System and method for optimizing material processing operations

Abstract

This invention provides a system and method for optimizing material processing operations. [Solution] A control system for a material processing application includes a sensor device and a controller. The sensor device is configured to receive actual motion data and actual process data from the equipment performing the material processing application. The controller determines adjusted motion parameters and adjusted process parameters based on the actual motion data and actual process data, and includes logic for executing the adjusted motion parameters and adjusted process parameters to achieve performance targets.

Description

Technical Field

[0001] Cross - reference to Related Applications This application claims the benefit of U.S. Provisional Patent Application No. 63 / 706,942, filed Oct. 14, 2024, and U.S. Non - Provisional Patent Application No. 19 / 321,245, filed Sep. 7, 2025.

[0002] This disclosure relates to systems and methods for material processing operations. More particularly, it relates to a control system configured to receive sensor data and determine operation parameters adjusted based on the sensor data via an artificial intelligence (AI) model to achieve performance goals of the material processing operations.

Background Art

[0003] Material handling operations (i.e., arc welding, laser welding, thermal cutting, laser cutting, plasma cutting, etc.) require repeatability to ensure that the parts or products produced therefrom are of consistent quality. Many material handling operations are automated to improve repeatability and increase product throughput. For example, many welding and thermal cutting applications use robots that operate a welding gun to perform one or more welds on the materials to be joined, or robots that operate a cutting torch to perform one or more cuts on a target material. Such applications may also utilize other forms of automation, such as material handling robots for positioning one or more workpieces on fixtures, tables, or positioners configured to hold the workpieces in place during such operations. However, automating such operations does not always achieve the desired repeatability or product throughput due to external factors, such as the work environment (e.g., temperature, air quality, etc.), the status of the equipment used (e.g., wear or component failure), or deviations in the material being processed. For example, one workpiece may differ from another workpiece of the same type dimensionally, metallurgically, or otherwise, based on differences in roughness (e.g., unevenness) or the presence of contaminants on it. Furthermore, deviations in the supplies, components, and accessories required to perform such operations may exist, such as cutting aid gases or shielding gases, welding or cutting consumables, or welding or cutting gun / torch components used. In many cases, such deviations may not be discernible to the human eye, and addressing such problems would require extensive downtime that hinders productivity, for example, by reprogramming the robot's planned motion path or adjusting process parameters (e.g., welding or cutting parameters). Moreover, these problems can be exacerbated by other issues, such as misalignment (e.g., misalignment of tack welds between workpieces being processed). Therefore, a solution is needed that takes such deviations into account in real time and ensures that the desired target performance attributes of the material handling operation (e.g., throughput, operating cost, repeatability) are achieved. [Overview of the Initiative] [Means for solving the problem]

[0004] The following summary provides a simplified overview to give a basic understanding of some aspects of the devices, systems, and / or methods described herein. This summary is not a comprehensive overview of the devices, systems, and / or methods described herein. It is not intended to identify key elements or to describe the scope of such devices, systems, and / or methods. Its sole purpose is to present some concepts in a simplified form as a prelude to the more detailed descriptions that will be presented later.

[0005] According to one aspect of the present disclosure, a control system for a material processing application includes a sensor device and a controller. The sensor device is configured to acquire actual motion data and actual process data derived from an instrument performing the material processing application. The controller includes logic for determining adjusted motion parameters and adjusted process parameters based on the actual motion data and actual process data via an artificial intelligence model, and for executing the adjusted motion parameters and adjusted process parameters via the instrument to achieve performance targets.

[0006] In one embodiment, the artificial intelligence model is a neural network comprising an input layer, at least one hidden layer, and an output layer.

[0007] In one embodiment, the sensor device is configured to receive workplace data during a material handling application, and the controller includes logic for determining adjusted motion parameters and adjusted process parameters based on actual motion data, actual process data, and actual workplace data.

[0008] In one embodiment, the workplace data includes at least one of workpiece data, supply data, or environmental data.

[0009] In one embodiment, the apparatus includes at least one of a robot, a torch, a positioner, a power supply, a laser source, a cutting table, a cutting torch, a plasma source, or an oxygen / fuel source.

[0010] In one embodiment, the control system includes an image capture device configured to extract spatial data, and the controller includes logic to determine workpiece information based on the spatial data.

[0011] In one embodiment, the control system includes an image capture device configured to extract spatial data, and the controller is configured to determine planned motion parameters and planned process parameters via the spatial data.

[0012] In one embodiment, the controller includes logic for generating generated content that embodies a work command, based on adjusted motion parameters or adjusted process parameters.

[0013] In one embodiment, the controller includes a timestamp component configured to receive and timestamp actual process data and actual motion data for later retrieval.

[0014] In one embodiment, the sensor device is configured to receive workplace data during a material handling application, and the controller includes a replay component configured to reproduce actual process data, actual motion data, and workplace data in a time-series and temporally synchronized manner.

[0015] In one embodiment, the controller includes logic for generating a composite dataset from actual process data, actual motion data, and workplace data.

[0016] In one embodiment, the controller also includes an emulator component configured to generate a digital twin of a materials processing application based on a synthetic dataset.

[0017] In one embodiment, the controller is communicatively coupled to at least a first input device and a second input device. The first input device is operable via a first programming language including first motion or process commands, and the second input device is operable via a second programming language including second motion or process commands. The controller also includes a decoding component configured to decode the first and second programming languages ​​and to commonize them into a common programming language used to execute a material handling application.

[0018] In one embodiment, the common programming language includes planned motion parameters and planned process parameters.

[0019] In one embodiment, the controller also includes motion trajectory components and process trajectory components. The motion trajectory components are configured to receive motion commands related to planned motion parameters, and the process trajectory components are configured to receive process commands related to planned process parameters. The motion trajectory components are also configured to define motion commands over time, and the process trajectory components are configured to define process commands over time.

[0020] In one embodiment, the artificial intelligence model includes a first mathematical model and a second mathematical model. The controller uses the first mathematical model to determine motion parameters adjusted based on actual motion data, and uses the second mathematical model to determine process parameters adjusted based on actual process data.

[0021] According to another embodiment, the material handling system includes a robot, a welding source, and a controller. The controller is operably connected to the robot and the welding source. The controller includes logic to receive sensor data from the robot and the welding source in real time and to generate process parameters and motion parameters based on the sensor data via an artificial intelligence model. The controller also includes logic to transmit the process parameters and motion parameters to the welding source and the robot in order to achieve performance targets.

[0022] In one embodiment, the sensor data includes actual process data and actual motion data.

[0023] In one embodiment, the artificial intelligence model includes a deep learning neural network comprising an input layer, at least one hidden layer, and an output layer. The input layer receives sensor data, and the output layer generates tuned process parameters and tuned motion parameters to achieve performance targets.

[0024] According to another embodiment, the material processing method includes receiving actual process data and actual motion data from a welding or cutting machine, and generating adjusted process parameters and adjusted motion parameters based on the actual process data and actual motion data via a controller. The method also includes transmitting the adjusted process parameters and adjusted motion parameters to the welding or cutting machine in real time.

[0025] In one embodiment, the welding equipment includes at least one of a robot, a laser source, a power supply, a welding torch, and a positioner, and the cutting equipment includes at least one of a cutting torch, a laser source, a cutting table, a cutting torch, a plasma source, or an oxygen / fuel source.

[0026] The above and other aspects of the present invention will become apparent to those skilled in the art to which the present invention pertains upon reading the following description with reference to the accompanying drawings.

Brief Description of the Drawings

[0027] [Figure 1] Shows a schematic block diagram of a control system for a robotic welding operation according to one embodiment. [Figure 2] Shows a weld joint shown in relation to an exemplary sensor in the form of an image capture device according to one embodiment. [Figure 3] Shows a schematic block diagram of a control system for a material handling application according to a further embodiment. [Figure 4] Shows a schematic block diagram of a control system for a material handling application according to another embodiment. [Figure 5] Shows an exemplary welding robot disposed around an exemplary hexapod actuator according to one embodiment. [Figure 6] Is a flowchart showing an exemplary method of optimizing operating parameters to achieve performance goals according to one embodiment. [Figure 7] Is a flowchart showing an exemplary method of optimizing process and motion parameters to achieve performance goals for a welding application according to one embodiment.

Modes for Carrying Out the Invention

[0028] This disclosure relates to material handling operations, and more particularly to a control system configured to optimize operating parameters (e.g., process parameters and motion parameters) for such operations to achieve desired performance attributes (e.g., desired productivity or repeatability) based on sensor data extracted from the material handling operations in real time. While specific embodiments of this disclosure are described in the context of automated gas metal arc welding (GMAW) applications, this disclosure is not limited thereto. For example, specific embodiments may also be applied to semi-automatic GMAW applications, or other material handling applications including, but not limited to, semi-automatic or fully automated gas tungsten arc welding (GTAW), flux core arc welding (FCAW), metal core arc welding (MCAW), additive 3D manufacturing, surface hardening, laser welding, cladding, thermal cutting processes (e.g., mechanized plasma cutting, laser cutting, oxygen fuel cutting), and assembly operations, such as robotic pick-and-place operations. It is also considered that various inventions disclosed herein may be adapted for use in other types of manufacturing operations such as machining (e.g., CNC machining), water jet cutting, casting, plastic injection molding, and equivalents.

[0029] This disclosure is described herein with reference to the drawings, and similar reference numbers are used throughout to refer to similar elements. It should be understood that the various drawings are not necessarily drawn to scale from one drawing to another, or within a given drawing, and in particular, the sizes of components are drawn arbitrarily to facilitate understanding of the drawings. The following description includes numerous specific details for illustrative purposes to provide a complete understanding of this disclosure. However, it may be apparent that this disclosure can be implemented without these specific details. Additionally, other embodiments of this disclosure are possible, and the various inventions disclosed herein can be implemented and performed in ways other than those described. The terms and expressions used in describing this disclosure are adopted for the purpose of facilitating understanding of the various inventions disclosed herein and should not be construed as limitations.

[0030] As used herein, “at least one,” “one or more,” and “and / or” are open-ended expressions that are both conjunctive and disjunctive in their operation. For example, each of the expressions “at least one of A, B, and C,” “at least one of A, B, or C,” “one or more of A, B, and C,” and “A, B, and / or C” means A only, B only, C only, A and B together, A and C together, B and C together, or A, B, and C together. Any disjunctive word or phrase that presents two or more alternative terms should be understood, whether in the description of embodiments, claims, or drawings, to be intended to include the possibility of including one of the terms, either of the terms, or both of the terms. For example, the phrase “A or B” should be understood to include the possibility of “A” or “B” or “A and B.”

[0031] As used herein, “controller” refers to the logic circuits and / or processing elements and associated software, programs, or artificial intelligence models used to perform the methods and systems disclosed herein. A controller may include processors and memory devices and may take various forms, such as workstations, servers, computing clusters, blade servers, server farms, or any other data processing system or computing device. Due to the constantly changing nature of computing devices and networks, the descriptions of various examples of controllers described herein and shown in the relevant drawings are intended only as examples to illustrate some embodiments. Many other configurations of controllers are considered to be within the scope of this disclosure.

[0032] As used herein, “components” may be defined as parts of hardware, parts of software, or a combination thereof. Parts of hardware may include at least a processor and parts of memory, and memory may contain instructions for executing software programs. Components may be associated with a device.

[0033] As used herein, “logic” is synonymous with “circuit” and includes, but is not limited to, hardware, firmware, software, and / or combinations thereof for performing a function or action. For example, depending on the desired application or need, logic may include discrete logic such as a software-controlled microprocessor, an application-specific integrated circuit (ASIC), or other programmed logic devices and / or controllers. Logic may also be fully embodied as software.

[0034] As used herein, “processor” includes, but is not limited to, one or more substantially any number of control systems or standalone processors, such as microprocessors, microcontrollers, central processing units (CPUs), suitable integrated circuits, one or more field-programmable gate arrays (FPGAs), one or more application-specific integrated circuits (ASICs), one or more digital signal processors (DSPs), and / or any combination thereof. A processor may be associated with a variety of other circuits that support the operation of the processor, such as random access memory (RAM), read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), clocks, decoders, memory controllers, or interrupt controllers. These support circuits may be inside or outside the processor or its associated electronic packaging. The support circuits may be able to communicate with the processor in an operable manner.

[0035] As used herein, “signal” includes, but is not limited to, one or more electrical signals, including analog or digital signals, one or more computer instructions, bits, or bitstreams. The term “signal” can also correspond to “data” and “information” communicated from one device or component to another.

[0036] As used herein, “software” includes, but is not limited to, one or more computer-readable and / or executable instructions (for example, stored on a local machine-readable medium or server) that cause a computer, processor, logic, and / or other electronic device to perform functions, actions, and / or operate in a desired manner. Instructions may be embodied in various forms, such as routines, algorithms, artificial intelligence models, modules, or programs, including separate applications or code from dynamically linked sources or libraries.

[0037] Referring here to the drawings, an exemplary control system 10 for an automated GMAW welding system is shown. The control system 10 is configured to dynamically adjust operating parameters to achieve desired performance targets, such as desired uptime %, cycle time, throughput (e.g., parts or assemblies per hour), and / or quality metrics (e.g., desired repeatability or dimensional tolerance). Generally, the control system 10 may include a controller 20, an input device 50, and a sensor device 70. The controller 20 may be communicatively coupled to the input device 50 and configured to receive planned operating parameters for welding operations performed by the equipment 5 in the work cell. The equipment may include, but is not limited to, a welding source 2, a wire feeder 3, a welding controller, a robot controller, a hot wire power supply, a workpiece positioner 7, a wire positioner, a welding gun or torch 8, a nozzle positioner / cleaner, etc. For the purposes of this disclosure, the term “welding source” is intended to refer to a laser source or power supply for a welding application.

[0038] Planned motion parameters may include planned motion parameters and planned process parameters. Planned motion parameters may include, but are not limited to, position, velocity, acceleration, deceleration, jerk, and the trajectory or path taken by the robot arm 6 (e.g., the arm operating the welding gun). In one embodiment, motion parameters may define a weave pattern, such as one or more welds. Planned motion parameters may also relate to the operation of other equipment used to perform the welding operation, such as the operation of a positioner 7 configured to operate one or more workpieces 12 attached thereto, or controlled vibration of a laser beam (e.g., laser "wobble"). In further embodiments, planned motion data may define motion parameters of other types of equipment (e.g., wire feeders, torch cleaners, or material handling robots).

[0039] Planned motion parameters can be defined according to a plan or a predetermined time sequence (e.g., a time series) that specifies the planned motion parameters at various stages of the welding operation. For example, a time series can define motion and process parameters for a first stage of positioning and fixing the workpiece on the positioner 7, a second stage of moving the positioner to the welding position, a third stage of moving the robot arm 6 to the welding position, a fourth stage of moving the robot arm 6 along a predetermined path to create one or more welds, a fifth stage of returning the robot arm 6 to its home position, and a sixth stage of removing the workpiece (e.g., via a material handling robot). In such embodiments, the time series can be measured in microseconds (μs), milliseconds (ms), seconds, hours, days, or months.

[0040] The planned process parameters may include, but are not limited to, parameters related to the welding process. Various non-limiting examples include cycle time, duty cycle, welding process type (e.g., gas metal arc welding - GMAW, gas tungsten arc welding - GTAW, flux-cored arc welding - FCAW, shielded metal arc welding - SMAW, etc.), welding wire type and size (e.g., diameter), wire feed rate, waveform, output current, output voltage, trim value, protrusion length, contact tip-to-workpiece distance, polarity (e.g., DC electrode positive - DCEP, DC electrode negative - DCEN, etc.), welding speed, transition mode (e.g., short circuit, globule transition, spray transition, pulsed spray transition, etc.), weld joint configuration (e.g., corner weld, butt weld, etc.), workpiece material type, laser source settings, wire feeder settings, welding gun settings, remote current / voltage settings, and shielding gas parameters (e.g., shielding gas flow rate, composition (e.g., 100% CO2, argon / CO2 blend, etc.)).

[0041] Planned process parameters can also be defined for the same time series to which motion parameters are assigned. For example, planned process parameters (e.g., welding current, wire feed rate) can be specified at a specific point in the time series to which planned motion parameters are also assigned. This allows for coordinated control of planned process parameters (e.g., current and wire feed rate) and planned motion parameters (e.g., position, velocity, acceleration, deceleration, trajectory, or path) for the same time series to perform the welding operation.

[0042] The above exemplary planned motion and process parameters relate to robotic GMAW welding operations, but it should be understood that such parameters may vary depending on the type of material handling operation being performed. For example, planned motion parameters for a thermal cutting process may include planned motion parameters or planned process parameters for robotic or mechanized cutting operations. For example, planned motion parameters may include, but are not limited to, the path or trajectory of the cutting torch, the cutting speed, and / or the position or orientation of the cutting torch relative to the workpiece or worktable. On the other hand, planned parameters may include, but are not limited to, drilling time, cutting current, feed rate, cutting depth, standoff distance, cutting or assist gas characteristics (e.g., composition, flow rate, pressure, etc.). As another example, planned motion parameters for an assembly operation may specify the path, position, speed, acceleration, deceleration, trajectory, or path taken by the robot arm, for example, to pick up a workpiece from one location and transport it to another location (e.g., a robotic welding fixture or position). Thus, it should be understood that the various inventions disclosed herein may be applicable to a wide variety of manufacturing and assembly processes.

[0043] In some embodiments, the controller 20 may be configured to dynamically determine some or all of the planned motion and / or planned process parameters based on acquired data, such as three-dimensional spatial data or depth-extended visual data. In certain embodiments, the controller 20 may acquire spatial data from a sensor, which in the illustrated embodiment represents an image acquisition device 72. For the purposes of this disclosure, “spatial data” may include, but is not limited to, point clouds, 2D or 3D images, videos, depth maps, two-dimensional or three-dimensional coordinate measurements, and any other data or dataset representing the spatial properties of an object. In such embodiments, the image acquisition device 72 may be located on a robot arm 6 (see, for example, Figure 1) or on a welding gun 8 mounted thereon (see, for example, Figure 2). The image acquisition device 72 may embody a two-dimensional or three-dimensional camera (e.g., a stereo camera), a laser tracking device, a laser scanning device (e.g., a high-fidelity laser scanner), a LiDaR sensor, or any other suitable example of the image acquisition device 72 disclosed herein.

[0044] As shown in Figure 2, the image capture device 72 can be configured to detect and scan the weld joint 14 to evaluate the tack weld between the two workpieces 12. This may include, for example, measuring the gap between the workpieces along the planned welding path P of the weld joint 14, or determining the distance along the vertical axis between the welding gun (e.g., its contact tip) and the workpiece. In such embodiments, the controller 20 may utilize spatial data to define motion parameters such as the trajectory of the robot arm 6 and / or the welding gun / torch attached thereto. For example, the trajectory may define a weave pattern for welding between workpieces via the welding gun, or the path the welding gun follows. In some embodiments, the controller 20 may utilize spatial data to determine planned process parameters. For example, the controller 20 may utilize spatial data to determine the electrode protrusion distance, the distance from the contact tip to the workpiece, or the weld joint configuration. In further embodiments, the controller 20 may utilize spatial data to determine information about the workpiece to be welded, such as the joint gap or the dimensions of the workpiece (e.g., thickness, length, width, etc.). In this way, spatial data can be used to derive workpiece-related information (for example, referred to herein as “workpiece data” or “workpiece information”). In some embodiments, the controller 20 can determine planned process parameters (e.g., current, voltage, etc.) based on the workpiece data. For this purpose, the controller 20 can refer to a lookup table (for example, stored in the storage device 24 or a remote server) that defines predetermined motion or process parameters corresponding to spatial data acquired by the image acquisition device 72.

[0045] The input device 50 may include a control panel or remote device (e.g., a smartphone, tablet, pendant, or laptop) having a user interface 52. The user interface 52 may include a graphical display that can be operated to input, upload, and / or modify planned motions and planned process parameters. In some embodiments, the user interface 52 may include, for example, one or more control knobs, buttons, sliders, touchpads (or screens), cameras for visual command recognition or image capture (e.g., tags or QR codes on scanned parts / workpieces to be welded), keyboards, pointing devices (e.g., mice, trackballs, touchpads), graphics tablets, scanners, audio input devices (e.g., voice recognition devices, microphones, etc.), and / or other types of known input devices that facilitate communication between the user and the control system 10.

[0046] In some embodiments, the input device 50 may be operable to receive software containing planned motion and process parameters via a communication component 54 that enables wireless communication between the input device 50 and one or more remote devices 60 (e.g., laptops, mobile phones, etc.) capable of inputting or uploading software. The communication component 54 may include a physical and / or logical interface for connecting the input device 50 to one or more remote devices 60 or another network. For example, the communication component 54 may enable Wi-Fi-based communication such as via frequencies defined by the IEEE 802.11 standard, short-range radio frequencies such as Bluetooth®, cellular communication (e.g., 2G, 3G, 6G, 6G LTE, 5G, etc.), or any suitable wired or wireless communication protocol that enables one or more remote devices to communicate with the control system 10. In some embodiments, the communication component 54 may include a cellular network card, Token Ring, FDDI, ArcNet, or other examples of dedicated network interface cards.

[0047] In some embodiments, the input device 50 may include a physical interface, such as a USB interface, serial interface, or wired interface (e.g., Ethernet, fiber optic interface), and equivalents, for uploading, inputting, or modifying planned motion and process parameters. In some embodiments, the input device 50 may include an encryption component 56 that can operate to encrypt the planned motion and process parameters for data security and / or privacy purposes (e.g., to protect the privacy of customer data). The input device 50 may also include a storage device 58 that can operate to store the planned motion and process parameters in memory. In various implementations, the memory may be implemented using any suitable memory technology, such as static random access memory (SRAM), synchronous dynamic RAM (SDRAM), non-volatile / flash type memory, or any other type of memory capable of storing information. In certain embodiments, the input device 50 may be communicably coupled to a remote cloud-based server for storing the planned motion and process parameters.

[0048] In the illustrated embodiment, the input device 50 is shown as a separate component configured to be communicatively coupled to the controller 20 and to supply therewith one or more data streams (e.g., via a system bus including a physical / wired connection). In some embodiments, the input device 50 may form part of the controller 20 or be communicatively coupled to the controller 20 via a communication component (e.g., 56) that enables wireless communication between them. In the illustrated embodiment, the input device 50 transmits a first data stream 59a containing planned motion parameters and a second data stream 59b containing planned process parameters. It should be understood that the first and second data streams 59a and 59b may be transmitted to the controller 20 via a single data stream or connection (instead of two). It is also conceivable that the input device 50 may transmit three or more data streams associated with different types of planned motion parameters.

[0049] Referring further to Figure 1, the sensor device 70 may include one or more sensors configured to measure and acquire sensor data from downstream material handling operations. The sensor device 70 may also include a data acquisition system (DAQ), a data logger, a programmable digital multimeter, or another suitable device configured to acquire, store, and timestamped sensor data in real time. In some embodiments, the sensor device 70 may include a system clock (e.g., a real-time clock) configured to automatically timestamp the acquired sensor data (e.g., UNIX time, datetime string, or relative time in milliseconds). The sensor device 70 may be operably connected to a work cell 5 (or equipment) to measure and retrieve sensor data acquired from one or more downstream processes (e.g., downstream robotic welding or cutting processes) in real time. For the purposes of this disclosure, “real time” may embody real time or near real time (e.g., considering network latency (e.g., milliseconds, seconds, etc.)). The sensor data may include deviations in the welding (or cutting) process, for example, deviations between one weld produced by work cell 5 and another weld produced by it, or actual motion data, actual process data, and workplace data associated with external factors that may cause variations in cutting quality.

[0050] Workplace data may include, but is not limited to, workpiece data, supply data, and environmental data. Supply data may relate to tools, supplies, or equipment. For example, supply data may relate to welding equipment including, but not limited to, a welding source (i.e., welding power supply or laser source), hot wire power supply, robot, torch, gun, wire, wire feeder, torch cleaner, shielding gas supply unit, or other components. In other embodiments, supply data may relate to cutting equipment including, but is not limited to, a laser source, plasma source, cutting table or gantry, robot, torch, or oxygen fuel supply source. Supply data may include, for example, equipment status information indicating whether a power supply, laser source, wire feeder, robot, positioner, or other peripheral device is fully operational or experiencing a failure due to, for example, a component failure. Supply data may also include technical specifications associated with each piece of equipment, tool, or supply. For example, supply data may define the update rate or communication frequency of one or more robots.

[0051] Environmental data may include data related to the surrounding environment (e.g., temperature, humidity, air quality, etc.).

[0052] The workpiece data may include information about the properties or physical characteristics (e.g., material dimensions, alloy composition, shape, etc.) of one or more workpieces to be welded. As described above, the workpiece data can be derived from spatial data acquired by a sensor in the form of an image capture device 72.

[0053] In some embodiments, the image acquisition device 72 can take various forms, including but not limited to LiDaR sensors, ultrasonic sensors, infrared (IR) distance sensors, cameras (2D or 3D), laser trackers, or laser scanning devices (e.g., high-fidelity laser scanners). In further embodiments, the image acquisition device 72 can embody a stereo 3D camera 74 (Figure 2) having two fields of view 74a and 74b, enabling the generation of three-dimensional images of one or more workpieces. In this way, the image acquisition device 72 can embody a suitable configuration for acquiring any suitable example of the spatial data disclosed herein. In some embodiments, the image acquisition device 72 may be mounted on a robotic arm or on a robotic welding gun (or cutting torch). In some embodiments, the image acquisition device 72 may be mounted in another location on the work cell 5 (e.g., adjacent to the work cell 5 or on a wall surrounding the work cell 5, as shown in Figure 2). For example, the image acquisition device 72 may be configured to acquire images (from two fields of view) in order to generate a 3D representation of a wider scene, including a positioner, a robotic arm, a welding gun (or cutting torch), and a workpiece.

[0054] As shown in Figure 2, the image acquisition device 72 can be configured to detect and scan the weld joint 14 to acquire information about one or more workpieces derived via spatial data. For example, the controller 20 can use the spatial data to determine workpiece data including, but not limited to, the two- or three-dimensional coordinates or dimensions of the workpiece or weld groove, joint dimensions and geometric shape (e.g., joint type), workpiece gap dimensions, texture or surface roughness characteristics, position and / or orientation of the welding gun / torch (e.g., its contact tip) or cutting torch relative to the workpiece (e.g., including the distance between them or its spatial coordinates), electrode protrusion distance, arc length, positioning of the workpiece relative to the worktable or positioner, and other physical characteristics. In some embodiments, the controller 20 may include a 3D dimensional measuring instrument configured to determine the dimensions of the workpiece based on the spatial data. Additionally or alternatively, the controller 20 can utilize a deep learning neural network (e.g., a convolutional neural network) to convert the spatial data (supplied as input to the network) into an output, for example, to define the pixel coordinates of the workpiece or the surrounding environment.

[0055] Returning to Figure 1, the sensors of the sensor device 70 may be configured to retrieve other examples of workpiece data. For example, the sensors may include an infrared digital thermometer or thermal imager (for determining the location of the heat-affected zone or the temperature of the workpiece). The sensors may also include proximity sensors (e.g., inductive, optical, capacitive, magnetic, or ultrasonic proximity sensors) configured to detect the presence of the workpiece relative to a welding gun / torch or cutting torch. In a further embodiment, the controller 20 may be communicably coupled to a storage device or cloud server that defines workpiece information, such as the workpiece material type, thickness, joint configuration (e.g., square weld, butt weld, etc.), material type (of the workpiece), etc., before starting a material processing application. In this embodiment, the controller 20 may adjust the workpiece information based on the workpiece data received (via an artificial intelligence model).

[0056] In some embodiments, the sensor may include environmental sensors for checking environmental data, such as ambient temperature or humidity in a workplace, or sensors for detecting the presence of specific gases (e.g., carbon monoxide, flammable gases, ozone, etc.) or contaminants in the air (e.g., electrochemical sensors, analyzers, catalytic diffusion sensors, or other known gas sensors).

[0057] The sensor device 70 may also include sensors configured to capture real-time motion data (for example, of a robotic arm, welding gun or cutting torch, positioner, or workpiece) via, for example, one or more rotary encoders or inertial measurement units (equipped with accelerometers, gyroscopes, and magnetometers). The inertial measurement unit may be configured to measure linear acceleration, rotational velocity, and orientation in 3D space. For example, the inertial measurement unit can capture and record data related to the motion of a robotic arm, a welding gun (for example, of its weave pattern), or another welding tool or accessory related to welding operations. In some embodiments, the sensors may include tilt sensors or yaw rate sensors for measuring pitch and roll.

[0058] In addition, the sensors may include sensors for retrieving actual process data in real time from equipment operably connected to the controller, including but not limited to welding process data transmitted from the welding equipment 5.

[0059] Referring further to Figure 1, the sensors may include, but are not limited to, voltmeters, shunt resistors (for measuring current), Hall effect or current transformers (e.g., clamp-on for measuring current), arc voltage probes, and oscilloscopes (for checking waveforms). For laser welding applications, the sensors may include, but are not limited to, laser power meters, beam profilers, beam alignment sensors, and photodiodes.

[0060] Actual process data may include, but is not limited to, cycle time, duty cycle, welding process type (e.g., gas metal arc welding - GMAW, gas tungsten arc welding - GTAW, flux-cored arc welding - FCAW, shielded metal arc welding - SMAW, etc.), welding wire type, wire size (e.g., diameter), wire feed rate, waveform, output amperage, output voltage, trim value, wire protrusion length, contact tip-to-workpiece distance, polarity (DC electrode positive - DCEP, DC electrode negative - DCEN), welding speed, transition mode (e.g., short circuit, globule transition, spray transition, pulsed spray transition, etc.), wire feeder settings, welding gun settings, remote amperage control settings, remote voltage control settings, shielding gas flow rate, shielding gas composition (e.g., 100% CO2, argon / CO2 blend, etc.), laser output, or any other examples of process data or process parameters disclosed herein. In some embodiments, the sensor device 70 may include sensors configured to measure the penetration of one or more welds (e.g., laser scanners, infrared cameras, acoustic emission sensors, etc.). In some embodiments, the sensor device 70 may include an optical scanner (or any other suitable example disclosed herein) for monitoring the appearance of the weld, e.g., bead shape / height. In some embodiments, the scanner may include an acoustic scanner configured to monitor the sound of the weld (e.g., sputter level deviation indicating a quality problem). It should be understood that the actual process data may vary depending on the type of material handling application being performed. For example, as described above, the actual process data for a thermal cutting application may include data relating to cutting amperage, assist gas flow rate / pressure, and / or other examples disclosed herein. In some embodiments, the actual process data and actual motion data may include actions and responses of welding or cutting equipment operably connected to the controller 20. In further embodiments, the process data may relate to laser welding parameters or hot wire power parameters.

[0061] The sensor device 70 can receive sensor data and communicate the sensor data (i.e., workplace data, actual motion data, and actual process data) to the controller 20 in real time.

[0062] In the illustrated embodiments, sensor data is shown as being transmitted in first, second, and third data streams 79a, 79b, and 79c, respectively, for transmission to the controller 20. However, the data streams may be combined into a single data stream (instead of three). The data streams may be transmitted to the controller 20 via a system bus (including, for example, a physical / wired connection). The sensor device 70 may also transmit sensor data wirelessly to the controller 20, for example, via a communication component (such as 56 described above). In some embodiments, the sensor device 70 may communicate sensor data to a remote storage device or a cloud-based server for subsequent retrieval by the controller 20. In the illustrated embodiments, the sensor device 70 is shown as a separate component from the controller 20. The sensor device 70 may be intended to form part of the controller 20. In some embodiments, the sensor data may be converted into a suitable format that the controller 20 can process, such as a numerical vector for a scalar sensor, a time-series tensor, a pixel array, etc.

[0063] Referring further to Figure 1, the controller 20 may include a processor 22 that includes logic for determining and executing a process and motion parameter adjusted based on sensor data in order to achieve performance objectives, such as a desired throughput (e.g., desired repeatability, cycle time of a welded assembly including one or more welds, uptime %, or target quantity of parts or assemblies produced per hour). In other embodiments, the performance objectives may relate to quality metrics (e.g., desired first-pass yield, reduction of defects or scrap rate, desired weld penetration, welding characteristics such as the absence of porosity or slag, acceptable weld profile, bead shape, or part tolerances). In particular, the processor 22 can predict the adjusted process and motion parameter by utilizing an artificial intelligence model that takes into account external factors, such as deviations in the workplace environment (e.g., workplace data acquired by the sensor device 70), or deviations in actual motion and process data from planned motion and planned process parameters. The artificial intelligence model may embody a supervised artificial neural network (e.g., a deep learning neural network including an input layer, one or more hidden layers, and an output layer). In further embodiments, the artificial neural network may embody a convolutional neural network (for image processing), an RNN / LSTM / transformer (for time-series patterns), or a reinforcement learning model (for control optimization). In some embodiments, the artificial intelligence model may embody an unsupervised model (e.g., the input to the model is not classified or defined). The input layer may be configured to receive some or all of the planned motion data (i.e., planned motion and process data) and sensor data (workplace data, actual process data, and actual motion data) as input and to detect patterns in the received sensor data based on external factors (the aforementioned deviations revealed by the sensor data).For example, external factors may result from deviations in workplace data, such as deviations in the material being processed, joint configuration, material dimensions, material shape, gas supply, welding gun and gun components, ambient environment, wire supplied for the welding application, and / or any other deviations that can be revealed by sensor data. For example, an artificial neural network may use sensor data as input to predict adjustments to specific process parameters (e.g., current, voltage, wire feed rate, welding rate) that have a greater impact in achieving desired performance targets (e.g., required throughput or weld quality metrics). As a non-limiting example, an artificial neural network may predict higher voltage, reduced movement speed (e.g., by a robot), weaving technique, or increased wire feed rate to perform welding on joints that define a wide gap (to ensure the joint is properly filled and joined). In some embodiments, the controller may determine the dimensions and shape of the wide gap based on spatial data (e.g., captured by an image acquisition device) to derive workpiece data / information in real time. In this way, the controller 20 can adjust process parameters (e.g., voltage, travel speed, weaving technique, wire feed speed) in real time based on workpiece data by detecting wide gaps, sending workpiece data / information to the input layer of an artificial neural network, and predicting the adjusted process parameters via its output layer. While the above example relates to adjusting process and motion parameters for welding in wide gaps, it should be understood that the controller can also adjust such parameters when welding in narrow gaps (e.g., based on spatial data) by modifying the process parameters to define, for example, a straight bead, reduced wire feed speed, lower voltage, or increased travel speed (such modifications are predicted via the artificial neural network).

[0064] Artificial neural networks may also utilize sensor data for other purposes, for example, to solve optimization problems that reduce robot arm travel time, or to discover optimal robot arm paths that improve overall throughput as opposed to planned robot arm paths. In some examples, the artificial neural network may generate generated content, such as work instructions (for example, to a welder performing a semi-automatic welding application). The generated content is intended to be transmitted in the form of video or audio recordings. In such embodiments, the instructions may be transmitted in the native language of the individual performing the welding operation (for example, to display or play the instructions in the welder's native language). In some embodiments, the controller 20 may provide human-perceived feedback (for example, tactile feedback via the welding or cutting torch, or via the welding or cutting helmet) when certain conditions exist (for example, when sensor data reveals that the air quality is no longer suitable for welding or cutting). The instructions may also instruct the welding machine to adjust the torch angle or travel speed. In such embodiments, the instructions may implement the adjusted motion parameters derived from the predictions or recommendations of the artificial neural network to achieve desired / performance targets. In some embodiments, the indication may be provided via an indicator light (for example, to show when the speed or angle is appropriate, or when such parameter is outside the acceptable range).

[0065] In some embodiments, the artificial neural network may be complemented by other forms of machine learning (e.g., a deep Q network) to provide predictive insights into how input or decision variables (e.g., process data, motion data, workplace data) affect the achievement of desired performance targets (e.g., throughput or repeatability) based on the sensor data it receives. The artificial neural network may also be cloud-based so that it can wirelessly receive sensor data and transmit tuned process and motion parameters to the controller 20 (e.g., via a communication component) to control the welding operation (e.g., to achieve desired performance targets). In some embodiments, the weights of the artificial neural network may be optimized during a supervised learning process, and the optimal model weights (e.g., for a given welding or cutting application) may be stored in a memory device 24 or a remote cloud-based server for later retrieval. In some embodiments, the artificial neural network can utilize backpropagation to adjust the network weights to improve prediction accuracy (by reducing errors in network weights and biases) and / or determine how each weight contributed to the error (e.g., based on the deviation between predicted data values ​​and target data values). In this way, the controller 20 can utilize the artificial intelligence model (e.g., a deep learning neural network) to identify the optimal set of weights to achieve a desired performance target via the model's decision variables (e.g., via tunable motion and / or process parameters).

[0066] Furthermore, each layer of the artificial neural network may have one or more nodes, each node connected to all nodes in the next layer and all nodes in the previous layer (if applicable). Nodes in the input layer may represent features corresponding to sensor data (e.g., actual process data, actual motion data, or workshop data). For example, one node may represent the characteristics of a welded joint, e.g., that the welded joint is a butt joint. In this example, assigning a value of 1 to that node may represent that the joint is a butt welded joint, and assigning a value of 0 may represent that the joint is not a butt welded joint. It is also intended that one or more input nodes may correspond to part dimensions, pixels (in an image), and / or other aspects of sensor data (i.e., workshop data, actual process data, or actual motion data). Each node in the hidden layer may receive output values ​​from nodes in the previous layer (e.g., the input layer) and associate each of the nodes in the previous layer with a weight. Next, each hidden node can multiply each value received from the node in the previous layer by the weight associated with that node, and output the sum of the products to each node in the next layer. The nodes in the output layer can then receive and process the input values ​​received from the nodes in the hidden layer in a similar manner. Exemplary outputs in the output layer may include predictions / recommendations that, when executed by the controller, will achieve performance targets, e.g., predictions / recommendations of tuned process parameters and / or tuned motion parameters. In another non-limiting example, the artificial neural network could receive sensor data paired with performance targets (e.g., desired productivity or quality metrics) as input and generate tuned process and / or tuned motion parameters that, when executed by the controller 20, will achieve performance targets.

[0067] In this way, the controller 20 can control the welding process (or cutting process) by applying or executing adjusted process and / or motion parameters (by sending motion or process commands to equipment operably connected to it) in order to achieve desired performance targets in real time. In some embodiments, an artificial neural network can be hosted on one or more servers and data stores and leverage a cloud-based deep learning infrastructure with artificial intelligence to analyze actual motion data, actual process data, and workplace data in real time. In some embodiments, the controller 20 can start a material handling application using nominal settings (e.g., planned motion parameters and / or planned process parameters), and the controller 20 can modify itself to generate adjusted motion or adjusted process parameters / data (e.g., on the fly) based on the sensor data it receives.

[0068] In some embodiments, the controller 20 can store the tuned process and / or motion parameters in a storage device 24 or a cloud-based server and retrieve these parameters according to a predetermined time sequence or time series. For example, the controller 20 may retrieve a specific subset of the tuned motion parameters and tuned process parameters at each stage of the welding operation. The controller 20 may also retrieve the tuned process parameters and tuned motion parameters during a later welding operation, for example, when it is desired to weld the same type of work assembly in the future. In other embodiments, the tuned process parameters and tuned motion parameters may be stored on a remote server or data store for retrieval by the controller via a wireless network, such as a WAN, LAN, LoRaWAN, NB-IoT, etc.

[0069] In some embodiments, the controller 20 can adjust process and / or motion parameters in response to detecting a failure of an element within the equipment via sensor data, such as a robot, positioner, wire feeder, gun / torch, laser source, power supply, or other peripheral devices associated therewith, to continue the material processing operation at a reduced speed (e.g., reduced throughput). In these embodiments, the controller 20 can adjust process and / or motion parameters (via an artificial neural network) to maintain the material processing operation at reduced throughput in consideration of the element failure. For example, the controller 20 can adjust motion parameters to the extent that the robot becomes immobile on one of its axes (e.g., one of its six axes) to complete the material processing operation using the remaining axes, but the cycle or takt time increases due to the limited range of motion. As another example, if the welding system is unable to maintain the target welding speed (for example, due to a wire feeder malfunction), the controller 20 may detect this condition via sensor data (e.g., supply data) and determine process parameters (e.g., adjusted wire feeding speed) adjusted to continue the welding operation at a reduced welding speed. As yet another example, sensor data may receive an audio signal (e.g., captured via an acoustic sensor) that identifies a failure in a particular component (e.g., a faulty gear on a positioner). In such an example, the controller 20 can maintain throughput by reducing the speed of the welding robot (at a reduced speed to compensate for the component failure) instead of completely stopping production. It is also intended that various inventions disclosed herein may be applied to other forms of manufacturing, such as an assembly line where a conveyor is not operating at full capacity due to a failure of a component somewhere along the line. In such an example, various inventions disclosed herein may enable the assembly line to adjust process parameters (e.g., robot material handling speed) to compensate for the component failure and maintain production at a lower throughput than expected (instead of stopping production).Advantageously, this aspect of the disclosure enables continued operation despite component failure, thereby avoiding production downtime. In certain implementations, manufacturers may choose to maintain output at reduced throughput rather than completely halting production.

[0070] In some embodiments, the artificial neural network may receive inputs specifying performance targets, including required levels of productivity or quality instructed by the customer. For example, the neural network may receive inputs indicating that an intermediate level of quality is acceptable to a particular customer (e.g., allowing for increased deviations such as larger part tolerances, weld consistency, bead profile, and weld penetration) in order to achieve higher throughput requirements (e.g., to meet required shipping dates). In this way, the artificial neural network can adjust its weights to target the specified level of quality while achieving increased throughput, for example, by minimizing the error between its predictions and the actual target values ​​for achieving the required levels of quality and throughput. In this way, the artificial neural network can map its recommended actions to system control variables, e.g., adjusted motion and / or process parameters, in order to achieve the required levels of quality and throughput. In some embodiments, training data for the artificial neural network may be derived from actual process and / or motion data (e.g., acquired during previous material handling operations). In other embodiments, training data may instead be generated within a simulated or virtual material handling environment, as further described below.

[0071] As shown in Figure 3, another exemplary control system 100 is shown, in which the input device 150 is operable to define a desired balance between quality and throughput. In the illustrated example, productivity is shown as taking precedence over quality. It is assumed that the input device 150 may be operable to receive other user inputs (e.g., increased quality level, required throughput level, required dimensional quality (e.g., repeatability, tolerance, through-hole, etc.)) via any example of a user interface or control panel disclosed herein. In some embodiments, the controller 120 may run a specific artificial intelligence model (e.g., comprising a mathematical model, depicted in example as a behavior modifier) ​​corresponding to a desired level of quality or productivity. For example, the controller 120 may invoke a specific mathematical model (e.g., a specific artificial neural network) previously determined to provide optimal part quality. It should be understood that the controller 120 may also invoke a specific model tailored to the type of manufacturing being performed, e.g., a specific mathematical model for a fully automated welding application 105a, or a mathematical model for a semi-automated welding application 105b. In such embodiments, the inputs to each model may differ. For example, a model for a semi-automatic GMAW welding application does not receive actual motion data for the robot arm as input because the welding gun is manually controlled by the operator.

[0072] In some embodiments, desired performance targets may be defined in other ways, for example, to reduce operating costs (e.g., reducing welding supply costs, gas costs, spatter reduction, etc.) or to achieve some other initiative (e.g., to improve safety or reduce emissions for applications that emit pollutants or contaminants into the air). In some examples, various insights and patterns learned by artificial intelligence models can reveal errors or problems that engineers or production managers had not previously considered, which can be insightful for the future when designing or developing new welding or cutting systems.

[0073] As described above, the controller 120 can implement or execute adjusted motion and / or process parameters to achieve desired or target performance attributes via commands sent to the welding system (e.g., shown as an action affecting Figure 2). Commands corresponding to adjusted motion parameters may include signals sent to the robot (e.g., to change the robot's path). Commands corresponding to adjusted process parameters may include signals sent to the welding power supply (e.g., to adjust current or voltage settings), the laser source (e.g., to adjust power), or the wire feeder (e.g., to adjust wire feeding speed). In some embodiments, adjusted motion parameters may include adjusted velocity, acceleration, jerk, and / or torque. Commands may also be transmitted to various other devices, such as a positioner (e.g., to change its tilt or motion), a cutting table (e.g., to change the cutting speed, cutting depth (or kerf / width)), a gas supply system for adjusting assist gas for laser cutting applications, a spray gun / atomizer for painting operations (to adjust the paint deposition rate based on the type and shape of the material being painted), or an electric grinder or sander for sanding or grinding operations (e.g., to adjust the grinding speed when sensor data reveals the presence of contaminants on the surface of the workpiece). Thus, it should be understood that the various inventions disclosed herein may be applicable to a wide variety of manufacturing operations. In some examples, commands may relate to predicting and recommending a fluid flow rate (e.g., a coolant supplied to a welding torch to dissipate heat from the welding torch). In such examples, the controller 20 can use computational fluid dynamics to determine the optimal flow rate based on sensor data (e.g., based on temperature data, flow rate data, or the type (rating) of the welding torch used in the application).

[0074] In some embodiments, the artificial intelligence model may include constraints, such as limitations or restrictions on decision variables used in the manufacturing application (e.g., process or motion parameters). For example, the movement of a robotic arm may be limited to a specific speed, reach, or free axis (as indicated by the robotic model). Other non-limiting examples of constraints may include a maximum wire feeding speed (determined by the type of wire feeder used) or a maximum current level (e.g., determined by the welding power source used).

[0075] Next, with reference to Figure 4, another exemplary control system 200 will be described. For brevity, descriptions of similar features have been omitted, and the following description will focus on the differences in this embodiment.

[0076] In the illustrated example, the controller 220 is configured to receive input data from one or more remote devices 260 (including software applications) configured to supply planned operating parameters (planned process and motion parameters) to the controller 220. Each of the remote devices 260 may include a specific software / interface language. For example, an automated welding application (e.g., for additive 3D metal printing) may utilize a different software or interface language than that used for a mechanized thermal cutting application. Furthermore, the controller 220 may include a decryption component 261 configured to decrypt software languages ​​(e.g., encrypted commands it receives) to facilitate secure communication between the controller 220 and the remote devices 260. In some embodiments, the decryption component 261 may also be configured to unify software and / or interface languages ​​(e.g., those associated with a welding system, a cutting system, a robot, etc.) into a single or common software and / or interface language for use by the controller 220. For example, the decoding component 261 can decode the software language and then commonize the planned motion and planned process parameters (received from various remote devices) into a single software and / or interface language. However, it should be understood that the order can be reversed, with the language being commonized first and then decoded. In this way, the controller 220 can be configured to interface and communicate with various different remote devices and software applications for various applications (e.g., thermal cutting, material handling, automated welding, etc.) regardless of the native software and / or interface language associated with it. For example, the decoding component can be coupled to communicate with at least a first input device and a second input device.A first input device may be operable via a first programming language or interface language including a first motion or process command, and a second input device may be operable via a second programming language or interface language including a second motion or process command. A decoding component may be configured to decode the first and second programming languages ​​and to unify them into a common programming language. The common programming language may include planned motions and planned process parameters for a material handling application, associated with a first motion or process command (derived from the first programming language) and associated with a second motion or process command (derived from the second programming language).

[0077] The controller 220 may also include a motion planner component 262a and a process planner component 262b. The motion planner component 262a can identify and extract all expected motion commands, including but not limited to robots, positioners, welding sources, or welding guns or cutting torches, to be taken by equipment 205 operably connected to the controller 220, based on planned motion parameters (e.g., arcs, trajectories).

[0078] In some embodiments, the motion planner component 262a may be configured to communicate with a remote device 260 and to identify where a particular weld path is located within a welded assembly (using a 3D model) and to define motion commands for that location. The motion planner component 262a may also include logic for converting a particular planned trajectory or arc into a vector.

[0079] The process planner component 262b can identify and extract all expected process commands based on the planned process parameters, such as welding source, welding gun or torch, cutting table, laser cutter, and other process commands.

[0080] The controller 220 may also include a profile planner component 264a configured to define and extract other motion commands, such as complex motion commands that a welding torch must perform on a weave pattern during an additive manufacturing process. In some examples, the profile planner component 264a may transmit a command indicating that a weld joint containing a particular gap should require different weaving techniques and resulting motions to properly fill the joint. In the shown embodiment, motion commands (for planned motion parameters) and process commands (for planned process parameters) may be transmitted to buffer components 266a and 266b, respectively, configured to temporarily store each command for later retrieval, for example, to first-in, first-out buffers where data / commands are processed in the same order in which they were received.

[0081] The controller 220 may also include a motion trajectory component 268a and a process trajectory component 268b. The motion trajectory component 268a is configured to extract motion commands (e.g., one at a time) from the buffer component 266a and translate the planned motion commands (e.g., the geometric path taken, acceleration, deceleration, etc.) from points in space to a time series. In embodiments, the motion trajectory component 268a may define motion commands (derived from planned motion parameters) over a time series. In some embodiments, this may include defining specific motion parameters, including but not limited to jerk (e.g., the rate of change of acceleration or deceleration of an object over time), acceleration, velocity, and position. The process trajectory component 268b may be configured to define welding or cutting commands (e.g., derived from planned process parameters) over a time series, i.e., the same time series to which the motion commands are associated. In certain non-limiting examples, this may include defining when a power source directs current to the torch, or when a laser source generates and directs a laser beam to the torch, or when shielding gas is supplied through the welding gun to shield the welding arc. For example, a welding source may direct current (or light) to the torch when the torch is moved to a starting position to begin welding. In this way, it should be understood that process parameters (and corresponding commands) can be synchronized with motion parameters (and corresponding commands) at each stage of the welding operation. Furthermore, in other embodiments, processing commands may be defined separately from motion commands (as two separate data streams executed in parallel at each stage). Motion trajectory components and process trajectory components may also receive information about equipment or components operably connected to the controller 220, e.g., motion characteristics of a robot or other equipment, or e.g., information about the process (e.g., current, voltage) capabilities of a welding source. For example, in certain non-limiting examples, motion trajectory component 268a may be configured to retrieve the update rate or communication frequency of one or more robots.

[0082] The controller 220 may also include first and second mathematical models 280a and 280b. The first mathematical model 280a may include a first artificial intelligence model (e.g., a first deep learning neural network) related to optimizing motion parameters to achieve performance targets, and the second mathematical model 280b may include a second artificial intelligence model (e.g., a second deep learning neural network) related to optimizing process parameters to achieve performance targets. The first and second mathematical models 280a and 280b can receive sensor data from a memory device 297 or a remote or cloud-based server (communically coupled with the first and second mathematical models 280a and 280b) to determine motion parameters and adjusted process parameters to achieve desired performance targets (e.g., repeatability, uptime, or cycle time target requirements). The sensor data may embody data acquired in real time by the sensors of the sensor device 270, or may embody any example of the sensor data disclosed herein.

[0083] The controller 220 may also include a motion command composer 282a and a process command composer 282b. The motion and process command composers 282a and 282b may be configured to generate motion and process commands (based on planned or adjusted motion and process parameters) taking into account the signal interfaces of equipment (e.g., robots, welding sources, torches, positioners) operably connected to the controller 220. For example, the motion command composer 282a may construct commands in the form of a programming language used by the robot. In some examples, this may require translating commands from one language to another. The process command composer 282b may similarly be configured to construct commands in the form of a communication protocol or interface language associated with the equipment, such as the ArcLink® communication protocol for Lincoln Electric welding systems.

[0084] The controller 220 may also include a synchronization component 284 configured to synchronize motion commands and process commands over a time series (controlled by the system clock). For example, the synchronization component 284 can determine that a particular command cannot be executed simultaneously and ensure that such command is staged (e.g., delayed) until the result from a previous command is available. The synchronization component 284 may also be configured to take into account the command or update frequency receiving capability of the equipment, e.g., the time frequency (milliseconds / hour) at which a robot can take commands. The controller 220 may also include a motion command component 286a and a process command component 286b configured to transmit motion and process commands to the equipment 205 (e.g., a positioner, robot, nozzle and wire positioner, manual or robotic torch, welding source (e.g., laser source or power supply), wire feeder, etc.) according to a time series defined for the welding application. In some embodiments, the motion and / or process commands may embody Ethernet commands (e.g., serial commands) transmitted to the equipment via control signals. In the illustrated embodiment, the equipment 205 includes welding equipment. It should be understood that the equipment may include cutting equipment, for example, any example of cutting equipment disclosed herein.

[0085] Controller 220 may also include other features, such as a synchronizer component 290 configured to interface with multiple different manufacturing operations. Each of the different manufacturing operations may be operablely connected to its respective controller (e.g., 292, 294, 296). Controllers 292, 294, and 296 may have the same structure and functionality as controller 220, so that all controllers 220, 292, 294, and 296 can communicate according to the same communication protocol.

[0086] The control system 200 may include a sensor device 270 comprising one or more sensors configured to receive sensor data (e.g., workplace data, actual motion data, actual process data) and transmit the sensor data to a controller 220 in real time. The sensors may take any suitable form, e.g., any example of the sensors disclosed herein. In this way, the controller 220 may include logic for receiving and processing the sensor data and, accordingly, determining the adjusted motion and process parameters, for example, via the first and second mathematical models 280a and 280b. For example, the controller 220 may control motion parameters such as jerk (i.e., differentiated acceleration), acceleration (i.e., differentiated velocity), or velocity (i.e., differentiated position). Similarly, the controller may control process parameters, e.g., welding amperage, wire feed rate, voltage, or any other suitable example of process parameters disclosed herein. In certain embodiments, the control system 200 may be configured to refresh tuned process and motion parameters based on predetermined frequencies, including, but not limited to, every 10, 100, or 500 nanoseconds, every 10, 100, or 500 microseconds, every 10, 100, or 500 microseconds, every 10, 100, or 500 milliseconds, every 10, 100, or 500 centiseconds, or every 10, 30, or 60 seconds. In some embodiments, the predetermined frequency may be determined or constrained by the frequency of sensor readings.In some embodiments, sensor data, actual process data, and actual motion data may be interpolated to correspond to the same time series, for example, by the controller 220 adjusting process and / or motion parameters based on sensor data (i.e., actual process data, actual motion data, and / or workpiece data) that has been measured or interpolated to correspond at the same command transmission point, including, but not limited to, every 100 microseconds, every 500 microseconds, and / or any other suitable example of a predetermined frequency.

[0087] The controller 220 may also include a timestamp component 285 that is communicatively coupled to the sensor device 270. The timestamp component 285 may be configured to receive and timestamp sensor data (e.g., actual process data, actual motion data, and workplace data) for later retrieval. In some embodiments, the timestamp component 285 may be configured to associate a high-resolution time reference marker (e.g., accurate to microseconds or milliseconds, e.g., an exact time or measurement) with actual process data, actual motion data, workplace data, and / or other events (e.g., actions and responses taken by the equipment 205) detected via the sensor of the sensor device 270 or otherwise retrieved. In certain examples, the timestamp component 285 may be configured to assign timestamps with a time precision of about 2 microseconds (2 μs), thereby enabling an accurate chronological order of events occurring within the control system 200 and the equipment 205 operably connected thereto.

[0088] The timestamp component 285 may be communicatively coupled to one or more data storage systems. In certain embodiments, the data storage system may embody local storage or a remote data center. Timestamped sensor data may be stored as a continuous time-series dataset that preserves the order, frequency, and fidelity of the input sensor data. In some embodiments, the sensor data may be recorded at a rate that matches the operational sampling capability of the device 205, for example, the update frequency of one or more robots operably connected to it.

[0089] The stored, timestamped sensor dataset may be retrieved by a replay component 291 configured to reproduce the actions and responses of the instrument 205 and peripheral devices (e.g., captured via telemetry / sensor data). In this way, the replay component 291 can replay actions and responses including, but not limited to, actual process data, motion data, or workplace data (e.g., workpiece data, supply data, environmental data) captured during the material handling application. For example, the control system 220 may, in time synchronization, reproduce motion stages, process parameters, and peripheral device signals corresponding to a selected historical interval (e.g., in parallel with the time series). The replay function of the replay component 291 can support time frames ranging from intervals of less than one second (e.g., microseconds, milliseconds, etc.) to multi-year archiving periods, enabling post-event analysis without requiring the re-execution of physical material handling operations.

[0090] In various embodiments, the replay operation may be performed in conjunction with a graphical visualization interface (e.g., facilitating dynamic spatial visualization, statistical visualization, or process parameter / data analysis visualization). In some embodiments, the graphical visualization interface may display synchronized motion phases and corresponding process signals, enabling accurate reproduction of system behavior for diagnostic, quality control, or optimization purposes. The replay component 291 enables the identification of performance degradation, cycle time deviations, and process anomalies with temporal accuracy, thereby facilitating root cause analysis and process improvement.

[0091] In addition to replaying historical data, the timestamp component 285 may be configured to generate and store a composite dataset representing predicted process and motion profiles derived from historical sensor data acquired by the control system 200. The composite dataset can be supplied to the control system emulator component 293 to generate a high-fidelity digital twin of a physical or actual material handling application, such as a physical semi-automatic or fully automatic welding or cutting system. In some embodiments, the emulator component 293 may be configured to reproduce the temporal and behavioral characteristics of an actual manufacturing system, enabling forward emulation (referred to herein as the “replay” function) to predict performance results before designing, producing, or integrating any equipment 205. The emulator component 293 can favorably enable the user to create emulated solutions, the emulation being tightly coupled to real-world time constraints.

[0092] Synthetic and historical datasets may also be used to train predictive maintenance algorithms, self-healing control models, and behavioral models, enabling the control system 200 to predict equipment performance without performing physical process operations.

[0093] Referring here to Figure 5, an exemplary six-axis articulated robot 15 is shown mounted on a hexapod actuator 9. The hexapod actuator 9 (also known as a Stewart platform) may comprise a mobile platform 9b that is movable with six independent degrees of freedom. The hexapod actuator 9 has six linear actuators 14, which may be formed by small electric linear motors. The linear actuators 14 may extend between the proximal base portion 9a of the hexapod actuator and the mobile platform 9b supporting the robot 6. Each linear actuator 14 has a leg that extends and retracts to control the spatial orientation of the mobile platform (and therefore the robot). The mobile platform 9b is movable with six degrees of freedom by the linear actuators 14. In this way, the hexapod actuator 9 may be configured to increase the range of motion of the robot 15 (up to 12 degrees of freedom) so that the robot 15 can weld in places that would otherwise not be implementable (e.g., based on six degrees of freedom). This aspect of the disclosure may be useful for increasing the throughput and productivity of material processing operations, for example, by assuming a motion path that is shorter (and therefore faster) than the motion path that the robot itself can take. The movement of the hexapod actuator 9 can be controlled by various examples of controllers disclosed herein based on tuned motion parameters to optimize productivity or quality in material processing applications.

[0094] Next, referring to Figure 6, an exemplary method for controlling a material processing application is shown. The method can be carried out using any suitable combination of the apparatus, systems, devices, components, and / or configurations disclosed above. The following example is intended for optimizing a welding application, but it should be understood that the following may be applicable to other forms of manufacturing applications (e.g., thermal cutting, laser cutting, etc.).

[0095] In step 302, the method includes receiving a desired performance objective (indicated as a performance attribute). This may include receiving information (e.g., via a controller) corresponding to a desired quality or productivity metric for a material handling application. In step 310, the method includes receiving sensor data related to the welding operation (e.g., work area data, actual process data, and actual motion data). The retrieval of sensor data may be performed in real time. Step 310 may also include receiving planned process data (i.e., parameters) and planned motion data (i.e., parameters) for the material handling application.

[0096] In step 320, the method includes determining the adjusted process parameters and the adjusted motion parameters. Step 320 may be performed by a controller utilizing an artificial intelligence model (e.g., a deep learning neural network), using at least one of the planned motion data, planned process data, actual motion data, actual process data, and workplace data as input to the model.

[0097] In step 340, the method includes implementing or performing adjusted process parameters and adjusted motion parameters (e.g., via equipment operably connected to a controller (e.g., 20, 120, 220, etc.)) in order to achieve a target quality (e.g., repeatability) metric and / or a target productivity metric.

[0098] Referring to Figure 7, another exemplary method for controlling a material handling application (e.g., a welding or cutting application) is shown. In one embodiment, step 402 may include receiving planned motion and planned process parameters (e.g., from one or more input devices) including, but not limited to, position, velocity, acceleration, deceleration, trajectory, or path taken by a robot arm 6 (e.g., an arm operating a welding gun). For example, motion parameters may define a weave pattern for one or more welds. Planned motion parameters may also relate to other aspects of the welding application, e.g., motion of a workpiece positioner (e.g., axis of rotation, tilt), or any other suitable example of motion parameters disclosed herein. In further embodiments, planned motion parameters may define the motion of a laser welding torch, e.g., vibration motion of a laser beam (e.g., sway).

[0099] The planned process parameters may include, but are not limited to, parameters related to the welding application, such as cycle time, duty cycle, welding process type (e.g., gas metal arc welding - GMAW, gas tungsten arc welding - GTAW, flux-cored arc welding - FCAW, shielded metal arc welding - SMAW, etc.), welding wire type, wire size (e.g., diameter), wire feed rate, waveform, output amperage, output voltage, trim value, wire protrusion length, contact tip-base metal distance, polarity (DCEP, DCEN), welding speed, transition mode (e.g., short circuit, globule transition, spray transition, pulsed spray transition, etc.), weld joint configuration (e.g., corner weld, butt weld, etc.), material type (of the workpiece), wire feeder settings, welding gun settings, remote amperage control settings, remote voltage control settings, shielding gas flow rate, and shielding gas composition (e.g., 100% CO2, argon / CO2 blend, etc.). In other embodiments, process parameters may relate to laser parameters of a laser source to facilitate the laser welding application, such as power, wavelength, beam quality, spot, welding speed, or pulse parameters. In further embodiments, process parameters may relate to hot wire power supply parameters (for example, for resistance heating of the welding wire).

[0100] In step 404, the method may include, for example, receiving a performance objective via an input device. The performance objective may also relate to product throughput, quality metrics (e.g., reproducibility), or any other suitable example of a performance objective or attribute disclosed herein. In certain embodiments, the performance objective may embody a target wire feed rate, a target takt or cycle time (for welding an assembly or one or more workpieces), or a target welding rate. On the other hand, in further embodiments, the performance objective may relate to yield, product / part conformance %, uptime %, cycle time, or defect rate (e.g., number of parts or assemblies per 1,000, 100,000, or 1,000,000).

[0101] In the embodiments shown, the planned motion parameters and planned process parameters are received before the performance targets are received. In certain embodiments, this may be reversed.

[0102] In step 406, the method may include, for example, receiving sensor data via a sensor device after initiating a welding application, such as actual motion data, actual process data, and optionally actual workplace data. Actual motion data, actual process data, and actual workplace data may represent any suitable examples of motion parameter / data, process parameter / data, or workplace data disclosed herein. In certain embodiments, the actual motion data, actual process data, and actual workplace data may be time-stamped and recorded in a central server 412 or a remote storage device 414 for later retrieval when it is desired to evaluate the actions and responses taken by the welding system (for example, to facilitate a replay or playback function as disclosed herein).

[0103] Step 406 may optionally include sending the actual process data, actual motion data, and actual workspace data to a remote server 412 or data storage device 414 for later retrieval. This may be done via any suitable example of a telemetry device. In embodiments, the actual process data, actual motion data, and actual workspace data can be used as input for exemplary playback or replay functions, as described above.

[0104] In step 408, the method includes determining tuned process parameters and tuned motion parameters based on actual process data, actual motion data, and workspace data. This can be done, for example, via a controller (e.g., 20, 120, 220) that receives actual process data, actual motion data, and actual workspace data as input to an artificial intelligence model (e.g., a deep learning neural network). The artificial intelligence model may include an input layer, one or more hidden layers, and an output layer. In such embodiments, the controller (via the output layer) can predict recommended process and motion parameters based on the actual process data, actual motion data, and optionally actual workspace data received (e.g., in real time) to achieve performance targets. In some embodiments, this can be done, for example, by adding biases and applying nonlinear activation functions via neurons in the hidden layer (by applying a weighted sum of their inputs). In further embodiments, this can be done by minimizing errors (e.g., mean squared error, mean absolute error, etc.).

[0105] Step 408 may optionally include transmitting the adjusted process parameters and adjusted motion parameters to a remote server 412 or data storage device 414 for later retrieval. This may be done via any suitable example of a telemetry device. In embodiments, the adjusted process parameters and adjusted motion parameters can then be used for playback or replay functions disclosed herein.

[0106] In step 410, the method includes transmitting motion and process commands to equipment (e.g., a welding power source, laser source, hot wire power source, robot, positioner, welding gun, etc.) to implement the adjusted motion and adjusted process parameters in real time, for example, to achieve performance targets. The motion and process commands may be derived from or embody the adjusted motion parameters and adjusted process parameters. In embodiments, equipment (e.g., 5, 105a, 105b, or 205) may embody cutting equipment, including at least some or all of equipment for other exemplary applications, such as a cutting torch, laser source, cutting table, plasma source, or oxygen / fuel gas supply controller. This disclosure is an example, and it will be apparent that various modifications can be made by adding, modifying, or deleting details without departing from the fair scope of the teachings contained herein. Accordingly, the present invention is not limited to the specific details of this disclosure, except to the extent that the following claims are necessarily limited. [Explanation of Symbols]

[0107] 2 Welding source 3 Wire feeder 5 Equipment 6 Arms 7 Workpiece Positioner 9. Hexapod Actuator 9a Proximal base portion 9b Mobile Platform 10 Control Systems 12 workpieces 14. Welded joints 15. Six-axis articulated robot 20 controllers 24 Storage Devices 50 Input Devices 52 User Interface 54 Communication Components 58. Storage Devices 59a First data stream 59b Second data stream 60 Remote Devices 70 Sensor Devices 72 Image acquisition devices 74 Stereo 3D Cameras 74a field of view 74b Field of view 79a First data stream 79b Second data stream 79c Third data stream 100 control systems 105a Fully Automated Welding Application 120 controllers 150 Input Devices 200 Control Systems 205 Equipment 220 Controllers 260 remote devices 261 Decryption Components 262a Motion Planner Components 262b Process Planner Components 264a Profile Planner Components 266a Buffer Components 266b Buffer Components 268a Motion trajectory components 268b Process orbital components 270 Sensor Devices 280a The first mathematical model 280b The second mathematical model 282a Motion Command Composer 282b Process Command Composer 284 Synchronization Components 286a Motion command components 286b Process command components 291 Replay Components 292 Controllers 294 Controllers 296 Controllers 412 Remote Server 414 Data Storage Devices

Claims

1. A control system for material processing applications, A sensor device configured to receive actual motion data and actual process data from equipment performing the aforementioned material processing application, Controller and The controller is equipped with, Based on the actual motion data and the actual process data, the adjusted motion parameters and adjusted process parameters are determined via an artificial intelligence model. To achieve performance targets, the adjusted motion parameters and adjusted process parameters are transmitted to the equipment. Including logic, Control system.

2. The control system according to claim 1, wherein the artificial intelligence model is a neural network including an input layer, at least one hidden layer, and an output layer.

3. The control system according to claim 1, wherein the sensor device is further configured to receive workplace data during the material processing application, and the controller further includes logic for determining the adjusted motion parameters and the adjusted process parameters based on the actual motion data, the actual process data, and the actual workplace data.

4. The control system according to claim 1, wherein the workplace data includes at least one of workpiece data, supply data, or environmental data.

5. The control system according to claim 1, wherein the equipment comprises at least one of a robot, a hexapod actuator, a torch, a positioner, a power supply, a laser source, a cutting table, a cutting torch, a plasma source, or an oxygen / fuel source.

6. The control system according to claim 1, wherein the sensor device comprises an image capture device configured to extract spatial data, and the controller includes logic for determining workpiece information based on the spatial data.

7. The control system according to claim 1, wherein the sensor device includes an image capture device configured to extract spatial data, and the controller determines planned motion parameters and planned process parameters via the spatial data.

8. The controller generates generated content that embodies the work command based on the adjusted motion parameters or the adjusted process parameters. A control system according to claim 1, including logic.

9. The control system according to claim 1, further comprising a timestamp component configured to receive and timestamp the actual process data and actual motion data for later retrieval.

10. The sensor device is further configured to receive workplace data during the material processing application, and the controller is configured A replay component configured to reproduce the actual process data, the actual motion data, and the workplace data in a time-series and temporally synchronized manner. The control system according to claim 1, further comprising the following:

11. The aforementioned controller, A composite dataset is generated from the actual process data, the actual motion data, and the workshop data. The control system according to claim 10, further comprising logic.

12. The control system according to claim 11, further comprising an emulator component configured to generate a digital twin of the material processing application based on the synthesized dataset.

13. The controller is communicatively coupled to at least a first input device and a second input device, the first input device being operable via a first programming language including first motion or process commands, the second input device being operable via a second programming language including second motion or process commands, and the controller is A decoding component configured to decode the first programming language and the second programming language, and to unify the first programming language and the second programming language into a common programming language used to execute the material processing application. The control system according to claim 1, further comprising the following:

14. The aforementioned controller, Motion trajectory component and process trajectory component, wherein the motion trajectory component is configured to receive motion commands related to the planned motion parameters, and the process trajectory component is configured to receive process commands related to the planned process parameters. The motion trajectory component further comprises, The process trajectory components are configured to define the motion commands for a time series, and the process trajectory components define the process commands for the time series. The control system according to claim 13, configured as described above.

15. The artificial intelligence model, A first mathematical model and a second mathematical model, wherein the controller uses the first mathematical model to determine the adjusted motion parameters based on at least the actual motion data, and uses the second mathematical model to determine the adjusted process parameters based on at least the actual process data. The control system according to claim 1, including the following:

16. A material processing system, Robot and welding source, A controller operably connected to the robot and the welding source. The controller is equipped with, Sensor data is received in real time from the robot and the welding source. Process parameters and motion parameters are generated based on the sensor data via an artificial intelligence model. To achieve performance targets, the process parameters and motion parameters are transmitted to the welding source and the robot. Including logic, Material processing system.

17. The material processing system according to claim 16, wherein the sensor data includes actual process data and actual motion data.

18. The material processing system according to claim 16, wherein the artificial intelligence model includes a deep learning neural network comprising an input layer, at least one hidden layer, and an output layer, the input layer receiving the sensor data, and the output layer generating tuned process parameters and tuned motion parameters to achieve the performance targets.

19. A material processing method, Receiving actual process data and actual motion data from welding or cutting equipment, The controller generates adjusted process parameters and adjusted motion parameters based on the actual process data and the actual motion data. The adjusted process parameters and the adjusted motion parameters are transmitted in real time to the welding equipment or the cutting equipment. A material processing method, including the following.

20. The material processing method according to claim 19, wherein the welding equipment comprises at least one of a robot, a hexapod actuator, a welding source, a workpiece positioner, a nozzle positioner, a wire positioner, and a welding torch, and the cutting equipment comprises at least one of a cutting torch, a laser source, a cutting table, a cutting torch, a plasma source, or an oxygen / fuel source.

Images