles), and rated furnace size. We have determined that the condition of a furnace can be described by its temperature, cleanliness, refractory type and condition, door condition (insulation and sealing), degree of fullness for a given cycle, construction materials (affecting heat transfer), environmental conditions surrounding the furnace (including adjacent operations), and degree of sealing of the furnace. The furnace operation (which can vary during a cycle) can include multiple parameters that may include, for example, burner firing rates, variations in burner operation (e.g., staging), any variability in fuel or oxidant composition, flue temperature, furnace pressure, furnace rotation speed and direction, tilt angle, furnace rotation motor current or hydraulics, door position, operators, operator rotation, slag, and the degree of metal oxidation (e.g., aluminum oxidation) at any point during the cycle.

[0007] We have also determined that secondary metal remelting (e.g., aluminum remelting, copper remelting, steel remelting, etc.) can rely on operator decisions and judgment to identify when the material inside the furnace has reached the appropriate tapping temperature. Conventionally, there is no consistent and accurate method for identifying when the metal is ready to tap. Typically, operators look for specific characteristics of the furnace or operation (such as the appearance of the furnace walls, effluent, or charge) that can be ambiguous and dependent on the operator's experience. We have determined that variability in the process and ambiguity in operator decision making can often lead to inconsistent furnace performance.

[0008] A mass and energy balance model can be used to calculate when the material inside the furnace is ready to be tapped (a "melt end" calculation). An example of this calculation method is described in U.S. Patent Application Publication No. 2019 / 0360067. We have determined that this method can have a certain amount of error depending on the charge material used. For example, individual dross materials with a range of aluminum contents can cause errors in the calculation of the energy required to melt the aluminum, which must be taken into account. For example, if the average aluminum content from a particular dross is 50%, but the range of aluminum content is 40% to 60%, then the actual aluminum content can be anywhere within the range. For a charge weight of 10 metric tons, the theoretical energy required to melt material with an aluminum content of 40%, 50%, and 60% (assuming the remaining material is alumina) is approximately 7,840 megajoules (MJ), 8,260Mj, and 8,680MJ, respectively. Therefore, if the mass and energy balance calculation assumes an average aluminum content of 50%, the calculated required energy may be + / - 5%, or a total potential range of up to about 10%. We have determined that the consequence of this type of error can be that at the end of the melting cycle, the charge is either too cold or too hot (i.e., under-heated or over-heated) if the required energy calculation based on the initial estimate of mass and composition is followed. For example, if the charge is under-heated for aluminum operation, then the aluminum may not have absorbed enough energy to completely melt, which can significantly reduce aluminum recovery. On the other hand, if the charge is overheated, then the aluminum may have absorbed too much energy and therefore become much hotter than necessary. This may also affect aluminum recovery because the higher the temperature, the more easily aluminum oxidizes. Overheated cycles also use more time and energy than necessary, thereby reducing productivity and increasing energy consumption, and therefore carbon emissions.

[0009] We have developed a controller and system that can utilize a control method for furnace operation that can solve the above-mentioned problems we have identified. Embodiments of our controller, system, and method can take into account charge variability by providing real-time correction calculations based on historical and field data. Embodiments can be used in conjunction with the recovery of metals, such as, for example, the processing of dross material, recycled aluminum, aluminum for remelting, melting of metal, melting of metal-containing materials, recovery of metals, and remelting of metal-containing materials. For example, in some embodiments that can process dross material, the processing of the material can include separating the oxide portion of the dross material from the metal component of the dross material by melting the metal component of the dross material, and subsequently separating the oxide portion of the dross material from the metal material (although the metal of the dross material can be melted into a liquid or molten state, the oxide portion of the dross material can still be in solid form). The metal material can then be poured into a container or other form for solidification and / or transportation so that the metal can be used in another type of process (e.g., manufacturing, etc.). For some other embodiments in which metal-containing materials may be processed, the process may include remelting the metal-containing material for recovery of the metal of the material or for some other end use of the metal of the material.

[0010] In some embodiments, a furnace control scheme that can be used by the controller can utilize real-time and historical data to model performance and determine relationships between different data and performance parameters for use in correcting suboptimal performance of the furnace in real-time (e.g., within seconds, minutes, minutes, 5 to 10 minutes, up to 15 minutes). Embodiments of the controller can be configured for use in an open-loop control scheme or a closed-loop control scheme. Embodiments of the controller can be provided for use via a cloud-based service that utilizes at least one Internet connection (e.g., the Internet and / or at least one enterprise network) or a more locally based service. The controller is located within the same network as the components of the furnace (e.g., within the same local network via a wired and / or wireless communication network, a local area network, etc.).

[0011] To establish a baseline data set, operating parameters may be recorded throughout the cycle for all cycles over a period of time for storage in a memory of the controller. The controller may then use this baseline data set to calculate the performance of the process. The controller may be configured to perform regression analysis to determine which parameters affect different aspects of performance. These relationships may then be used to predict performance during a single cycle in real time, and provide closed-loop or open-loop feedback to control the furnace operation, thereby improving the performance of the furnace and making the operation of the furnace more profitable. For example, at least one parameter that significantly affects aluminum oxidation may be identified from the development of at least one reference model, and the identified one or more parameters may be modified by changing the operating parameters in real time to reduce aluminum oxidation for the operation of the furnace in which the aluminum is being melted. Embodiments of the control scheme used by the controller may also be applied to other types of metals to be melted by a furnace (such as those containing iron, lead and / or copper), and to different types of furnaces (including, for example, reverberatory furnaces, dual-chamber furnaces, and side-cylinder furnaces).

[0012] Embodiments of a method for controlling the operation of a furnace to melt a material including a metal (e.g., a material containing a metal). The method may include: storing data related to a charge, furnace conditions, and operating data of the furnace in a non-transitory computer-readable medium of a computer device for a plurality of different operating cycles of the furnace; creating or determining one or more cycle performance parameters based on the stored data; creating a historical data repository based on the stored data parameters and the cycle performance parameters; determining x variables for one or more operating cycles of the furnace and feeding the x variables into a regression model to determine a relationship between at least one of the x variables and at least one y variable to define at least one reference cycle, each y variable corresponding to a respective one of the cycle performance parameters; receiving real-time data from sensors of the furnace; comparing the real-time data from the sensors of the furnace with the at least one reference cycle to determine whether an adjustment to one or more furnace operating parameters is required; and upon determining that there is a difference from the at least one reference cycle that meets or exceeds a significance threshold based on the comparison of the real-time data from the sensors of the furnace with the at least one reference cycle, adjusting the operation of the furnace so that the operation of the furnace is adjusted to approximate a preselected furnace performance.

[0013] In some embodiments of the method, the preselected furnace performance may be defined by at least one reference cycle. The at least one reference cycle may be an expected operation of the furnace, which may be based on a model derived with historical data of past performance of the furnace. Such historical data may be stored in a historical data repository that includes empirical data related to past operating cycles of a furnace melting material including metal. In some embodiments, the reference cycle may be considered to be an optimal performance operating curve for the furnace. In other embodiments, the reference cycle may be a preselected furnace performance target based on furnace historical operating data for furnaces having similar configurations (e.g., the same or similar furnace models melting similar types of materials, etc.).

[0014] Embodiments of the method may also include other steps. For example, the method may include generating the at least one reference cycle. The generating of the at least one reference cycle may include characterizing the cycle into one or more material groups to generate the at least one reference cycle for the desired performance of the furnace.

[0015] The determination of the x-variables for the one or more reference cycles may include removing anomalous data from the population of materials prior to feeding data parameters to a regression model.

[0016] In some embodiments, the at least one y variable may include: a first y variable for a specific fuel consumption, a second y variable for a melting rate, and a third y variable for a yield. In other embodiments, there may be only a single y variable, only two y variables, or more than three y variables. In some embodiments, the one or more y variables may include at least one of a specific fuel consumption, a melting rate, and a yield, or a combination of these variables.

[0017] The at least one reference cycle for each y variable may be determined by identifying one or more best case operating cycles of the furnace. Best case performance cycles may be derived from historical data and / or other data stored in a historical data repository. In some embodiments, best case performance cycles may be derived from historical data for the furnace and further modified to define best case performance cycles. The modifications may be provided by one or more models used in conjunction with historical data of past furnace operations of past operating cycles of the furnace and / or past operations of furnaces of similar types.

[0018] In some embodiments, the at least one reference cycle is a single optimal reference cycle. In other embodiments, the at least one reference cycle includes a plurality of optimal reference cycles defined based on the charge to be fed to the furnace. Each reference cycle may be an average of the best case furnace operating cycles for a particular type of charge.

[0019] The regression model may be configured to determine relationships between x variables and y variables, and may also be used to identify non-significant x variables so that the non-significant x variables may be removed for development of at least one reference cycle and / or other analysis.

[0020] The real-time data from the sensors of the furnace may include charge data, furnace condition data, and operation data. The real-time data may be real-time data within a few seconds, minutes, ten minutes, or 15 minutes of the current time when the furnace is operating. In some embodiments, the real-time data may be data up to 15 minutes or 20 minutes of the current furnace operation time. In other embodiments, the real-time data may be data up to 1 minute or up to 3 minutes of the current furnace operation time. The delay of the network environment or the system environment may affect the closeness of the real-time data to the current operation time of the furnace.

[0021] The method may further include transmitting data communications for adjusting the operation of the furnace to a computer device of an operator such that the operation of the furnace is adjusted to approximate the furnace performance defined by the at least one reference cycle, thereby adjusting the operation of the furnace to approximate the desired performance of the furnace. In other embodiments, the communication of data for adjusting the operation of the furnace may be automatically supplied in a manner that is immediately implemented without operator approval of the change(s). In some embodiments, adjusting the operation of the furnace such that adjusting the operation of the furnace may include transmitting data communications for adjusting the operation of the furnace to a first computer device operably connected to the furnace such that the operation of the furnace is adjusted.

[0022] An embodiment of a non-transitory computer-readable medium having a program stored thereon is also provided. The program may define a method performed by a computer device running the program. The method may include: (i) storing data related to furnace charges, furnace conditions, and operating data of the furnace for multiple different operating cycles of the furnace; (ii) creating or determining one or more cycle performance parameters based on the stored data; (iii) creating a historical data repository based on the stored data parameters and the cycle performance parameters; (iv) characterizing the cycle for generating one or more reference cycles; (v) determining x variables for the one or more reference cycles and feeding the x variables into a regression model to determine a relationship between at least one of the x variables and at least one y variable, thereby defining at least one reference cycle, each The y variable corresponds to a corresponding one of the cycle performance parameters; (vi) when the furnace operates to melt a charge, receiving real-time data from a sensor of the furnace; (vii) comparing the real-time data from the sensor of the furnace with the at least one reference cycle to determine whether one or more furnace operating parameters need to be adjusted; and (viii) when a deviation from the at least one reference cycle is determined based on the comparison of the real-time data from the sensor of the furnace with the at least one reference cycle, adjusting the operation of the furnace so that the operation of the furnace is adjusted to approach a preselected furnace performance target.

[0023] Embodiments of the method defined by the program or the code of the program may include embodiments of the method discussed above or elsewhere herein. For example, in some embodiments, a regression model may determine a relationship between an x ​​variable and a y variable, and also be used to identify non-significant x variables so that the non-significant x variables are removed from the analysis. As another example, the method defined by the program may also include: communicating data for adjusting the operation of the furnace to a computer device of an operator so that the operation of the furnace is adjusted to approximate the furnace performance defined by the at least one reference cycle.

[0024] As another example, a preselected furnace performance target may be defined by the at least one reference cycle. As discussed above, in some embodiments, the preselected furnace performance may be defined by the at least one reference cycle. The at least one reference cycle may be an expected operation of the furnace, which may be based on a model derived from historical data of past performance of the furnace. Such historical data may be stored in a historical data repository that includes empirical data related to past operating cycles of a furnace melting material including metal. In some embodiments, the reference cycle may be considered to be an optimal performance operating curve for the furnace. In other embodiments, the reference cycle may be a preselected furnace performance target based on furnace historical operating data for furnaces having similar configurations (e.g., the same or similar furnace models that melt similar types of materials, etc.).

[0025] A system for controlling the melting of a charge by a furnace is also provided. The system may be configured to implement an embodiment of the method and / or an embodiment utilizing a non-transitory computer-readable medium. In some embodiments, the system may include: a computer device that is communicatively connected to a sensor of the furnace to obtain measurement data during the operation of the furnace. The computer device may be configured to store data related to the charge of the furnace, the furnace condition, and the operating data of the furnace in a non-transitory computer-readable medium that is communicatively connected to a processor of the computer device for a plurality of different operating cycles of the furnace. The computer device may also be configured to create or determine one or more cycle performance parameters based on the stored data, create a historical data repository based on the stored data parameters and the cycle performance parameters, and generate one or more reference cycles. The computer device may also be configured to determine x variables for the one or more reference cycles, and feed the x variables into a regression model to determine a relationship between at least one of the x variables and at least one y variable, thereby defining at least one reference cycle. Each y variable may correspond to a corresponding one of the cycle performance parameters. The computer device may also be configured such that when a deviation from the at least one reference cycle is determined based on a comparison of the real-time data from the sensor of the furnace with the at least one reference cycle, the computer device initiates adjustment of the operation of the furnace such that the operation of the furnace is adjusted to approach a preselected furnace performance.

[0026] In some embodiments of the system, a computer device may be considered a first computer device. The first computer device may be communicatively connected to a computer device of an operator (which may be considered a second computer device). The first computer device may initiate adjustments to the operation of the furnace such that, by sending communications to the computer device of the operator, the operation of the furnace is adjusted to approximate the furnace performance defined by at least one reference cycle. The communications may facilitate input entered by the operator to cause one or more operating parameter adjustments to be changed. In other embodiments, the communications may be sent such that the changes occur automatically without operator input.

[0027] In some embodiments, the computer device can be configured to adjust at least one control parameter for a furnace to initiate adjustment of the operation of the furnace so that the operation of the furnace is adjusted to approach a preselected furnace performance. As discussed above, the preselected furnace performance can be defined by at least one reference cycle. For example, in some embodiments, the preselected furnace performance can be defined by at least one reference cycle. The at least one reference cycle can be the expected operation of the furnace, and the expected operation can be based on a model derived from historical data of the past performance of the furnace. Such historical data can be stored in a historical data repository, which includes empirical data related to past operating cycles of a furnace melting material including metal. In some embodiments, the reference cycle can be considered to be the optimal performance operating curve of the furnace. In other embodiments, the reference cycle can be a preselected furnace performance target based on the furnace historical operating data of a furnace with a similar configuration (e.g., the same or similar furnace model that melts similar types of materials, etc.).

[0028] Embodiments of the system may also include other components. For example, the system may include: a furnace and / or sensors for the furnace. The system may also include other components, such as, for example, one or more program logic controllers for the furnace and / or one or more network nodes (e.g., routers, gateway access points, bridges, etc.) that connect the furnace sensors to a computer device for communicating sensor data to the computer device, etc.

[0029] Other details, objects, and advantages of our controller for a furnace, systems utilizing the controller, and methods of making and using the controller will become apparent as the following description of certain exemplary embodiments proceeds. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Exemplary embodiments of our system for controlling the operation of a furnace, a controller for a furnace, and methods of making and using the same are shown in the drawings included herein. It should be understood that the same reference symbols used in the drawings may identify the same components. For example, in some of the drawings, the term "Temp." refers to temperature.

[0031] Figure 1 1 is a schematic diagram of a first exemplary embodiment of a system for controlling the operation of a furnace. An intermediate node 11 and a communication connection between the intermediate node 11 and the controller 3 and the computer device 10 and the furnace 2 (e.g., a furnace sensor and a programmable logic controller of the furnace, etc.) are provided in Figure 1 Shown by dotted lines.

[0032] Figure 2 is a block diagram illustrating a control method that may be used in an embodiment of a controller of a first exemplary embodiment of a system for controlling the operation of a furnace.

[0033] Figure 3 is a graph illustrating the end of an exemplary melt correction.

[0034] Figure 4 is a graph showing an exemplary correction that can account for metal oxidation (eg, aluminum oxidation).

[0035] Figure 5 is a graph showing that as the yield of the material increases, the effect of dumping temperature on yield loss decreases until about 60% yield, beyond which the yield begins to increase again. Or in other words, the 60% yield material has the least yield loss sensitivity to dumping. Figure 5 The y-axis of the graph of is the exponential coefficient, which is the pour temperature divided by 680°C (exponential coefficient (pour temperature / 680°C)).

[0036] Figure 6 is a graph showing the effect of material yield on melting rate coefficient on yield loss. The graph shows that, on average, as the material yield increases, the sensitivity of yield loss to melting rate also increases.

[0037] Figure 7 is a graph illustrating the extent to which yield loss (y-axis) is affected by the amount of aluminum present in a charge (x-axis) in a first exemplary embodiment of a system for controlling operation of a furnace when the charge including aluminum is being melted.

[0038] Figure 8is a graph showing the relationship between yield loss (y-axis) and tapping temperature (x-axis) when a charge including aluminum is melting, which may exist in a first exemplary embodiment of a system for controlling the operation of a furnace. The graph provides an example showing the importance of tapping temperature to the loss of metal, here aluminum. For example, tapping liquid aluminum at 900°C will result in a loss of molten aluminum compared to 750°C for a charge containing 80% aluminum. Figure 8 An additional yield loss of about 3% is achieved in the example. DETAILED DESCRIPTION

[0039] refer to Figures 1 to 8 , the system 1 for controlling the operation of the furnace 2 may include a controller 3 that is communicatively connected to the furnace's sensors that collect data related to various operating parameters of the furnace. In some embodiments, the controller 3 may be a computer device that is communicatively connected to a plurality of sensors and one or more program logic controllers (PLCs), or other types of controllers that are connected to the furnace to control the operation of the furnace.

[0040] The controller 3 may be directly connected to a PLC or other controller via a directional wired and / or wireless connection, or may be indirectly connected via one or more intermediate nodes 11 included in a network connection (e.g., an Internet connection, a local area network connection, a wireless local area network connection, etc.). For example, the controller 3 may be a remote cloud-based controller 3 connected via an Internet connection, wherein there is at least one type of intermediate node 11 that facilitates a communication connection between the controller 3 and one or more PLCs of the furnace 2, and communicatively connects the controller 3 to one or more sensors of the furnace 2. Each intermediate node 11 may be a bridge, a router, an access point, a gateway, or other type of intermediate node. There may be one or more such intermediate nodes that facilitate a communication connection via a network connection. In some embodiments, the intermediate node 11 may be configured to collect sensor data from sensors of the furnace and then provide the data to the controller 3. Such data collection and transmission may occur periodically (e.g., every few seconds, every minute, at least once every 5 to 10 minutes, etc.). The controller 3 may then use the sensor data to provide one or more control messages to one or more of the PLCs and / or the operator's computer device 10 to facilitate adjusting the furnace operation when the controller 3 determines that the adjustment is approved.

[0041] In other embodiments, the controller 3 may be a computer device within the same enterprise network of the furnace sensor, PLC or other controller, for communicating with the controller and the sensor. In this type of arrangement, there may still be one or more intermediate nodes 11 involved to facilitate such network connections. The intermediate node 11 may include one or more routers, access points or other types of network node components of a local area network.

[0042] In some embodiments, the controller 3 may be communicatively connected to a computer device 10 of an operator who uses the computer 10 to supervise and / or monitor the operation of the furnace. The controller 3 may communicate with the operator's computer device 10 to communicate information about operational adjustments of the furnace implemented by the computer device 10 and / or the operator. For example, the controller 3 may send one or more messages to the operator's computer device 10 to suggest changes to one or more operational set points in an open-loop control scheme implementation for the operator to use to actuate the changes to the one or more set points. Such a communicative connection between the controller 3 and the operator's computer device 10 may be facilitated by at least one application programming interface (API) and / or other communication connection interface. In some embodiments, the messages sent to the operator's computer device 10 may be configured to run with an automated process control program running on the operator's computer device 10 to allow the operator to easily adjust the set points recommended in the messages sent by the controller 3. One or more graphical user interfaces displayed via the operator's computer device 10 may facilitate displaying one or more messages that may be sent by the controller to facilitate input of operator input for accepting one or more recommended adjustments to the operation of the furnace that may be communicated by the controller 3.

[0043] The furnace may include a door 2a that can be opened and closed. The door 2a can be moved between a fully closed position and a fully open position. The door 2a can also be moved to a partially open position for pouring material and / or feeding charge into the chamber of the furnace. For example, the door 2a can be fully opened or at least partially opened so that charge can be positioned in the furnace chamber for melting in the chamber of the furnace. The door 2a can be closed to enclose a chamber 2c for melting metal. The door 2a can be fully opened or at least partially opened to open the chamber for pouring molten metal from the chamber of the furnace and pouring it into at least one container 2v.

[0044] The furnace 2 may also include a flue 2f through which combustion gases may leave the furnace chamber when the material in the chamber is heated by at least one burner 2b. The (multiple) burners 2b of the furnace may be configured to burn fuel by air or other types of gas streams including oxygen fed to the burners for heating metals and other materials in the chamber to melt the materials (e.g., metal to be melted). The flue 2f may be in fluid communication with the chamber so that hot combustion gases may be discharged from the chamber. One or more flue gas treatment mechanisms may be connected to the flue for receiving flue gas so as to treat the gas before the flue gas is discharged from the furnace and / or recycled into the furnace as a component of combustion gas, etc.

[0045] The sensors of the furnace 2 may include temperature sensors, pressure sensors, component detection sensors, operating parameter sensors (e.g., flow rate of flue gas, composition of flue gas, temperature of flue gas, composition of flue gas, etc.). The controller 3 may be communicatively connected to the sensor via a wired connection and / or a wireless communication connection. In some embodiments, the controller may be directly connected to one or more sensors via wiring and an interface 5 of the controller. The controller 3 may also or alternatively be communicatively connected to some or all of the sensors via at least one network. In some embodiments, the intermediate node 11 may store sensor data and then periodically transmit the data to the controller 3, or it may immediately transmit the data to connect the controller 3 to the sensor. In some embodiments, there may be one or more intermediate nodes 11 between the sensor and the controller that facilitate the communication of sensor data to the controller 3 for storage in a memory of the controller 3 or in a memory accessible to the controller 3 (e.g., an example memory of a server that may be connected to the controller 3). The controller 3 and the sensor may be communicatively connected via at least one network (e.g., a local area network, the Internet, an enterprise network, etc.).

[0046] The controller 3 may be a computer device, such as a controller device, a workstation, a server, an array of servers hosting a furnace control service, a desktop computer, a laptop computer, or other type of computer device. The controller 3 may be on-site with the furnace 2, or may be located remotely from the furnace and connected to the furnace sensors and other data input mechanisms via a remote communication connection (e.g., a cloud-based service connectable via the Internet, a web-based service connectable via a wide area network, etc.).

[0047] The controller 3 may include hardware including at least one interface 5 connected to at least one processor (CPU) 7 and non-transitory memory 9. The processor 7 may be a central processing unit, at least one core processor, or other type of hardware processor component. The memory 9 may be a flash memory, a hard drive, a solid-state drive, or other type of memory. Each interface 5 may be configured to facilitate a communication connection (e.g., including a wireless network adapter, an Ethernet network adapter, a near field communication adapter, a Bluetooth adapter, and / or at least one network interface card (NIC), etc.).

[0048] The memory 9 may have stored thereon at least one application (App) 9a and at least one data repository 9b (also referred to in the figures as at least one data repository "DS"). The (multiple) data repository 9b may include data collected over time from one or more of the sensors, program functions, data files, databases, databases containing sensor data, and / or other information. The application 9a may include instructions defining methods that may be performed by the controller 3 when the processor 7 runs the application. The application may include code requiring the processor 7 to utilize information contained in one or more data repositories 9b when running the application 9a. In some embodiments, the controller 3 may be a server hosting a cloud-based furnace control service for communicating to one or more computer devices 10 of a process control system that can monitor and / or supervise the control operation of the furnace. In other embodiments, the controller 3 may be incorporated into or as a computer device of a more locally existing automatic control system.

[0049] The controller 3 may include or be connected to a display device or a computer device 10 having a display device (e.g., a laptop computer with a display, a workstation with a display, a tablet computer with a touch screen display, etc.). Data generated by the controller 3 may be communicated to the computer device 10, which has a display for generating graphical content to be displayed in a user interface, so as to display data to a user (e.g., an operator), which data may be used by the user to monitor the operation of the furnace and / or supervise the operation of the furnace. In some embodiments, the computer device 10 may be part of a workstation running an automated process control program for supervising and / or monitoring the operation of the furnace 2, and the computer device 10 may display a graphical user interface containing data received from the controller 3, for displaying at least a portion of this data with a specific type of graphical element or display element to communicate relevant data about the operation of the furnace to the user for use in monitoring and / or adjusting the operation of the furnace (e.g., identifying the desired steel tapping time, adjusting the temperature set point of the furnace, etc.). The controller 3 may also be configured to communicate data to other communication devices (e.g., a desktop computer, a laptop computer, a smart phone) to provide data related to the operation of the furnace. In some embodiments, the controller 3 may communicate with the operator's or other personnel's computer device 10 via at least one application programming interface (API) or other type of interface for communicating such data.

[0050] The controller 3 may be configured to receive data from the sensors and store the data in at least one data repository 9b for a preselected period of time to store a baseline data set for a preselected number of operating cycles of the furnace 2. Some of the data obtained may also come from non-sensor devices, such as from manual input of data provided by a user using an input device, or from a computer device connected to a scale that may be used to detect the weight of material to be fed to the furnace. In some embodiments, this baseline data set may be stored in at least one data repository 9b by copying data from furnace operation stored in an automatic process control system that may monitor previous cycles of furnace operation, so that this baseline data set is stored and accessible by the controller without any delay in the event that the controller 3 is installed for use with a pre-existing furnace 2. In other embodiments, the baseline data set may include data based on model furnace measurement data collected from similar types of furnace use that are considered applicable to the furnace 2 on which the controller 3 will be used. Once this baseline data set is in a memory accessible to the controller (e.g., the controller's memory, or the memory of a workstation, database server, or other device communicatively coupled to the controller 3 so that the controller can access the data), the controller can be placed in an active state for monitoring and / or controlling furnace operations.

[0051] The data stored in the memory accessible to the controller 3 may include data from sensors and other sources relating to a large number of parameters that may fall into a number of different categories. Figure 2 As can be understood in , these categories can include, for example, charge, furnace condition, and operation. Operational data can be operational data obtained at any point during a continuous operating cycle of the furnace or during batch or semi-batch processing of the furnace 2. This sensor data, as well as other data in these categories, can include all or a selected subset of the following:

[0052] I . Charge:

[0053] a. Type (e.g., scrap, dross, metal, flux);

[0054] b. Weight / mass;

[0055] c. Source;

[0056] d. size and shape (e.g. volume, surface area, length, width, diameter, etc.); and

[0057] e. Composition (e.g., weight % of a particular type of metal, weight % of different elements of the charge, etc.).

[0058] II . Furnace Condition

[0059] a. Type (e.g. rotary, reflective, side-tank, double-chamber, etc.);

[0060] b. Configuration (e.g. one-way, two-way, etc.);

[0061] c. Construction materials (e.g. refractory type, thickness, etc.);

[0062] d. Doors (e.g. type, size, shape, location, number);

[0063] e. Flues (e.g. type, size, shape, location, number);

[0064] f. Burners (e.g., type, size, location, orientation, fuel type, oxygen percentage in oxidant, quantity, etc.); and

[0065] g. Status (e.g. temperature, cleanliness, refractory condition, door condition, furnace fullness, environmental conditions, sealing effect, etc.)

[0066] III . operate

[0067] a. Burner firing rate (high, medium, low, off);

[0068] b. Stoichiometric ratio;

[0069] c. Stack temperature (e.g., measurement, correction for inaccuracies, etc.) and control (e.g., damper or filter / bag house settings, etc.);

[0070] d. Furnace rotation speed and direction (clockwise, counterclockwise, rotation speed in revolutions per second or revolutions per minute, or in other units such as m / s, etc.);

[0071] e. Tilt angle;

[0072] f. Furnace drum rotation resistance (e.g., motor current, hydraulic pressure);

[0073] g.Furnace pressure;

[0074] h. Door position;

[0075] i. Operators and rotation;

[0076] j. Slag (weight, condition, temperature); and

[0077] k. Variability in the fuel or oxidant supplied.

[0078] In addition to the data described above, the data store may include other data. Non-limiting examples of different data that may be stored in a memory accessible to the controller 3 are provided above so that the controller may utilize the data in its operation to manage and / or supervise the operation of the furnace 2.

[0079] As from Figure 2 As will be appreciated, based on this baseline data set, the controller may be configured to process the data to define performance parameters during and after the cycle is complete. Examples of cycle performance parameters may include all or a selected subset of the following:

[0080] Cycle performance parameters

[0081] 1. Specific fuel consumption (SFC), which can be defined as the fuel used for the cycle divided by the total charge weight.

[0082] 2. The stoichiometric ratio, which may be defined as the ratio of fuel to oxygen compared to the ratio of fuel to complete combustion of oxygen in air, substantially pure oxygen, or any combination thereof.

[0083] 3. Metal recovery, which may be defined as the mass of metal recovered (e.g., aluminum, copper, steel, iron, lead, etc.) divided by the mass of metal-containing scrap charged to the furnace. Metal recovery may also be defined as the output from a furnace operation (e.g., the amount of metal recovered from the mass of charge fed to the furnace).

[0084] 4. Salt dosage (mass or percentage of charge).

[0085] 5. Slag quantity (mass or percentage of charge).

[0086] 6. Melt rate, which can be defined as the total charge weight divided by the time from the start of firing to the start of tapping.

[0087] 7. Energy losses (e.g., flue losses, convection losses, radiation losses, Figure 1 heat loss shown in FIG.

[0088] 8. Useful energy (e.g., energy absorbed by the charge to melt the charge).

[0089] 9. Metal oxidation (estimated), e.g. aluminum oxidation, copper oxidation, etc., to make specific metals molten for operation of the furnace.

[0090] In addition to the parameters described above, these cycle performance parameters may include other parameters. Non-limiting examples of different cycle performance parameters are provided above, which are determined by the controller and stored in a memory accessible to the controller 3 so that the controller can use the parameters to manage and / or supervise the operation of the furnace 2 during its operation.

[0091] refer to Figure 2 , the baseline data and cycle performance parameters may be stored in a memory accessible to the controller 3. In addition, the furnace 2 may be operated for some number of cycles, and the controller may store this operating data and the cycle performance parameters of said data in a memory accessible to the controller 3 for subsequent use. The controller may store the data and performance parameters to create a historical data repository 9b.

[0092] The controller 3 may also use a historical data repository so that the data categories and data of the cycle performance parameters for each furnace operation cycle are characterized into a material characterization data repository 9b, which groups furnace charges of similar size, shape, composition and / or origin. The controller may be configured to form and adjust the material characterization data repository 9b from the historical data repository so that anomalous data for each material type utilized from the historical data repository is removed to ensure that the data and performance parameters fall within a specific variance of the mean. The selected variance range for selecting and removing anomalous data may depend on the material type as well as other factors (e.g., the desired range of accuracy of the historical data set, the desired statistical significance of the data set in history, etc.).

[0093] After the anomalous data is removed (for embodiments in which the anomalous data is removed), each material type in the material type data repository 9b can then be processed by the controller to form one or more reference cycle data repositories 9b. The reference cycle data repository can be formed to define at least one reference cycle that demonstrates the pre-selected performance of the furnace using a particular type of charge or an average value of different conditions and materials. In one example, the reference cycle can represent the best performance under specific conditions. In another exemplary example, the reference cycle can represent an average performance or some other benchmark for comparison purposes. For example, the average of the cycles with the highest SFC or the lowest SFC can be used to calculate relative performance. The calculated relative performance can then be used to compare the calculated relative performance with other cycles and with the current cycle performance.

[0094] The material characterization data repository 9b may also be used by the controller to define multiple x-variables for each type of material for regression analysis. The determined x-variables may define multiple different data parameters for furnace conditions and operation for each classified charge type for analyzing the relationship between those data parameters (x-variables) and cycle performance parameters, which may be considered y-variables for a particular type of charge being fed into the furnace.

[0095] Once the x-variables and y-variables are defined, the controller 3 may perform a regression analysis to determine the relationship between the key performance indicators and the data recorded in the historical data repository 9b. In conjunction with this process, x-variables of low significance may be determined to have minimal impact on a particular y-variable in order to update the subsequent use of the regression model and its evaluation of various y-variables.

[0096] The regression model results can then be used to provide real-time adjustments to control furnace operation. For example, the results from the regression model can be used to utilize significant x-variable data obtained in real-time for a particular furnace operating cycle, and the data can be used to calculate desired cycle performance parameter values ​​for improving furnace operation of the cycle. This can be done for one or more cycle performance parameters of the furnace.

[0097] The regression model used by the controller 3 may define a regression analysis that results in the generation of equations for each cycle performance parameter (y variable) to describe the relationship of the y variable to different data parameters (x variables). Key performance indicators (e.g., one or more performance parameters that may each be a corresponding y variable) may be generated for all reasonable data parameters (x1, x2, x3, ... x n ) to understand the statistical significance of each x variable on the y variable. The p-value of the coefficient (A N , B N , C N ...N N ) can be directly related to the statistical significance and relationship of each x variable to the y variable, thereby determining which x variables are statistically significant to the y variable and which are not. The value and sign (+ / -) of the coefficient determines the relative sensitivity of the data parameter to the performance parameter. The regression model can be designed so that it only considers x variables with high significance (e.g., significance equal to or exceeding a pre-selected statistical significance threshold) and therefore ignores x variables with low significance coefficients. An example of a low significance coefficient is one or more x variables that are determined to have a non-significant effect (e.g., an effect below a significance threshold) on the operating parameter y variable of potential interest. Each y variable (Y1, Y2, Y3, ..., Y n Examples of equations for ) may include, for example:

[0098] Y1=A1x1+B1x2+C1x3+…+N1x n

[0099] Y2=A2x1+B2x2+C2x3+…+N2x n

[0100] Y3=A3x1+B3x2+C3x3+…+N3x n

[0101] …

[0102] Y N =A N x1+B N x2+C N x3+…+N N x n

[0103] Nonlinear and correlated terms can also be included in a regression model to help provide high prediction accuracy. Examples of these terms are shown below, where higher value exponents are not shown for simplicity:

[0104] Y1=A1x1+B1x1 2 +C1x1x2+D1x2+E1x2 2 +F1x1x2 2 +G1x1 2 x2+H1x1 2 x2 2 +…

[0105] The regression model may include higher value exponential terms and exponential terms.Other techniques such as partial least squares, principal components, or other techniques may also be used to model performance.

[0106] The regression model can be used by the controller 3 to predict the performance of the furnace 2 based on data related to the operation of the furnace received in real time, and to provide real-time feedback to change the operating parameters of the furnace. The live x variable data recorded during the current cycle can be input into the regression equations of the regression model, and these regression equations are compared with the corresponding reference melting of the data of the historical data repository 9b as the cycle proceeds. The regression model can be used by the controller 3 to identify whether the y variable (e.g., cycle performance parameter) begins to diverge from the reference situation of the determined reference cycle. If divergence is detected, the controller 3 can determine which (which) control parameters should be changed to mitigate the divergence. The regression model can be used to optimize all y variables or any number of individual y variables based on the x variable data of the currently ongoing furnace operation, depending on the data details related to the specific ongoing cycle of the furnace operation. For example, SFC can be optimized. As another example, the yield of the furnace can be optimized. As yet another example, through the controller's adjustment of various control parameters determined by using an embodiment of our process, the yield and SFC can be optimized for a specific cycle based on the data of the x variables of that ongoing cycle of the furnace operation.

[0107] For example, the results of the regression analysis performed by the controller 3 can be used to adjust the furnace operation in real time by using the historical data log and the significance of the data determined using the regression model during the current cycle of the furnace operation, by the controller calculating the relative degree of aluminum oxidation (a cycle performance parameter, the y variable of the regression model) from the real-time sensor data. This aluminum oxidation estimate can then be used in real time during the cycle to provide feedback information to the control loop of the controller 3 to control the furnace operation in order to mitigate aluminum losses and improve production. This approach can also address mass energy balance errors that may occur due to variability in a given charge.

[0108] As from above and Figure 2 It will be appreciated that an embodiment of the control process performed by the controller 3 may include storing data parameters relating to charge material, furnace conditions and operating data parameters in a first step S1. This stored data may then be used to create a historical data repository in a second step S2. The data parameters stored from step S1 are also used to calculate performance parameters which are also recorded in a historical record during the second step S2 to form a historical data repository 9b.

[0109] These cycles are then characterized into one or more material groups in a third step S3, and the one or more material groups are analyzed in a fourth step S4 to calculate the mean and variance of the data and performance parameters. In the fourth step S4, outliers are determined by identifying values ​​that fall outside a specific variance of the mean, which may vary depending on the material type. Outliers can be removed from the data set, and the data from the remaining cycles can constitute the x variables for one or more reference cycles that will be fed into the regression model in the fifth step S5. The y variables can be any data or performance parameters that can be modeled based on the use of the regression model in the fifth step S5 for optimization. For example, specific fuel consumption, melting speed and output. One method of defining a reference cycle for each y variable can be determined by identifying "best case" cycles that convey near-ideal performance. The reference cycle data can be data for only one best reference cycle, or there can be multiple best reference cycles defined based on the charge to be fed to the furnace 2. The reference cycle case can be an average of these "best case" furnace operating cycles for a specific type of charge, or there is only a single best case furnace operating cycle defined for all types of charge. Even if it is not the "best case", there may be other reference cycles that achieve the expected performance. Still other types of reference cycles may be defined that allow the current cycle to be compared to historical performance and allow the controller 3 to utilize real-time sensor data and compare the data to the determined reference cycle to determine how the ongoing cycle operates compared to other cycles. The controller 3 can use this comparison to determine how to adjust the furnace operating parameters (e.g., various set points, etc.) accordingly to achieve one or more cycle goals (e.g., maximize yield, optimize SFC, optimize SFC and also yield, etc.) for processing material during the ongoing cycle.

[0110] In a sixth step S6, a regression model may be used by the controller 3 to determine the relationship between the x-variable and the y-variable, as explained above and herein. Low significance x-variables may also be removed from the analysis. In a seventh step S7, live charge, furnace condition and operating data (x-variables) obtained in real time (e.g., within seconds or minutes of actual ongoing operation of the furnace 2, within a time period of 1 to 15 minutes of data collected from ongoing operation of the furnace 2, etc.) is input into the model so that the controller can calculate performance parameters (y-variables) as the cycle proceeds based on this x-variable data obtained during operation of the furnace. In an eighth step S8, the calculated y-variable (cycle performance parameter) value may be compared in real time with the corresponding y-variable value from a reference case so that the controller can indicate whether and when a particular real-time calculated y-variable deviates significantly from the reference case. If the cycle performance follows the reference conditions within a reasonable error range throughout the cycle (e.g., within + / - 5% or + / - 10% of a preselected performance defined by at least one reference cycle calculated by the controller), the performance is considered to be within expectations and no corrective action is taken or recommended.

[0111] However, once it is determined that the performance of the ongoing cycle begins to deviate and continues to deviate significantly from the reference condition or from the pre-selected desired performance, the controller may determine that the furnace performance has deviated from the pre-selected furnace performance target and an adjustment in operation is required. This determination may be made by the controller based on a comparison of the real-time sensor data of the furnace for the ongoing cycle with at least one calculated reference cycle to determine that the furnace operation has diverged from the reference condition or from the desired performance. For example, the detection of divergence may be made by the controller detecting the presence of a difference from the furnace performance that meets or exceeds a significance threshold based on the comparison thereof based on the real-time sensor data. Such a furnace performance benchmark may be defined by one or more reference cycles calculated by the controller based on historical data in a memory accessible to the controller 3.

[0112] Upon determining that adjustments in furnace operation are warranted, the controller may initiate corrective action by providing adjustment data to the control loop to adjust control parameters of the furnace (such as burner firing rate, furnace rotation speed, furnace tilt angle, etc.) to correct furnace performance so that the corrected furnace x variable data received becomes closer to the desired reference condition(s) or performance. Such initiation may be performed by the controller 3 automatically sending one or more control messages to the furnace's PLC. Such corrective action may also or alternatively be performed by the controller communicating one or more messages to the operator computer device 10 to facilitate the operator providing input through the operator computer device 10 to adjust the set points and / or other operational control parameters determined to be required by the controller 3.

[0113] Exemplary embodiments of the controller 3 and its operation for controlling furnace operation are provided below. These non-limiting examples can help further convey how the controller 3 can be implemented to correct calculations made of the energy required for furnace operation, thereby improving production and other cycle performance parameters of the furnace, enabling the furnace to be operated to improve profitability, and also reduce emissions, fuel usage, electricity consumption, and improve other aspects of the operation of the furnace. Embodiments of the controller 3 can be implemented so that the required energy calculation can be corrected as needed so that the required energy calculation causes the performance of the furnace to approach or reach a defined desired performance level (e.g., the reference case performance level defined in the fifth step S5 of the embodiment of the method discussed above, the desired operation defined by the reference cycle model derived based on historical data from the historical data repository 9b).

[0114] Example 1

[0115] Figure 3 A graph showing a cycle in which the required energy calculation is corrected due to divergence using an embodiment of our method implemented in an exemplary embodiment of the controller 3 is shown. Burner gas flow rate, furnace rotation speed, furnace door position, required energy calculation and flue temperature are shown in FIG. Figure 3 . Initially, the furnace door is open (100%) for charging, with the burners off (0%), the furnace is not rotating (0%), and the flue temperature can be seen to decrease as it moves to equilibrium with ambient conditions. At time 77, the furnace door is closed, and the burners are then turned on to an 80% firing rate, and the furnace is set to a 40% rotation speed. Over time, at approximately 435 time units, the required energy calculation can steadily increase to approximately 65% ​​of the total required energy (percentage completed). At this point, the controller 3 identifies a significant deviation from the reference or expected performance. A significant deviation is determined when a difference that meets or exceeds a significance threshold is detected. The extent of the material composition change error is determined, and the "melting end" calculation is corrected by the controller 3 so that the furnace operating parameters are adjusted to bring the furnace performance back to consistent with the reference situation. At time 435, the required energy calculation can be seen to "jump" from 65% completion to 93%, where the controller's use of the regression model corrects for the amount of aluminum that is lower than the expected composition. In this case, with less aluminum content, there is more aluminum oxide, which does not undergo a phase change at the same temperature as the aluminum in the charge. Therefore, less energy is required to bring the charge to tapping temperature, and therefore, the required energy calculation is corrected for a higher completion percentage.

[0116] Example 2

[0117] Figure 4The example shown in shows how the controller 3 is used to correct for aluminum oxidation (yield) losses when the furnace 2 is used to melt aluminum. In this example, the controller defines a regression equation that relates yield loss (y variable) to data parameters (x variables) to mitigate yield losses during a given cycle of the furnace. Generally, due to the nature of the melting process, it is expected that the aluminum will always oxidize to some extent during the cycle. A regression model is defined for use by the controller 3 to compare the performance of a live cycle to a reference case to determine the extent to which aluminum oxidation has occurred throughout the cycle. In this exemplary case, the reference cycle is considered to have experienced minimal aluminum oxidation. When significant aluminum oxidation is identified, corrective measures are taken, which generally can reduce the firing rate of the burners and / or the rotation speed of the furnace, and sometimes additional flux (salt) is required. Examples of such corrections are shown in FIG. Figure 4 Shown in.

[0118] Figure 4 A cycle is shown where significant aluminum oxidation is detected by the controller 3 and the controller can provide a control signal so that mitigation actions can be implemented. In addition to detecting aluminum combustion, the model also corrects the charge composition as in the previous example to maximize the accuracy and performance of the required energy calculation. Figure 4 As can be seen, burner gas flow rate, furnace rotation speed, furnace door position, required energy calculation and flue temperature are shown. Initially, the furnace door is open (100%) for charging, with the burner closed (0%), the furnace rotates intermittently, and the flue temperature can be seen to decrease as it moves to equilibrium with ambient conditions. At time 54, the furnace door is closed, the burner is then turned on to a firing rate of 78%, and the furnace is set to a rotation speed of 40%. Over time, the "melting end" calculation can be seen to steadily increase to approximately 50% completion at time 290. After this point, the controller 3 determines that aluminum oxidation is significantly diverging from the desired conditions and has to take mitigating actions. After the reduction in firing rate occurs, it can be seen that the required energy calculation method for completion increases at a slower rate due to the reduced energy input. The required energy calculation increases to approximately 70% completion, at which point it can be seen that it jumps to 100% completion at time 425.

[0119] exist Figure 4In this example of , the corrections made are due to deviations in the composition assumptions and the aluminum oxidation approximation y variable. The energy release caused by aluminum oxidation contributes to the energy required to bring the charge to tapping temperature. The mitigating action is to turn off the burner and open the furnace door to verify the required energy calculation. The burner is fired again for 8 time units before the mitigating action stops the burner. The furnace is then rotated without the burner on for 85 time units before the operator opens the furnace door again at time 550. At this point, it can be seen that the required energy is reduced from 100% completion to 99%, with the burner being activated for a shorter period of time to bring the required energy back to 100% completion.

[0120] Example 3

[0121] To help further demonstrate the improved performance that embodiments of controller 3 may provide, refer to Figures 6 to 8 As an example, a sample of nearly 700 cycles from an aluminum tilting rotary furnace is discussed. In the third step S3 of this method of this example, the materials are characterized into 9 different material groups. The expected aluminum content of these material groups ranges from 20 to 80% aluminum. The number of cycles in each material group ranges from 25 to 86 cycles, and the percentage of aluminum in each material does not vary by more than + / -5% of the average aluminum content.

[0122] The y variable selected in step S6 of the embodiment of the method of this example is aluminum oxidation (yield) loss, and the resulting statistical model shows that there are a number of x variable parameters that significantly affect yield loss, including the ratio of salts used, aluminum and oxide content in the charge, melting rate, metal tapping temperature, and a number of operating and furnace conditions. The controller 3 uses a regression equation that was found to have an average R of 0.76. 2 Value (and R 2 The R value is a well-known statistical measure of how close the data are to the fitted regression line. It is also called "R squared", the coefficient of determination, or the multiple coefficient of determination for multiple regression). For each material group, R 2 The values ​​range between 0.62 and 0.89, and the effect of different x variables on the y variables may vary between different material groups. Figures 5 to 7 Shown in.

[0123] For example, Figure 5 is a graph showing that as the yield of the material increases, the effect of pouring temperature on yield loss decreases until about 60% yield, beyond which the yield begins to increase again. Or in other words, the 60% yield material has the least yield loss sensitivity to pouring temperature.

[0124] Figure 6is a graph showing the effect of material yield on melting rate coefficient on yield loss. The graph shows that, on average, as the material yield increases, the sensitivity of yield loss to melting rate also increases.

[0125] Figures 5 and 6 It helps to show that many of the variables are interrelated with each other and therefore a partial least square (PLS) regression model using independent predictors (x variables) created from the supplied set of variables may be more suitable for the purpose of modeling yield loss for use by the controller 3. In the final regression model, containing linear, higher order and crossed variable terms for these basic predictors, only three principal components were sufficient to account for 78% of the data variability. Additionally, the linear regression coefficients constructed by the model have the same sign as in the simple group-specific linear regression model, which helps to show that the effects of the selected variables are significant.

[0126] Example 4

[0127] Figure 7 and Figure 8 Control results obtained from an exemplary embodiment of a controller 3 used in conjunction with a tilting rotary furnace 2 for remelting a charge including aluminum-containing dross material are shown. Furnace cycle data from a large number of cycles were analyzed to determine aluminum oxidation losses over a large number of cycles, and regression analysis performed showed an exponential relationship between yield loss and metal tapping temperature. Baseline data was compared with data from cycles completed using an embodiment of the controller 3, which showed that an accurate model end of melt prediction combined with timely burner shut-off control could be provided by the controller 3. This can provide significant benefits in production and yield, as well as energy savings.

[0128] The controller 3 is able to calculate when the aluminum inside the furnace 2 reaches the tapping temperature and predict when the melting part of the cycle is complete. The results discussed herein show that this combined with timely shut-off of the burners prevents overheating of the aluminum, resulting in less aluminum being burned, which results in increased production. Furthermore, it was found that the system used in Example 4 provides a reduction in overall cycle time and energy consumption, which increases productivity while also reducing carbon emissions.

[0129] The embodiment of our controller 3 used for Example 4 was included on a dual pass 12 metric tonne (MT) tilt rotary furnace (TRF) burning oxygen-natural gas processing aluminum slag with different compositions. The furnace and burner system were all controlled by an Allen Bradley programmable logic controller (PLC), the charge was loaded using a manual system, and the liquid aluminum was batched into casting molds (example of container 2v).

[0130] The user interface and communication panel are installed locally. The communication system consists of a central data collection device, wired and wireless communications, and internet connectivity. Data is collected multiple times per minute from different parts of the system using various methods based on how the data is accessible. Modbus TCP is used to transmit data that is digitally available from the PLC to the central data collection device. Additional sensors are installed and connected directly to the communication panel using standard wired protocols or wirelessly using Zigbee. Once the data is centralized, it is encrypted and securely transmitted over the internet to a cloud system for remote processing, storage, and access. Once the data is processed, real-time results are sent back to the user interface for operator access and to assist in decision making.

[0131] In addition to the immediate results available to operators via a human machine interface (HMI) (e.g., a graphical user interface on a display of the computer device 10), managers can also access real-time and historical data via a web browser-based dashboard accessible on desktop and mobile devices. Figure 1 1 and 2. Other examples of computer devices 10 that may be communicatively connected to the controller 3 are shown in FIG. Additional analytical reports summarizing the previous day's activity in terms of material processing, tapping temperatures, cycle times, etc. are sent to plant management daily. For historical purposes, these results are also compiled and archived on a monthly basis.

[0132] The controller 3 is configured to utilize a real-time mass energy balance based on historical and live data. To help ensure the model accuracy of the regression model used by the controller 3, the data in the historical data repository 9b includes data over a sufficiently long period of time to capture the variability of the process. The mass inputs and outputs, including charge size, composition, and weight, are recorded for each cycle, as well as the outputs of liquid aluminum and salt slag for multiple operating cycles of the furnace 2. Based on the energy input through the burners, furnace losses, and energy absorbed into the refractory (see, e.g. Figure 1 The useful energy entering the process is calculated using the losses identified in

[0133] Combustion efficiency is calculated using the burner gas flow rate, which is measured directly from a mass flow meter. Furnace energy losses, including shell losses, door losses, and flue losses, are all measured using temperature sensors. Liquid aluminum temperature is measured during tapping, where the tapping temperature is fed back into the model to verify the mass-energy balance calculation. Over time, as more and more data is collected, the accuracy of the model is improved. For this study, the digital twin model was designed to reduce variability and improve yield. The technology can also be applied to focus on improving throughput and productivity.

[0134] For the controller of Example 4, we assume that during the secondary aluminum remelting, a portion of the aluminum in the charge inevitably oxidizes and generates heat, which increases the energy supplied from the burner and contributes to the energy required for melting. The energy generated by aluminum oxidation and the resulting aluminum loss can be large, depending on a variety of factors, including material composition, salt usage, burner firing rate, and furnace conditions. Aluminum oxidation losses can be referred to as yield losses, which are expressed as a percentage, where the weight of aluminum oxidized is divided by the total weight of aluminum present in the charge. In the embodiment of Example 4, the extent of aluminum oxidation during the cycle is understood by analyzing the energy balance, thereby estimating the yield loss for each cycle. Using this yield loss estimate, a thorough analysis of the important parameters that affect yield loss is performed using regression methods (see, e.g. Figure 2 Steps S5 to S6 of the embodiment shown in FIG.

[0135] A number of statistical methods were used to understand the impact of the different factors that affect aluminum oxidation (yield loss). Initially, a number of reasonable predictors such as charge weight, melting time, salt ratio, etc. (x variables) were used as inputs in the statistical model for the controller 3 to determine their significance for yield loss (the selected y variables). The coefficients for each x variable in the statistical model that were directly related to their effect on yield loss and any terms with low statistical relevance were eliminated (see e.g. Figure 2 The resulting statistical model shows that there are a number of parameters that significantly affect yield loss, including the ratio of salt used, the aluminum and oxide content in the charge, the melting rate, the metal tapping temperature, and the furnace condition, where the furnace condition is a calculated parameter that compares the furnace condition to the previous cycle. The results are R using data from 9 different groups of primary materials for a total of 674 cycles. 2 The model with a value of 0.83. The sign and magnitude of the coefficient of each x variable parameter was studied to better understand its influence on aluminum oxidation (selected y variables) and in particular two factors were found to be highly influential, namely, aluminum content of the charge and tapping temperature.

[0136] Figure 7 How the amount of aluminum present in the charge (x-axis) determined in Example 4 affects yield loss (y-axis) is shown. Yield loss is the amount of aluminum that is oxidized for a given cycle, expressed as a percentage of the aluminum present in the charge. The graph shows that, with all other predictors held constant, as the percentage of aluminum in the charge increases, the chance of aluminum oxidation decreases, and vice versa. Thus, when processing material with an aluminum content of 20%, yield loss can increase by 7.5% compared to material with an aluminum content of 90%.

[0137] This can be attributed to a number of physical factors. As the percentage of aluminum present in the charge decreases, the presence of oxides increases, and therefore salts are often added to the process. The energy input into the process is then used to heat the oxides and salts, as well as the aluminum. The aluminum dross material traps the aluminum in an oxide layer, which requires a large amount of energy to release by breaking down the oxide layer. The result is that more energy per unit of aluminum is required to heat the charge, resulting in overheating of the aluminum trapped in the dross. Furthermore, the aluminum is released from the oxide layer as small droplets, which have a high surface area to volume ratio compared to a solid block of aluminum that can melt more evenly, providing more opportunities for oxidation to occur.

[0138] Figure 8 Shown is the relationship between the yield loss (y-axis) and the tapping temperature (x-axis) determined in Example 4. For the embodiment of the controller used in Example 4, an exponential model was found to be a better fit than a simple linear model. Figure 8 It helps to illustrate the importance of the tapping temperature on aluminum loss identified in Example 4. For example, for a charge with 80% aluminum content, tapping liquid aluminum at 900°C instead of 750°C can result in an additional yield loss of about 3%.

[0139] In Example 4, Controller 3 was able to reduce the tapping temperature, thereby increasing aluminum production, by better predicting when the material inside the furnace was ready to tap. Initially, baseline data was collected in a first step S1 over a seven month period prior to the implementation of Controller 3 in order to record and assess process variability and to form a historical data repository in a second step S2. The majority of the material charged into the furnace was dross, which was characterized as a material type based on size, composition, and origin in a third step S3. Over 1,000 cycles were analyzed to determine the average tapping temperature. After the implementation of Controller 3, 350 cycles were analyzed, with Controller 3 monitoring and / or adjusting furnace control parameters based on real-time data for those cycles, and the overall average tapping temperature was successfully reduced by 25°C, with furnace operation for melting some materials improving more than others. Based on Figure 8 In the graph in , the average reduction in temperature corresponds to a 0.30% yield improvement. Some materials have increased by more than 50°C, corresponding to a 0.5% to 0.75% yield improvement.

[0140] Improvements in furnace operation can provide a reduction in melting time, a reduction in energy usage (e.g., fuel costs), and an increase in production. This type of improvement can have a significant impact on the profitability of furnace operation. For example, for a furnace with a capacity of 10 million tons, melting 3 million tons per hour, and operating for 48 weeks of the year, and having an estimated annual production of 24 billion tons per year, an embodiment of the controller can reduce the melting time by 10%, reduce energy usage (e.g., reduce the amount of oxygen and natural gas used for the operation of burner 2b) by at least 5%, and increase production by 0.5%. We estimate that this improvement will result in an increase in the profitability of the operation of such a furnace of approximately $580,000 per year, or more than £461,000 (British pounds sterling) per year. This example is provided to help quantify the significance of the operational improvement embodiments of the controller 3 and embodiments of our furnace operation method. Different furnaces of different sizes and types may achieve greater improvements than this exemplary improvement, or may not experience much greater improvements than this exemplary improvement, depending on many factors (e.g., size, operating capacity, value of metals being processed in the charge, etc.).

[0141] By focusing the controller 3 in Example 4 on reducing the tapping temperature and increasing the output, there is a significant reduction in the variability of the tapping temperature, resulting in more consistent tapping temperatures and melting times. As a knock-on effect, energy usage is reduced by an average of 5% compared to the baseline cycle of stored data collected before the controller 3 was implemented, because the burners were shut down in a more timely manner. More consistent melting times can also lead to improvements in productivity, which can be seen in some furnace charges, where the melting time is reduced by up to 10.5%, and even higher in other cases where material properties vary significantly. The reduced variability in tapping temperature allows for more precise control of the furnace operation, as well as the ability to reduce the tapping temperature even more, leaving room for further optimization in the future.

[0142] We believe that the results from Example 4 are consistent with how aluminum readily oxidizes in air and how quickly a thin layer of aluminum oxide forms around newly exposed solid aluminum surfaces. Therefore, aluminum likely always contains at least a small amount of aluminum oxide. Once the thin layer of aluminum oxide has formed, the aluminum is protected from further oxidative attack. However, as the aluminum is heated, it becomes more susceptible to oxidative attack because the aluminum oxide "shell" begins to react with the atmosphere and decompose, forming cracks in the protective aluminum oxide outer shell. At higher temperatures and during the melting stage in the furnace, we determined that the oxidation rate may increase as oxygen further penetrates into the metal due to the changing crystal structure of the aluminum oxide with increasing temperature.

[0143] In tilt-rotary furnace processing of aluminum dross and scrap, the oxide layer may not remain intact during the high temperature and melting phase. The constant movement of the furnace can provide a mechanical stimulus to break up the oxide layer. Also, salt is often added to the process, providing a chemical means of breaking down the oxide layer. The purpose of adding salt is to extract as much liquid aluminum as possible by allowing the aluminum to escape from the oxide layer. Once the aluminum becomes liquid, the oxide and salt float to the surface, providing a physical barrier to protect the molten aluminum from oxidative attack. However, the constant movement and mixing can continually break down the protective layer of oxide and salt, leading to further oxidation.

[0144] We determined that at temperatures above the melting point of aluminum (660°C), the oxidation reaction rate can be significantly accelerated. We found that an exponential increase in oxidation rate with increasing temperature can occur, indicating that lowering the tapping temperature can increase aluminum production. Therefore, by focusing the embodiment of the controller 3 used in Example 4 on lowering the tapping temperature, aluminum production can also be increased by preventing the metal from overheating and being kept at a higher temperature for a long period of time.

[0145] Examples 1 through 4 are exemplary in nature, as are the other embodiments discussed herein. It should be understood that different embodiments of our controller 3 may be used to be incorporated into a system 1 for use with a furnace 2 to adjust the operation of the furnace 2 as the furnace melts a metal (such as aluminum, lead, iron, steel, or copper), which may be determined to be necessary to control the operation of the furnace so that one or more control parameters may be dynamically adjusted to better meet a particular set of cycle performance parameters (e.g., yield, yield and SFC, yield, melt rate and SFC, etc.). Embodiments of our system 1, controller 3, and method of controlling furnace operation may be used to help improve yield and other performance metrics. Embodiments of our controller, system, and method may employ significantly different approaches for determining adjustments to furnace operation that may uniquely address mass and energy balance calculation errors that may exist when melting metal materials where the metal content of the material varies significantly or is otherwise not well understood.

[0146] Embodiments of our method, controller 3, and system 1 can provide a unique way to improve process performance for the furnace 2. Key performance and operating parameters can be recorded and used over time so that historical and real-time data can be used to determine when the metal being melted in the furnace 2 has reached a target temperature. Real-time feedback can be provided to the operator with information about when the metal is ready to be tapped, so that the opening of the door 2a and the feeding of sufficient molten metal into one or more containers 2v can be done more consistently, and production can be improved while reducing emissions and lowering costs. As new data continues to be fed into the control for updating the historical data repository and updating the regression model in steps S2 to S6, the controller's predictions and the control parameter adjustments provided by the controller 3 can be further improved over time to provide even better improvements in furnace operation.

[0147] It should be understood that the embodiments explicitly shown and discussed herein can be modified to meet a specific set of design goals or a specific set of design criteria. For example, the arrangement of valves, pipes, sensors, controllers, communication connection components (e.g., wiring, intermediate network nodes, network layout, etc.), furnaces, sensors, and other components can be arranged, sized, and designed to meet a specific plant layout design that takes into account the available area of ​​the plant, the specific sensor array, controller hardware, and furnace operating requirements, as well as other design considerations. It should be understood that embodiments of system 1 can be configured to include various process control components that are positioned and configured to monitor and control operations (e.g., temperature sensors, pressure sensors, flow sensors, target component concentration sensors, an automated process control system having at least one workstation (which includes a processor, non-transitory memory, and at least one transceiver for communicating with the sensor component), valves, and a controller for providing a user interface for an automated process control system that can be run at a workstation and / or another computer device, etc.).

[0148] As another example, it is contemplated that specific features described, whether alone or as part of an embodiment, may be combined with other individually described features or parts of other embodiments. Thus, the elements and actions of the various embodiments described herein may be combined to provide additional embodiments. Thus, while certain exemplary embodiments of a system for controlling the operation of a furnace, a controller for a furnace, and methods of making and using the same have been shown and described above, it should be clearly understood that the invention is not limited thereto, but may be otherwise variously embodied and practiced within the scope of the following claims.

Claims

1. A method for controlling the operation of a furnace to melt a material comprising metal, the method comprising: storing data parameters related to charge material, furnace conditions, and operating data about the furnace in a non-transitory computer readable medium of a computer device for a plurality of different operating cycles of the furnace; creating or determining one or more cycle performance parameters based on the stored data parameters; creating a historical data store based on the stored data parameters and the cycle performance parameters; determining x variables for one or more operating cycles of the furnace from the historical data store and feeding the x variables into a regression model to determine a relationship between at least one of the x variables and at least one y variable to define at least one reference cycle, each y variable corresponding to a respective one of the cycle performance parameters; receiving real-time data from sensors of the furnace; comparing the real-time data from the sensors of the furnace to the at least one reference cycle to determine whether adjustments to one or more furnace operating parameters are required; Upon determining, based on a comparison of the real-time data from the sensor of the furnace with the at least one reference cycle, that there is a difference from the at least one reference cycle that meets or exceeds a significance threshold, adjusting the operation of the furnace so that the operation of the furnace is adjusted to approach a preselected furnace performance defined by the at least one reference cycle.

2. The method of claim 1 , comprising generating the at least one reference cycle, the generating of the at least one reference cycle comprising: The operating cycle of the furnace is characterized into one or more material groups for use in generating the at least one reference cycle for desired performance of the furnace.

3. The method of claim 1 , wherein the determining of the x variables for the one or more reference cycles comprises: Before feeding the data parameters to the regression model, outlier data, which are data parameters that fall outside a certain variance from the mean, are removed from the material population.

4. The method of claim 1, wherein the at least one y variable comprises: A first y variable for specific fuel consumption, a second y variable for melt rate, and a third y variable for production.

5. The method of claim 1, wherein the at least one reference cycle for each y variable is determined by identifying one or more best case operating cycles of the furnace. 6 . The method of claim 1 , wherein the at least one reference cycle is a single optimal reference cycle or comprises a plurality of optimal reference cycles defined based on the charge to be fed to the furnace.

7. The method of claim 6, wherein each reference cycle is an average of best case cycles for furnace operation for a particular type of charge.

8. The method of claim 1 , wherein the regression model determines a relationship between an x ​​variable and a y variable and is further used to identify non-significant x variables such that the non-significant x variables are removable, the non-significant x variables being x variables determined to have an effect below a significance threshold.

9. The method of claim 1, wherein the real-time data from the sensors of the furnace includes charge data, furnace condition data, and operational data.

10. The method according to claim 1, comprising: Data for adjusting operation of the furnace is communicated to an operator's computer device such that operation of the furnace is adjusted to approximate furnace performance defined by the at least one reference cycle such that operation of the furnace is adjusted to approximate desired performance of the furnace.

11. The method of claim 1 , wherein said adjusting the operation of the furnace such that adjusting the operation of the furnace comprises: Data communications for regulating the operation of the furnace are transmitted to a first computer device operably connected to the furnace such that the operation of the furnace is regulated.

12. The method of claim 1 , wherein the preselected furnace performance is defined by the at least one reference cycle, wherein adjusting the operation of the furnace to approximate the preselected furnace performance defined by the at least one reference cycle comprises adjusting the operation of the furnace using the calculated desired cycle performance parameter value.

13. A non-transitory computer-readable medium having a program stored thereon, the program defining a method performed by a computer device running the program, the method comprising: storing data parameters relating to a charge to a furnace, furnace conditions, and operational data relating to the furnace for use in a plurality of different operating cycles of the furnace; creating or determining one or more cycle performance parameters based on the stored data parameters; creating a historical data store based on the stored data parameters and the cycle performance parameters; characterizing the operating cycle of the furnace for use in generating one or more reference cycles; determining x variables for the one or more reference cycles from the historical data store and feeding the x variables into a regression model to determine a relationship between at least one of the x variables and at least one y variable to define at least one reference cycle, each y variable corresponding to a respective one of the cycle performance parameters; receiving real-time data from sensors of the furnace as the furnace operates to melt charge; comparing the real-time data from the sensors of the furnace to the at least one reference cycle to determine whether adjustments to one or more furnace operating parameters are required; Upon determining, based on a comparison of the real-time data from the sensor of the furnace with the at least one reference cycle, that there is a deviation from the at least one reference cycle, adjusting the operation of the furnace such that the operation of the furnace is adjusted to approach a preselected furnace performance target defined by the at least one reference cycle.

14. The non-transitory computer readable medium of claim 13, wherein the regression model determines a relationship between an x ​​variable and a y variable and is further used to identify non-significant x variables such that the non-significant x variables are removed from the analysis, the non-significant x variables being x variables determined to have an effect below a significance threshold.

15. The non-transitory computer readable medium of claim 13, wherein the method further comprises: Data communications for adjusting the operation of the furnace are transmitted to a computer device of an operator such that the operation of the furnace is adjusted to approximate the furnace performance defined by the at least one reference cycle.

16. The non-transitory computer-readable medium of claim 13, wherein the preselected furnace performance target is defined by the at least one reference cycle, wherein adjusting the operation of the furnace to approach the preselected furnace performance target defined by the at least one reference cycle comprises adjusting the operation of the furnace using the calculated desired cycle performance parameter value.

17. A system for controlling the melting of charge by a furnace, comprising: a computer device communicatively coupled to sensors of the furnace to obtain measurement data during operation of the furnace; The computer device is configured to store data parameters related to charge material, furnace conditions, and operating data about the furnace for a plurality of different operating cycles of the furnace in a non-transitory computer-readable medium communicatively connected to a processor of the computer device; The computer device is configured to create or determine one or more cycle performance parameters based on the stored data parameters, create a historical data store based on the stored data parameters and the cycle performance parameters, and generate one or more reference cycles; the computer device being configured to determine x variables for the one or more reference cycles from the historical data store and feed the x variables into a regression model to determine a relationship between at least one of the x variables and at least one y variable to define at least one reference cycle, each y variable corresponding to a respective one of the cycle performance parameters; The computer device is configured such that when a deviation from the at least one reference cycle is determined based on a comparison of real-time data from the sensor of the furnace with the at least one reference cycle, the computer device initiates adjustment of the operation of the furnace so that the operation of the furnace is adjusted to approach a preselected furnace performance defined by the at least one reference cycle.

18. The system of claim 17, wherein the computer device is a first computer device communicatively coupled to an operator's computer device, and the first computer device initiates adjustment of the operation of the furnace such that the operation of the furnace is adjusted to approximate the furnace performance defined by the at least one reference cycle by sending communications to the operator's computer device.

19. The system of claim 17, wherein the computer device adjusts at least one control parameter for the furnace to initiate adjustment of the operation of the furnace such that the operation of the furnace is adjusted to approach the preselected furnace performance, the preselected furnace performance being defined by the at least one reference cycle.

20. The system of claim 17, further comprising the furnace and sensors of the furnace.

Images