Systems and methods for improving computer-aided manufacturing process

Abstract

A method for computer aided manufacturing of an item, comprising: receiving a project setup and a current state of a manufacturing process for the item, the current state comprising a representation of a workpiece and an operation tree; selecting, for each candidate option, a working tool, additional parameters and computing a toolpath force for a toolpath associated with the candidate option; computing, for each candidate option of a subsequent step of manufacturing the item, a score using a scoring function, wherein an input to the scoring function comprises a simulation of an operation corresponding to the candidate option; performing a simulation based on the selected working tool and the computed toolpath force; selecting a sequence of manufacturing operations and corresponding working tools to manufacture the item; generating machine-executable instructions configured to cause a manufacturing machine to perform the selected sequence of manufacturing operation.

Description

P-654716-PCSYSTEMS AND METHODS FOR IMPROVING COMPUTER-AIDED MANUFACTURING PROCESSFIELD AND BACKGROUND OF THE INVENTION

[0001] The invention, in some embodiments thereof, relates to improving computer-aided manufacturing process.

[0002] Computer-aided manufacturing (CAM), also known as computer-aided modeling or computer-aided machining is the use of software to control machine tools in the manufacturing of workpieces. CAM processes aim to create a faster and cost-effective production process and components and tooling with more precise dimensions and material consistency, while simultaneously reducing energy consumption. CAM is widely used in design studios and factories. Large factories employ hundreds of designers to use CAM and computer-aided design (CAD) systems, and inefficient design processes lead to waste of labor time and budget.SUMMARY OF THE INVENTION

[0003] Embodiments may include a computer-implemented method for computer aided manufacturing of an item, the method comprising: receiving, by one or more processors, a project setup and a current state of a manufacturing process for the item, the current state comprising at least a representation of a workpiece and an operation tree; selecting, for each candidate option, a working tool, additional parameters and computing a toolpath force for a toolpath associated with the candidate option; computing, for each candidate option of a subsequent step of manufacturing the item, a score using a scoring function, wherein an input to the scoring function comprises at least a simulation of an operation corresponding to the candidate option; performing a simulation based on the selected working tool and the computed toolpath force; selecting a sequence of manufacturing operations and corresponding working tools to manufacture the item; generating machine-executableP-654716-PC instructions configured to cause a manufacturing machine to perform the sequence of manufacturing operations according to the selected sequence of manufacturing operations.

[0004] In some cases, the method further comprising generating a graph data structure comprising nodes representing respective candidate options for the subsequent step of manufacturing the item and edges representing permissible transitions between the candidate options. In some cases, the method further comprising performing an initial filtering to retain, from among the candidate options, a set of candidates prior to computing the score. In some cases, the input to the scoring function further comprises an estimated next state of the workpiece that is based on at least one of a machine configuration, a tool configuration, or a dynamic material state in the machine.

[0005] In some cases, the method further comprising selecting a subset of the candidate options based on the computed score, and wherein selecting the working tool and computing the toolpath force are performed for each candidate option in the selected subset. In some cases, the method further comprising computing, for each candidate option of at least one subsequent step reachable from the candidate options, a respective score. In some cases, the method further comprising removing, from the graph, unfeasible candidate options based on at least one result of the simulation and the computed toolpath force, thereby pruning the graph, and wherein selecting a route of actions is performed along the pruned graph. In some cases, selecting the route of actions along the graph comprises selecting a route that reduces at least one of: predicted tool breakage, predicted fixture failure, chatter occurrence, or dimensional deviation. In some cases, removing the unfeasible candidate options comprises removing candidate options that violate one or more machine constraints comprising at least one of: axis travel limits, kinematic reachability limits, or collision avoidance constraints. In some cases, removing the unfeasible candidate options comprises removing candidate options that violate one or more force-related constraints comprising at least one of: a spindle torque capacity, a cutting force threshold, a tool breakage threshold, or a fixture load capacity. In some cases, removing the unfeasible candidate options comprises removing candidate options associated with a predicted chatter instability based on a modality analysis that identifies natural frequencies of a tool setup and a rest material geometry.P-654716-PC

[0006] In some cases, performing the simulation comprises performing the simulation based on the selected working tool and the computed toolpath force. In some cases, performing the simulation comprises performing physical simulation including at least one of: deformation analysis for the workpiece, deformation analysis for a tool assembly caused by a cutting force vector, vibration amplitude estimation, fixture load calculation, or surface finish geometric calculation. In some cases, the simulation comprises a kinematic simulation that determines whether a corresponding operation is allowed without collisions or impossible machine movements. In some cases, the simulation uses geometrical and kinematical state snapshots of rest material and tool locations relative to a rest material geometry as inputs to a physical feedback model that estimates a force regime and outcomes comprising deformation and vibration amplitude.

[0007] In some cases, the method further comprising transmitting the machineexecutable instructions to a controller of the manufacturing machine and causing, by the controller, the manufacturing machine to perform the sequence of manufacturing operations. In some cases, the machine-executable instructions comprise toolpath geometry for at least one operation, the toolpath geometry comprising one or more of: a tool center point trajectory, a surface engagement trajectory, or a layer path. In some cases, the machineexecutable instructions encode one or more process parameters comprising at least two of: feed rate, spindle speed, step-over, axial depth of cut, layer thickness, acceleration limit, jerk limit, or tool orientation.

[0008] In some cases, selecting the working tool for a candidate option comprises selecting, from a tool catalog, a tool assembly including a cutting tool and at least one of a tool holder or a tool extender. In some cases, selecting the working tool is further based on at least one geometric dimensioning and tolerancing (GD&T) requirement for a geometry of the workpiece, including a surface finish requirement. In some cases, the scoring function comprises a machine learning model comprising at least one of a neural network or a large language model configured to output a numeric value for the candidate option. In some cases, the project setup comprises one or more of: machine CAD and specification, fixture CAD and specification, stock material CAD and specification, or special GD&T requirements. InP-654716-PC some cases, the current state comprises at least: a rest material CAD representation, a target CAD representation, and a current operation tree. In some cases, computing the toolpath force comprises using a force model that takes into account at least: tool geometry, feed and speed, and an engagement area between a tool and the workpiece at a plurality of points along a toolpath curve. In some cases, generating the machine-executable instructions comprises generating numerical control instructions configured for execution by a CNC controller. In some cases, the method is performed in real time during a CAM programming session responsive to a change in the operation tree.

[0009] Embodiments may include a system for computer aided manufacturing of an item, comprising (a) a database storing at least one of CAD files, CAM files, machine specifications, tool specifications, or fixture specifications; and (b) one or more processors configured to perform the methods disclosed herein. Embodiments may include a non-transitory computer-readable medium storing instructions that, when executed by one or more processors, cause the one or more processors to perform the methods disclosed herein.BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Some embodiments of the invention are described herein, by way of example only, with reference to the accompanying drawings. With specific reference now to the drawings in detail, it is stressed that the particulars shown are by way of example and for purposes of illustrative discussion of embodiments of the invention. In this regard, the description taken with the drawings makes apparent to those skilled in the art how embodiments of the invention may be practiced.

[0011] In the drawings :

[0012] Figure 1 shows a computerized environment used in executing the processes elaborated herein, in accordance with some embodiments of the subject matter;

[0013] Figure 2 shows a method of designing the manufacture of an article, in accordance with exemplary embodiments;P-654716-PC

[0014] Fig. 3 shows a computer-implemented method for computer aided manufacturing of an item, in accordance with exemplary embodiments;

[0015] Fig. 4 shows a computer-implemented method for designing and executing the manufacture of an item using graph-based scoring, in accordance with exemplary embodiments;

[0016] Fig. 5 shows a computer-implemented method for pruning a graph of optional actions to manufacture an item, in accordance with exemplary embodiments;

[0017] Fig. 6 shows a computer-implemented method for the generation and execution of machine-executable instructions for controller-driven manufacturing, in accordance with exemplary embodiments.DETAILED DESCRIPTION OF THE INVENTION

[0018] At least some implementations described herein address the technical problem of CAM programming workflows that rely heavily on a user’s manual selection of 3D toolpaths and machining parameters. Current CAM programming workflows lack physical feedback mechanisms to predict or prevent machine overload, tool breakage, fixture failures, or dimensional deviations. Historically, these risks were managed by seasoned CNC programmers who accumulated extensive trial-and-error knowledge over the years. However, the shortage of such seasoned CNC programmers leaves manufacturing companies without the necessary insights to confidently program CNC operations. Moreover, for widespread adoption of a new, more automated approach, the optimization of each step in the CAM program must occur in real time during the programming session — rather than as a standalone post-processing tool — so that users can proactively adjust machining parameters and ensure dimensional accuracy, operational safety, and overall efficiency in a single integrated workflow.

[0019] Embodiments include a computing system comprising a database configured to store CAD / CAM / PDF files that represent history projects for CNC production as well as databases of accessories and required elements involved in each production project such as clamping devices (a.k.a fixtures) list, cutting tools list , tools holders, tool extenders andP-654716-PC additional types of tools. In addition, the database may store textual informational material, such as specifications of machines, tools and fixtures, for example, in the format of PDF brochures. The computing system may also comprise a processor for managing the processes disclosed herein. The processor may run on a server accessible from a working terminal of a CAM programmer. The processor may access a neural network, language models and additional models to execute at least some of the processes described herein.

[0020] Figure 1 shows a computerized environment used in executing the processes elaborated herein, in accordance with some embodiments of the subject matter.

[0021] The computerized environment may include a models layer configured to receive prompts from users and / or components of the systems disclosed herein and output text or a numeric value. The models layer may comprise at least one of a Feature Recognition model, an operation generation model and a Physical Feedback model.

[0022] The Feature Recognition model 110 may be configured to receive a CAD file in the format of STEP / X T / IGES or other variations of B-reps. The CAD file may represent the final desired CAD file or represent a middle stage of the rest material in case there is a multi-stage production process that involves more than one fixture and one machine process to produce the part. The Feature Recognition model 110 may output a segmentation of the CAD into different features. The features may be either user-defined types (e.g., holes, grooves, faces) or a result of a process that clusters them automatically. The Feature Recognition model 110 may be trained for example, using a Graph Neural Network (GNN) on CAD B-reps, a Convolutional Neural Network (CNN) on a voxel representation of the CAD, or any other 3D processing tool.

[0023] The Operation Generator model 112 may be configured to receive a CAD file, its segmentation into features, the current state of the material in the machine, and other required metadata (machine information, material type, tool set, etc.) and output a recommended operation and a tool used to perform the recommended operation. The operation generator model 112 may also include (RL & Search Tree) and a Physical Feedback model (PINNs based on Deformation & Modality Simulation and equations).P-654716-PC

[0024] The models layer may interact with a machining kinematic Simulator 120, for example when required to execute an operation suggestion including a toolpath in order to receive feedback from the simulator if from a kinematic perspective, a specific operation is allowed and does not create collisions or impossible machine movements. Moreover, the machining kinematic simulator 120 enables the creation of geometrical and kinematic state snapshots of the rest materials and tool locations related to the rest material geometry to then be processed by the physical feedback 115 layer to estimate the force regime between the tool and rest material and the outcomes of deformation and vibration amplitude of the two.

[0025] The computerized environment may include an agent layer 130 having an interface for users to manage manufacturing and design processes. The agent layer 130 may comprise a CAD DFM agent, a process designer agent and a CAM co-pilot agent.

[0026] The CAD DFM agent may be a third-party plug-in to CAD software that can analyze the CAD part in the design stage and indicate geometrical features and dimensional tolerances that are problematic or challenging to produce as well as estimating the production cost of a part and the estimated lead time. The agent has a local software representation side to interact with the CAD and user and cloud / server side to run the Al models involved in computing and analyzing the 3D model and data. The CAD DFM Agent may be configured to drive to conclusions of Design for Manufacturing (DFM) feedback due to integration with the Process Design and CAM Agent in a way that allows understanding the manufacturing capabilities of the relevant company where the CAD and CAM users belong.

[0027] The Process Design Agent may be a standalone web-based platform (no CAD / CAM dependency) that allows users to upload a CAD file with PDF or PMI requirements and to automatically build the different stages involved in producing the part (a process is a list of stages). Each stage generation includes choosing a machine, fixture, stock material and special additional requirements from the available elements of the client's production floor. The platform estimates the machining time of the process and each stage to allow accurate process quoting afterwards. The platform allows the user to easily edit each Al decision. At the end each process can be exported directly to the CAM and start the CAMP-654716-PC programming session with the guidance of the CAM agent. The user of this platform is the CAM programmer and the CNC programming team manager that is responsible for process development and machining time estimation for each request.

[0028] CAM Agent - a third-party plug-in to CAM software that can analyze the CAD part with its dimensional requirements and stage setup environment (fixture, stock material, machine) and suggest a best next operation to create and on user approval can also automatically generate this operation in the CAM environment. The agent has a local software representation side to interact with the CAM and user and cloud / server side to run the Al models involved in computing and analyzing the 3D model and data.

[0029] The computerized environment may include a client layer 140. The client layer 140 may comprise a library of tools and specifications of the tools. The client layer 140 may also comprise the specification of the machines used to manufacture the article. The client layer 140 may also comprise metrics and performance information of the machines.

[0030] Figure 2 shows a method of designing the manufacture of an article, in accordance with exemplary embodiments. The method may be performed on the system disclosed herein.

[0031] In operation 210, project setup and a current state of the design process may be received. The project setup may include machine CAD and specifications, fixture CAD and specifications, stock material CAD and specifications, and special GD&T requirements. This type of input may be defined at the beginning of a CAM session to set the proj ect environment and limitations. This data enables an increase in the likelihood that the recommendations are within the limitations of each machine, fixture and material type. The current ‘state’ may be sent to the Al every time there is a change in the operations tree in the CAM session or in case the user's activity requests a different suggestion than what has been offered by the Al in the first place. A state may include may rest material CAD, target CAD, and current operation tree.

[0032] In operation 220, a working tool may be selected for a plurality of optional operations. When selecting a working tool, the tool selection module may receive as inputP-654716-PC the current setup limitations, such as fixture load capacity, machine spindle RPM and torque capacity, material type. The tool selection module may also receive as input specific geometric dimensioning and tolerancing (GD&T) requirement for the specific geometry. For example, a high surface finish will require a very stable and rigid tool and tool holder to reduce the vibration amplitude. The tool selection module may also receive as input the applied geometry in the rest material where the tool is required to operate. For example, if there is a deep cavity with an inner corner radius of 1.5 mm, the selected tool is required to be long enough, rigid enough and to have the right tool tip radius to create the 1.5mm corner. The tool selection module may select the tool from an existing tool catalogue, or from a closed list of tools. The selection of the tool may include a combination of a cutting tool, tool holder, and sometimes a harbor extender.

[0033] In operation 230 additional parameters may be selected. The additional parameters may include feed, speed, step over, axial depth of cut. The additional parameters may be selected based on at least the following data sources: 1. The tool's recommendation and limitations, for example as appearing in a database. 2. The client's history of parameters for that type of tool, for example based on a user's behavior profile created based on prior operations and selections made by a specific user or a group of users. 3. The physical feedback that takes into consideration properties such as the specific rest material geometry rigidity and shape, the tool assembly rigidity and the machine limitations. The feedback may be provided from a physics-informed neural network (PINN).

[0034] The additional parameters' selection may vary with the type of operation selected. Some operations are simple, for example, 3 -axis planar milling or 2-axis turning. Other operations are more complex, for example 5 axis simultaneous that require additional inputs to complete the operation. The additional inputs may include a type of toolpath curve, what to do with the feed, speed in the corners, etc.

[0035] In operation 240, the toolpath force may be computed. The toolpath force may include cutting forces expected along a candidate toolpath. The computing may be done based on inputs such as tool geometry, engagement conditions, and the selected process parameters (for example, feed and speed). This toolpath force estimate may then be used asP-654716-PC a technical basis for downstream physical evaluation, such as predicting overload conditions or deformation / vibration risk.

[0036] In operation 250, a score for each of the plurality of optional operations on the raw material may be computed. The score may be computed using a scoring function that takes into account at least a simulation of the operation (and, in some embodiments, an estimated next workpiece state). The resulting scores enable ranking or selecting among alternative operations to guide the manufacturing plan toward the target geometry with better predicted outcomes (for example, efficiency, safety, or accuracy)

[0037] In operation 260, a physical simulation may be performed. The physical simulation may include deformation analysis for the rest material shape in the relevant area where the force is applied by the tool. The physical simulation may include deformation analysis for the tool assembly caused by the cutting force vector, for example provided from the PINN. The physical simulation may include modality analysis to identify natural frequencies of the tool setup and the rest material geometry to avoid feed and speed conditions that will resonate the setup, for example provided from the PINN. The physical simulation may include load calculations on the fixture and on the tool based on defined equations. The physical simulation may include surface finish geometrical calculations.

[0038] Fig. 3 shows a computer-implemented method for computer aided manufacturing of an item, in accordance with exemplary embodiments.

[0039] In operation 310, the system may receive project setup and current state of the manufacturing process. The project setup may define the manufacturing environment, such as machine configuration, fixture configuration, stock material information, and any applicable requirements (e.g., tolerances). The system may also receive a current state of the manufacturing process, which includes a representation of the workpiece (e.g., rest-material geometry and / or target geometry) and an operation tree reflecting what operations have been planned or performed.

[0040] In operation 320, the system may select for each candidate action, a working tool + compute toolpath force. Selecting a tool assembly may be done from a tool catalog (e.g.,P-654716-PC cutting tool with a tool holder and optionally a tool extender). The system may also compute a toolpath force for a toolpath associated with the candidate action, for example based on tool geometry, feed and speed, and engagement area between the tool and workpiece along the toolpath.

[0041] In operation 330, the system may perform a simulation based on the selected tool and toolpath force. The simulation can include physical effects such as deformation of the workpiece, deformation of the tool assembly caused by cutting-force vectors, vibration amplitude estimation, fixture load calculations, and / or surface finish geometric calculations. The simulation may also include a kinematic simulation to determine whether the operation is feasible without collisions or impossible machine movements

[0042] In operation 340, the system may compute score per candidate action using simulation of the operation. The score may be computed using at least a simulation of the operation corresponding to the candidate action and may also account for an estimated next state of the workpiece resulting from the candidate action.

[0043] In operation 350, the system may select a sequence of actions to manufacture the item. The sequence of manufacturing actions may produce the item from the current state toward the target geometry. The selected sequence may include multiple operations and may specify corresponding working tools for the respective operations.

[0044] In operation 360, the system may generate machine-executable instructions to manufacture the item. The machine-executable instructions may be numerical control (NC) instructions suitable for execution by a CNC controller and may encode toolpath geometry and process parameters for one or more operations. The generated instructions may be output in a controller-interpretable format so that the manufacturing machine can execute the operations.

[0045] Fig. 4 shows a computer-implemented method for designing and executing manufacture of an item using graph-based scoring, in accordance with exemplary embodiments.

[0046] In operation 410, the system may receive project setup and current state. The project setup may define the manufacturing environment, such as machine configuration,P-654716-PC fixture configuration, stock material information, and any applicable requirements (e.g., tolerances). The system may also receive a current state of the manufacturing process, which includes a representation of the workpiece (e.g., rest-material geometry and / or target geometry) and an operation tree reflecting what operations have been planned or performed.

[0047] In operation 420, the system may generate a graph of candidate options for a subsequent manufacturing step and permissible transitions. The graph data structure may include nodes that represent candidate options for a subsequent manufacturing step, and edges that represent permissible transitions between those candidate options. Each node may encode an operation type (e.g., drilling, face milling, slotting), a target feature or region of the workpiece, and associated state context such as the rest-material condition and setup parameters. The graph may provide a structured representation of alternative “next-step” choices and the allowed ways those choices may be sequenced

[0048] In operation 430, the system may perform initial filtering to retain candidate options. The filtering process retains a set of candidate options from among a larger set of possible options represented in the graph. The filtering may remove options that are considered inapplicable given the project setup and current state, for example options inconsistent with the machine configuration, fixture configuration, stock geometry, or stated requirements.

[0049] In operation 440, for each candidate option in the subset, the system may select a working tool and compute toolpath force.

[0050] In operation 450, for each retained candidate option, the system may simulate the candidate operation and estimate a next state. The system simulates the candidate operation to predict effects of executing that operation on the workpiece and / or manufacturing process. The system may estimate a next state corresponding to performing the candidate option, where the next state can be represented as an updated rest-material geometry, tool location state, or other state snapshot usable for later evaluation. The simulation and estimated next state may provide predictive inputs for scoring, enabling the system to compare candidate options using anticipated outcomes rather than only static geometry. The system simulation may be based on the selected tool and computed toolpath force. The simulation may includeP-654716-PC physical effects such as deformation of the workpiece, deformation of the tool assembly due to force vectors, vibration amplitude estimation, fixture load calculations, and surface finish geometric calculations, and may also include kinematic feasibility checks. The simulation outputs may be used for determining feasibility and for excluding unsafe or non-executable candidate options.

[0051] In operation 460, the system may remove unfeasible candidate options from the graph based on simulation results and / or toolpath force, thereby pruning the graph. Unfeasible options can include options predicted to cause collisions, exceed axis travel or kinematic reachability, overload spindle torque capacity, exceed cutting force or fixture load thresholds, risk tool breakage, or induce chatter instability. Pruning can be implemented by deleting nodes / edges or marking them disabled so that later route selection does not traverse them.

[0052] In operation 470, the system may compute a score for each retained candidate option using a scoring function based on the simulation and the estimated next state. The scoring function may output a numeric value indicating the relative desirability of selecting the candidate option for the subsequent manufacturing step, such as favoring options that reduce machining time, reduce risk, or improve accuracy. The scoring can be implemented using a machine learning model (e.g., a neural network or large language model) or another programmed scoring routine.

[0053] In operation 480, the system may select a subset of the candidate options for the subsequent step based on the score. The selection may be performed by ranking candidates according to score and choosing a top N set, choosing candidates above a threshold, or applying additional selection constraints such as diversity across operation types. The selection may focus subsequent, more computationally intensive evaluation on the most relevant options

[0054] In operation 490, the system may select a route of actions along the pruned graph, the route specifying a sequence of operations and corresponding working tools. The selected route may be chosen to reduce predicted adverse outcomes such as tool breakage, fixtureP-654716-PC failure, chater occurrence, or dimensional deviation while still achieving progress toward the target geometry.

[0055] Fig. 5 shows a computer-implemented method for pruning a graph of optional actions to manufacture an item, in accordance with exemplary embodiments.

[0056] In operation 510, the system may receive a graph of candidate options and transitions. The graph data structure may represent or model alternative manufacturing choices as nodes (candidate options) and allowed sequencing relationships as edges (permissible transitions). Each node may encode, for example, a specific machining operation type, a target feature / region of the workpiece, and associated setup context (machine, fixture, stock, and rest-material state). The graph may serve as the decision space that subsequent constraint checks and pruning operations evaluate to remove unsafe or infeasible options before selecting a route

[0057] In operation 520, the system may receive simulation outputs and computed toolpath force values for candidate options. The simulation outputs may include, for example, kinematic feasibility indicators, predicted collisions, deformation estimates, vibration / chatter indicators, and predicted next-state snapshots, while toolpath force values may characterize predicted cutting loads along the toolpath.

[0058] In operation 530, the system may evaluate machine constraints for each candidate option. Machine constraints for each candidate option may include kinematic reachability and collision avoidance, and whether the machine can reach required poses without exceeding axis travel limits and whether the planned motion would result in collisions. This evaluation may be performed using a kinematic model of the machine and a collisionchecking routine that accounts for the machine structure, tool assembly, fixture geometry, and workpiece / rest-material geometry.

[0059] In operation 540, the system may evaluate force-related constraints for each candidate option. Force-related constraints may include spindle torque capacity, cutting force thresholds, tool breakage thresholds, and fixture load capacity, evaluates force-related constraints may include comparing the computed toolpath force values against limits such asP-654716-PC spindle torque capacity, maximum allowable cutting force, tool breakage thresholds, and fixture load capacity. The comparison can be performed per toolpath segment, per operation, or using aggregate statistics (e.g., peak load, RMS load, or safety factor) to determine whether executing the candidate option would overload the machine, tool, or fixture.

[0060] In operation 550, the system may evaluate chatter stability for each candidate option based on modality analysis and natural frequency identification. The evaluation may be performed to predict whether the tool-workpiece-fixture system is likely to enter unstable vibration during the proposed operation. The evaluation may consider the dynamic stiffness of the setup, the operating spindle speed and feed, and the excitation frequencies implied by tooth passing or periodic engagement, to determine stability margins. The output may include a stability assessment (e.g., stable / unstable classification or a stability score) that is used alongside machine and force constraints to decide whether the option should be pruned

[0061] In operation 560, the system may identify unfeasible candidate options based on one or more constraint evaluations. For example, a node may be marked unfeasible if it violates any required constraint (collision predicted, axis over-travel, torque overload, fixture overload, or chatter instability), or if it fails to satisfy a minimum safety margin.

[0062] In operation 570, the system may remove unfeasible candidate options (nodes and / or transitions) from the graph to generate a pruned graph. The system may remove the identified unfeasible candidate options by deleting corresponding nodes and / or edges from the graph, thereby generating a pruned graph. Removal can be implemented as physical deletion from the data structure or as logical disabling (e.g., setting feasibility flags) such that later route selection will not traverse those nodes / transitions. This pruning reduces the search space to candidate sequences that satisfy the evaluated constraints, enabling per forming simulation of more steps ahead of the current state of manufacture.

[0063] In operation 580, the system may output the pruned graph for route selection. The output may be provided to another software module, stored for later use, and / or presented to a user interface so a CAM programmer can review remaining options and their feasibility rationale.P-654716-PC

[0064] Fig. 6 shows a computer-implemented method for Generation and execution of machine-executable instructions for controller-driven manufacturing, in accordance with exemplary embodiments.

[0065] In operation 610, the system may receive a selected route of actions specifying a sequence of operations and corresponding working tools. The route of actions may define an ordered sequence of manufacturing operations (for example, facing, pocketing, drilling, additive deposition, or inspection) and associates each operation with a corresponding working tool (for example, a specific end mill, drill, or nozzle). The route may be represented as structured data identifying operation order, tool identifiers, and any required setup context such as work coordinate system or fixture selection. This operation may provide the high-level plan that downstream operations convert into controller-ready, machine-executable instructions.

[0066] In operation 620, the system may generate machine-executable instructions for the sequence of operations. In this process, the system may convert the received route of actions into machine-executable instructions that a machine controller can interpret and run, such as numerical control (NC) instructions (e.g., G-code, M-codes, or another controllerspecific format). This may include translating each planned operation into a corresponding instruction sequence, including tool-change commands, motion commands, and control commands, organized in the same order as the route. Operation 620 may depend on operation 610 for the operation sequence and tool mapping and may create the instruction “container” into which geometry and process parameters are encoded in operations 630 and 640.

[0067] In operation 630, the system may encode toolpath geometry in the machineexecutable instructions. The system may encode toolpath geometry into the machineexecutable instructions by specifying the spatial motion the tool should follow, such as lines, arcs, splines, and / or point-to-point moves in a defined coordinate frame. The geometry encoding may include explicit axis positions (e.g., X / Y / Z and optional rotational axes), interpolation types, and transitions between segments so the controller can generate continuous motion. This operation may supply the “where to move” portion of theP-654716-PC instructions produced in operation 620, which may later be combined with the “how to move” parameters encoded in operation 640

[0068] In operation 640, the system may encode process parameters in the machineexecutable instructions. The process parameters may include feed rate, spindle speed, step-over, axial depth of cut, layer thickness, acceleration limit jerk limit, and / or tool orientation. The process parameters may constrain or command the controller’s execution behavior so that the toolpath geometry from operation 630 is traversed using appropriate cutting or deposition conditions. Operation 640 is related to operation 630 because both sets of data must be coherent (e.g., feed rate compatible with curvature and acceleration limits), and together they define an executable program derived from operation 620 including at least one of feed rate, spindle speed, step-over, axial depth of cut, layer thickness, acceleration limit, jerk limit, or tool orientation.

[0069] In operation 650, the system may validate the machine-executable instructions against machine and fixture limitations. The machine limitations may include axis travel, maximum spindle speed, maximum feed / acceleration / jerk, kinematics, and controller constraints and fixture limitations may include clamps, vise boundaries, keep-out zones, and collision risk. The validation may include detecting unreachable poses, singularities, overtravel conditions, or predicted collisions between the tool / toolholder and the fixture or workpiece.

[0070] In operation 660, the system may transmit machine-executable instructions to a controller of a manufacturing machine. Transmitting may be performed via a network connection, removable media, or a direct communications interface. The transmission can include the full program as a file or a streamed sequence of commands, optionally accompanied by metadata such as program identifiers, tool lists, or versioning

[0071] In operation 670, the system may cause, by the controller, the manufacturing machine to execute the operations according to the transmitted instructions. Causing may be done by issuing drive commands to motors / actuators, coordinating axis motion, and controlling process components such as spindle speed or deposition rate. The controller mayP-654716-PC interpret the encoded geometry and process parameters to produce time-sequenced actions, including tool changes, position control, and any auxiliary functions required by the program.

[0072] In operation 680, the system may monitor execution feedback and record performance results. The system may record performance results, which can include actual cycle time, deviations from commanded trajectories, tool wear indicators, quality measurements, or error logs, in a persistent store for traceability and later analysis.

[0073] Embodiments also include a method of selecting a working tool to perform a subsequent step in manufacturing an article, in accordance with exemplary embodiments. The method may be performed on the system disclosed herein. Estimating the deformation of the workpiece may be trained on a dataset of force-deformation FEA (finite elements analysis) use cases of variety of basic geometrical shapes. The tool selection model may also be trained on a neural network that can take a complex geometry as an input and create a combination of simple shape geometries and estimate the total deformation for a given force based on the combination of simple shape geometries. The tool selection model may utilize a set of physical force-deformation equations and use the equations as input of a loss function of the neural network (for example simple beam shape geometry bending equation).

[0074] Embodiments also include a method of computing the Toolpath force to perform a subsequent step in manufacturing an article, in accordance with exemplary embodiments. The method may be performed on the system disclosed herein. In some cases, the force calculation may be defined with close equations that take into consideration the tool geometry, feed and speed and the engagement area between the tool and the rest material for every point in the curve of the toolpath. The method may employ a dataset of this calculation for different tools type and geometries and train a neural network to predict the force regime for a given toolpath and rest material geometry.

[0075] The descriptions of the exemplary embodiments have been presented for purposes of illustration but are not intended to be exhaustive or limited to the embodiments disclosed. Many modifications and variations will be apparent to those of ordinary skill inP-654716-PC the art without departing from the scope and spirit of the embodiments described. The terminology used herein was chosen to best explain the principles of the embodiments, the practical application or technical improvement over technologies found in the marketplace, or to enable others of ordinary skill in the art to understand the embodiments disclosed herein.

[0076] It is expected that during the life of a patent maturing from this application many relevant files and / or records will be developed, and the scope of the term file and / or record is intended to include all such new technologies a priori.

[0077] As used herein the term “about” refers to ± 10 %.

[0078] The terms "comprises", "comprising", "includes", "including", “having” and their conjugates mean "including but not limited to". This term encompasses the terms "consisting of' and "consisting essentially of'.

[0079] The phrase "consisting essentially of' means that the composition or method may include additional ingredients and / or steps, but only if the additional ingredients and / or steps do not materially alter the basic and novel characteristics of the claimed composition or method.

[0080] As used herein, the singular form "a", "an" and "the" include plural references unless the context clearly dictates otherwise. For example, the term "a compound" or "at least one compound" may include a plurality of compounds, including mixtures thereof.

[0081] The word “exemplary” is used herein to mean “serving as an example, instance or illustration”. Any embodiment described as “exemplary” is not necessarily to be construed as preferred or advantageous over other embodiments and / or to exclude the incorporation of features from other embodiments.

[0082] The word “optionally” is used herein to mean “is provided in some embodiments and not provided in other embodiments”. Any particular embodiment may include a plurality of “optional” features unless such features conflict.

[0083] Throughout this application, various embodiments may be presented in a range format. It should be understood that the description in range format is merely for convenienceP-654716-PC and brevity and should not be construed as an inflexible limitation on the scope of the invention. Accordingly, the description of a range should be considered to have specifically disclosed all the possible subranges as well as individual numerical values within that range. For example, description of a range such as from 1 to 6 should be considered to have specifically disclosed subranges such as from 1 to 3, from 1 to 4, from 1 to 5, from 2 to 4, from 2 to 6, from 3 to 6 etc., as well as individual numbers within that range, for example, 1, 2, 3, 4, 5, and 6. This applies regardless of the breadth of the range.

[0084] Whenever a numerical range is indicated herein, it is meant to include any cited numeral (fractional or integral) within the indicated range. The phrases “ranging / ranges between” a first indicate number and a second indicate number and “ranging / ranges from” a first indicate number “to” a second indicate number are used herein interchangeably and are meant to include the first and second indicated numbers and all the fractional and integral numerals therebetween.

[0085] It is appreciated that certain features, which are, for clarity, described in the context of separate embodiments, may also be provided in combination in a single embodiment. Conversely, various features, which are, for brevity, described in the context of a single embodiment, may also be provided separately or in any suitable subcombination or as suitable in any other described embodiment. Certain features described in the context of various embodiments are not to be considered essential features of those embodiments, unless the embodiment is inoperative without those elements.

[0086] Although the specific embodiments has been described, it is evident that many alternatives, modifications and variations will be apparent to those skilled in the art. Accordingly, it is intended to embrace all such alternatives, modifications and variations that fall within the spirit and broad scope.

[0087] It is the intent of the applicant(s) that all publications, patents and patent applications referred to in this specification are to be incorporated in their entirety by reference into the specification, as if each individual publication, patent or patent application was specifically and individually noted when referenced that it is to be incorporated hereinP-654716-PC by reference. In addition, citation or identification of any reference in this application shall not be construed as an admission that such reference is available as prior art to the invention. To the extent that section headings are used, they should not be construed as necessarily limiting. In addition, any priority document(s) of this application is / are hereby incorporated herein by reference in its / their entirety.

Claims

P-654716-PC CLAIMS:

1. A computer-implemented method for computer aided manufacturing of an item, the method comprising:receiving, by one or more processors, a project setup and a current state of a manufacturing process for the item, the current state comprising at least a representation of a workpiece and an operation tree;selecting, for each candidate option, a working tool, additional parameters and computing a toolpath force for a toolpath associated with the candidate option; computing, for each candidate option of a subsequent step of manufacturing the item, a score using a scoring function, wherein an input to the scoring function comprises at least a simulation of an operation corresponding to the candidate option; performing a simulation based on the selected working tool and the computed toolpath force;selecting a sequence of manufacturing operations and corresponding working tools to manufacture the item;generating machine-executable instructions configured to cause a manufacturing machine to perform the sequence of manufacturing operations according to the selected sequence of manufacturing operations.

2. The method of claim 1, further comprising generating a graph data structure comprising nodes representing respective candidate options for the subsequent step of manufacturing the item and edges representing permissible transitions between the candidate options.

3. The method of claim 1, further comprising performing an initial filtering to retain, from among the candidate options, a set of candidates prior to computing the score.

4. The method of claim 1, wherein the input to the scoring function further comprises an estimated next state of the workpiece that is based on at least one of a machine configuration, a tool configuration, or a dynamic material state in the machine.P-654716-PC 5. The method of claim 1, further comprising selecting a subset of the candidate options based on the computed score, and wherein selecting the working tool and computing the toolpath force are performed for each candidate option in the selected subset.

6. The method of claim 1, further comprising computing, for each candidate option of at least one subsequent step reachable from the candidate options, a respective score.

7. The method of claim 2, further comprising removing, from the graph, unfeasible candidate options based on at least one result of the simulation and the computed toolpath force, thereby pruning the graph, and wherein selecting a route of actions is performed along the pruned graph.

8. The method of claim 7, wherein selecting the route of actions along the graph comprises selecting a route that reduces at least one of: predicted tool breakage, predicted fixture failure, chatter occurrence, or dimensional deviation.

9. The method of claim 7, wherein removing the unfeasible candidate options comprises removing candidate options that violate one or more machine constraints comprising at least one of: axis travel limits, kinematic reachability limits, or collision avoidance constraints.

10. The method of claim 7, wherein removing the unfeasible candidate options comprises removing candidate options that violate one or more force-related constraints comprising at least one of: a spindle torque capacity, a cutting force threshold, a tool breakage threshold, or a fixture load capacity.

11. The method of claim 7, wherein removing the unfeasible candidate options comprises removing candidate options associated with a predicted chatter instability based on a modality analysis that identifies natural frequencies of a tool setup and a rest material geometry.

12. The method of claim 1, wherein performing the simulation comprises performing the simulation based on the selected working tool and the computed toolpath force.

13. The method of claim 12, wherein performing the simulation comprises performing physical simulation including at least one of: deformation analysis for the workpiece,P-654716-PC deformation analysis for a tool assembly caused by a cutting force vector, vibration amplitude estimation, fixture load calculation, or surface finish geometric calculation.

14. The method of claim 12, wherein the simulation comprises a kinematic simulation that determines whether a corresponding operation is allowed without collisions or impossible machine movements.

15. The method of claim 12, wherein the simulation uses geometrical and kinematical state snapshots of rest material and tool locations relative to a rest material geometry as inputs to a physical feedback model that estimates a force regime and outcomes comprising deformation and vibration amplitude.

16. The method of claim 1, further comprising transmitting the machine-executable instructions to a controller of the manufacturing machine and causing, by the controller, the manufacturing machine to perform the sequence of manufacturing operations.

17. The method of claim 1, wherein the machine-executable instructions comprise toolpath geometry for at least one operation, the toolpath geometry comprising one or more of: a tool center point trajectory, a surface engagement trajectory, or a layer path.

18. The method of claim 1, wherein the machine-executable instructions encode one or more process parameters comprising at least two of: feed rate, spindle speed, step- over, axial depth of cut, layer thickness, acceleration limit, jerk limit, or tool orientation.

19. The method of claim 1, wherein selecting the working tool for a candidate option comprises selecting, from a tool catalog, a tool assembly including a cutting tool and at least one of a tool holder or a tool extender.

20. The method of claim 1, wherein selecting the working tool is further based on at least one geometric dimensioning and tolerancing (GD&T) requirement for a geometry of the workpiece, including a surface finish requirement.P-654716-PC 21. The method of claim 1, wherein the scoring function comprises a machine learning model comprising at least one of a neural network or a large language model configured to output a numeric value for the candidate option.

22. The method of claim 1, wherein the project setup comprises one or more of: machine CAD and specification, fixture CAD and specification, stock material CAD and specification, or special GD&T requirements.

23. The method of claim 1, wherein the current state comprises at least: a rest material CAD representation, a target CAD representation, and a current operation tree.

24. The method of claim 1, wherein computing the toolpath force comprises using a force model that takes into account at least: tool geometry, feed and speed, and an engagement area between a tool and the workpiece at a plurality of points along a toolpath curve.

25. The method of claim 1, wherein generating the machine-executable instructions comprises generating numerical control instructions configured for execution by a CNC controller.

26. The method of claim 1 , wherein the method is performed in real time during a CAM programming session responsive to a change in the operation tree.

27. A system for computer aided manufacturing of an item, comprising:(a) a database storing at least one of CAD files, CAM files, machine specifications, tool specifications, or fixture specifications; and (b) one or more processors configured to perform the method of any one of claims 1 - 26.

28. A non-transitory computer-readable medium storing instructions that, when executed by one or more processors, cause the one or more processors to perform the method of any one of claims 1 -26.

Images