Sensorless feed optimization

Abstract

This provides a sensorless method for optimizing the feed rate of a cutting tool. [Solution] This method calculates a feed rate profile that increases the feed rate at points in the machining cycle where the spindle torque falls below a target value. Time-series data, including position and spindle torque command value, is recorded for a reference machining cycle. The spindle torque command target value is determined based on the maximum value in the reference machining cycle or the machine's limits. At each time step, an optimal feed rate profile is calculated to determine the new feed rate by multiplying the original feed rate by the ratio of the spindle torque command target value to the spindle torque recorded in the reference machining cycle. The optimal feed rate profile is used by the control device in the production machining operation. The control device can use feedback control in conjunction with the optimal feed rate profile.

Description

[Technical Field]

[0001] This disclosure relates more broadly to the field of feed rate control for cutting tools, and more particularly to a feed rate optimization method that collects spindle torque command data at a number of position points within a reference machining cycle, and calculates an optimized feed rate profile by increasing the feed rate at positions within the machining cycle where the spindle torque falls below a target value, and vice versa. [Background technology]

[0002] The use of computer-controlled devices to perform machining operations such as drilling and milling on parts is well known in the art. In some applications, computer numerical control (CNC) machines are used to move a tool along three main directions, with or without changing the orientation of the tool. In other applications, multi-axis industrial robots equipped with machining heads can move the tool along any spatial path while controlling the orientation of the tool to any value.

[0003] Regardless of the type of machine tool or robot used to perform machining operations, the quality of the finished workpiece is always important, and conditions that could negatively affect workpiece quality or the lifespan of the machine tool must be avoided. At the same time, for manufacturers who must provide cost-competitive products, machine productivity is also extremely important. Therefore, the feed rate of the cutting tool must be determined in a way that satisfies both quality and productivity goals.

[0004] In a typical machining operation, the path the cutting tool takes to remove material from the workpiece is defined by a control program. In many machining operations, the amount of material removed by the cutting tool changes as it moves along the tool path, depending on the shape of the raw or finished workpiece. To prevent damage to the machine tool or workpiece, users can refer to the cutting tool catalog and select a tool feed rate appropriate for the tool path based on the depth of cut and the type of workpiece material. [Overview of the project] [Problems that the invention aims to solve]

[0005] Conventional machine control programs specify a constant feed rate throughout the entire machining operation, thereby preventing the maximum cutting load (and therefore spindle torque) from exceeding acceptable levels. While constant feed rate technology is easy to program, it is inefficient because the material removed by the cutting tool changes along the tool path. This means that to shorten cycle time while maintaining an acceptable cutting load, the cutting speed needs to be increased at many points along the tool path.

[0006] Techniques for improving the feed rate of cutting tools are well known in this art, but all of these techniques have certain drawbacks and limitations. One well-known technique is the use of a simulator, which uses a 3D model of the machining operation and the workpiece to estimate the volume of workpiece material being cut at every position along the programmed toolpath, and uses that cutting volume to calculate the feed rate at each point along the toolpath. However, these machining operation simulator systems are expensive, and simulating small-scale machining can be very time-consuming.

[0007] Another well-known technique for improving tool feed rate is to maintain a constant cutting torque by adjusting the feed rate in real time using online feedback control. While feed rate feedback control is effective in some applications, the nature of feedback control can sometimes inevitably lead to overshoot from the target torque value. Furthermore, since machining processes generally change over time, parameter adjustment of proportional-integral-derivative (PID) controllers can be unintuitive.

[0008] Given the above circumstances, there is a need for an improved cutting feed rate optimization method that can accurately calculate a feed rate profile that satisfies both load management and cycle time requirements in machining operations, without requiring simulation software. [Means for Solving the Problem]

[0009] The present disclosure describes a method for optimizing the feed rate of a cutting tool. The method collects spindle torque command data in a reference machining cycle and calculates an optimized feed rate profile that increases the feed rate at positions where the spindle torque falls below a target value in the machining cycle. For a plurality of time steps of the reference machining cycle, time series data including a tool center point position and a corresponding spindle torque command value is recorded. The spindle torque command target value is determined as a maximum spindle torque command value from the reference machining cycle, or based on a mechanical limit of a machine. A new optimized feed rate profile is then calculated, where a new feed rate is determined for the tool center point position at each time step by multiplying the original feed rate by the ratio of the spindle torque command target value to the recorded spindle torque at the time step of the reference machining cycle. The optimized feed rate profile is used by a machine control device in production machining operations. Feedback control can be incorporated into the control device to adjust the feed rate in real time based on actual spindle torque.

[0010] Additional features of the systems and methods of the present disclosure will become apparent from the following description and the claims when taken in conjunction with the accompanying drawings. [Brief Description of the Drawings]

[0011] [Figure 1] 1 is a schematic diagram of a system including a computer-controlled machine tool that performs a machining operation on a workpiece, of a type applicable to the technology of the present disclosure.

[0012] [Figure 2A] 2 is a diagram illustrating a cutting operation of a machine tool in which a tool moves at a constant feed rate using a conventional programming technique. [Figure 2B] 3 is a diagram illustrating a cutting operation of a machine tool in which a tool moves at a feed rate that varies based on an amount of material to be removed, according to an embodiment of the present disclosure.

[0013] [Figure 3] It includes a graph of spindle torque versus time and a graph of feed rate versus time, each graph comprising a data trace of conventional constant feed rate and a data trace of feed rate varying based on the amount of material to be removed according to an embodiment of the present disclosure.

[0014] [Figure 4] It is a block diagram of a sensorless feed rate optimization system with auxiliary feedback control, which comprises providing a reference feed rate from an optimized feed rate profile and adjusting the reference feed rate using feedback control according to an embodiment of the present disclosure.

[0015] [Figure 5] It is a flow chart of a sensorless feed rate optimization method, which comprises calculating an optimal feed rate profile based on spindle torque data from a reference machining cycle according to an embodiment of the present disclosure.

[0016] [Figure 6] It is a plan view of a feed angle working area of a machine tool in implementation of sensor-based feed rate optimization according to an embodiment of the present disclosure, in which a plurality of sectors are defined, and a target value is selected at each time step based on which sector the tool feed angle vector falls within.

[0017] [Figure 7] It is a flow chart of a sensor-based feed rate optimization method, which comprises calculating an optimal feed rate profile based on axial acceleration data from a reference machining cycle according to an embodiment of the present disclosure.

[0018] [Figure 8] It is a diagram of a machining operation provided to illustrate the concept used in the surface roughness prediction technology of the present disclosure, showing how the movement of a tool generates a scallop shape on the surface of a workpiece.

[0019] [Figure 9] This figure shows an ideal cutting tool with perfect shape and an actual cutting tool with runout, illustrating how tool runout affects the scallop shape of the workpiece after machining and how this effect can be simulated in embodiments of this disclosure.

[0020] [Figure 10] This figure shows the trajectory of the tip of a cutting tool groove on the operating surface and the corresponding surface profile, illustrating how the surface roughness of the workpiece is simulated in the embodiments of this disclosure.

[0021] [Figure 11] This is a flowchart of a method for calculating a machining feed rate profile according to an embodiment of the present disclosure, which first performs sensorless or sensor-based feed rate optimization, then evaluates surface roughness, and adjusts the feed rate profile as necessary.

[0022] [Figure 12] This figure shows a graphical user interface (GUI) screen of a software application configured to perform feed rate optimization for a machining operation, according to an embodiment of the present disclosure.

[0023] [Figure 13] Figure 12 shows the graphical user interface (GUI) screen after the feed rate optimization calculation has been performed according to an embodiment of this disclosure.

[0024] [Figure 14] This figure shows a workpiece machined by a tool path defined by a program including command lines having sparsely spaced control points, according to an embodiment of the present disclosure, and a workpiece in which the tool path is redefined by increasing the control point density.

[0025] [Figure 15]This figure shows a GUI screen of a software application configured to perform feed rate optimization and surface roughness prediction for machining operations, according to an embodiment of the present disclosure. [Modes for carrying out the invention]

[0026] The following descriptions relating to embodiments of the present disclosure concerning sensorless and sensor-based feed optimization are illustrative in nature and are not intended to limit the disclosed apparatus and technology, or their applications or uses.

[0027] Controlling the feed rate in machine tool operation is extremely important. This is because a feed rate with an excessive material removal rate can cause damage to the workpiece and / or tool, while a feed rate with an insufficient material removal rate unnecessarily lengthens the cycle time of the machining operation. This disclosure describes a feed rate optimization technique that can be implemented in an easy and cost-effective manner without requiring external sensors or using only a simple accelerometer with the machine tool, and without relying on expensive and cumbersome simulator software.

[0028] Figure 1 is a schematic diagram of a system 100 including a computer-controlled machine tool of a type applicable to the technology of this disclosure, which performs a machining operation on a workpiece. The machine tool 110 rotates a spindle 112 on which a cutting tool (in this case, an end mill 120) is fixed. The machine tool 110 causes the end mill 120 to machine the workpiece 130. The machine tool 110 communicates with a control device 140, which is a computer that provides the machine tool 110 with operation commands and spindle motor speed commands. In a typical example, the machine tool 110 moves the rotating end mill 120 from a starting point along a path to cut material from the workpiece 130, moves the end mill 120 away from the workpiece 130, returns the end mill 120 to a position near the starting point, and then generates another path to cut further material from the workpiece 130. The end mill 120 is shown in detail in an inset, with teeth or grooves shown at its tip 122. In this example, the end mill 120 includes four cutting edges or grooves. The sensor 150 will be discussed later in relation to sensor-based feed rate optimization technology.

[0029] As detailed below, the technology of this disclosure is applicable to system 100 in Figure 1. Specifically, the sensorless feed rate optimization method of this disclosure can be programmed into the control unit 140 using data readily available in existing controller architectures. In some embodiments, sensors, microphones, and other data acquisition devices are not required for data acquisition, and there is no need to integrate a separate data acquisition subsystem or sensor subsystem into the control unit 140.

[0030] The components in Figure 1 are shown in a relatively simplified manner, and the machine tool 110 is movable along three main operating axes, namely the "vertical direction" (parallel to the axis of the end mill 120) and two "horizontal directions" (perpendicular to the axis of the end mill 120). It should be understood that the sensorless and sensor-based feed rate optimization methods of this disclosure are applicable to all types of machine tools, such as multi-axis machine tools with tool positioning and orientation functions, and robot-controlled milling and drilling machines with articulated robot arms that provide complete flexibility in tool positioning and orientation.

[0031] Figure 2A is a diagram illustrating a cutting operation of a machine tool using conventional programming techniques, where the tool moves at a constant feed rate. The workpiece 210 is being machined by the cutting tool 220 along the tool path 230, as shown in the diagram. The machining operation in Figure 2A is one type of operation that can be performed by the system in Figure 1, where the workpiece 210 in Figure 2A corresponds to the workpiece 130 in Figure 1, and the cutting tool 220 corresponds to the end mill 120. In Figure 2A, the workpiece 210 is shown in a partially machined state, and the material to be removed from the workpiece 210 is shown in a darker color in the upper right portion of the workpiece 210. As shown in the diagram, the tool path 230 is configured such that the amount of material removed from the workpiece 210 changes as the cutting tool 220 moves along the tool path 230.

[0032] In conventional computer numerical control (CNC, or simply NC) machining operations, the feed rate of the cutting tool is constant and predetermined so that the tip load (especially torque) on the cutting tool 220 is maintained below a predetermined target value. The constant feed rate is indicated by a series of arrows 240 in Figure 2A. Graph 250 shows the relationship between material removal rate (MRR) and time in the machining operation shown in Figure 200. MRR is the volume of material removed from the workpiece 210 per unit time (e.g., cubic millimeters / minute). MRR is a function of the feed rate (or feed speed) of the cutting tool and the working volume (essentially the cutting thickness (vertical direction in Figure 2A) and the axial cutting depth (backward direction in the paper)). The torque on the cutting tool 220 (which is also the spindle torque) changes approximately proportionally to the MRR, which will be described in detail later. Therefore, both MRR and torque are labeled on the vertical axis of Graph 250.

[0033] As shown in Figure 200, when the feed rate is kept constant and the amount of material removed along the tool path changes, the MRR and spindle torque change in accordance with the amount of material removed. This is shown in Graph 250. That is, when the amount of material removed is small, as illustrated at the left end of the cutting tool 220 in Figure 2A, the MRR and spindle torque are small, as shown at time 252 in Graph 250. The reverse is also true; that is, when the amount of material removed is large, the MRR and spindle torque are also large.

[0034] The constant feed rate used in the machining operation shown in Figure 2A can keep the spindle torque below the target value, but it does not maximize the machine's productivity because higher cutting tool speeds can be used at many points along the tool path without exceeding the target spindle torque. The technology of this disclosure has been developed to satisfy both the spindle torque limiting (and cutting tool force) requirements and the machine's productivity requirements, and achieves this using a simple method that does not require an external sensor used with the machine tool.

[0035] Figure 2B is a diagram illustrating a cutting operation of a machine tool, according to one embodiment of the present disclosure, in which the tool moves at a feed rate that varies based on the amount of material being removed. In Figure 2B, the workpiece 210, cutting tool 220, and tool path 230 are the same as in Figure 2A. However, in Figure 2B, the feed rate of the cutting tool 220 changes as it moves along the tool path 230. Arrow 270 indicates that the feed rate decreases when a relatively large amount of material is being removed from the workpiece 210, and arrow 280 indicates that the feed rate increases when a relatively small amount of material is being removed from the workpiece 210.

[0036] Graph 290 shows the relationship between material removal rate (MRR) and time in the machining operation shown in Figure 260. By using the technology of this disclosure, which will be described later, the change in feed rate along the tool path is designed to satisfy the maximum spindle torque requirement by moving the cutting tool faster in areas with small depths of cut and slowing it down in areas with large depths of cut. As a result, while the MRR in Graph 250 fluctuates significantly, the MRR in Graph 290 is relatively uniform throughout the entire machining operation. This means that the spindle torque (and therefore the cutter tip load) is also relatively uniform in variable feed rate machining operations. As a result of the feed rate increasing in the area with small depths of cut, the machining cycle time is shortened in variable feed rate operations, which explains why the curve in Graph 290 is shorter (finishes earlier) than the curve in Graph 250.

[0037] Figures 2A and 2B, and their corresponding graphs, are conceptual and provided to illustrate the conventional constant feed rate and the feed rate optimization concepts of this disclosure. Details of the disclosed method and the results exhibiting the same characteristics as those shown in Figures 2A and 2B are described below.

[0038] JPEG2026142547000002.jpg61170

[0039] JPEG2026142547000003.jpg23170

[0040] In equation (2), the only variable on the right-hand side is the cutting feed distance c per tooth. All other values ​​are constants of the given machining operation (workpiece material, axial cutting depth, cutting tool characteristics (size, shape, and material)). Furthermore, if the spindle rotation speed is constant, the cutting feed distance c per tooth is proportional to the feed rate. Therefore, from equation (2), it can be seen that the cutting torque is basically proportional to the feed amount per tooth (i.e., T∝c), and thus the cutting torque is basically proportional to the feed rate.

[0041] Based on the relationship described above, the sensorless feed rate optimization technology of this disclosure uses spindle torque command data recorded from a reference machining cycle to calculate the ratio of the target spindle torque at each position step along the tool path to the recorded spindle torque command value, and uses this ratio to calculate a new feed rate to be used at each position step. The new feed rate profile along the tool path approximates the optimal feed rate profile for the reasons described above and below.

[0042] The first step in the disclosed technology is to record time-series spindle torque command data in a reference machining cycle. In a preferred embodiment, the machine tool is set up with a pre-programmed NC tool path, typically at a constant feed rate. As those skilled in the art will understand, the tool path may include many different cutting steps and "air-cut" steps, in which the cutting tool repeatedly performs positioning, cutting material from the workpiece, repositioning, performing the next cut, etc., until the machining operation is complete. Large air-cut steps are usually programmed to be completed using the maximum machine speed and acceleration, so there is no need to modify such air-cut steps for time optimization. Feed rate optimization is performed only for the cutting steps and for the small air-cut steps that exist within the programming of the cutting steps.

[0043] To understand the baseline machining cycle, consider a machining operation at a constant feed rate along the tool path 230, as shown in Figure 2A. At each incremental position of the tool tip along the tool path 230, the spindle torque is recorded along with the tool tip position. This generates a time-series dataset in which the spindle torque and tool tip position at each time step are recorded. The feed rate along the tool path 230 in the baseline machining cycle is also known. The time step increments (i.e., the time and distance between each time-series data point) can be selected according to application requirements.

[0044] After the reference machining cycle is completed, time-series data can be evaluated to determine the maximum value of the spindle torque command throughout the entire reference machining cycle. This maximum value is the target value of the spindle torque command SPTCMD. TARG It can be defined as: or the target value of the spindle torque command SPTCMD TARG This can also be determined based on machine requirements, such as a value equal to 20% of the maximum spindle torque. In either method, a constant feed rate for the reference machining cycle is selected so that the maximum spindle torque is equal to the value based on the machine requirements, thus SPTCMD TARG The values ​​should be similar.

[0045] JPEG2026142547000004.jpg79170

[0046] Note that the new feed rate profile completes the machining cycle in a shorter time than the standard machining cycle (as shown in the graphs in Figures 2A and 2B). Therefore, the new feed rate profile includes a set of positions (along the tool path) and their corresponding feed rates. The feed rate profile, including positions and feed rates, can be converted into a time-step-based operation program, where each time step includes a corresponding position on the tool path trajectory and a corresponding feed rate.

[0047] Referring again to Figure 2B, it can be seen that equation (3) provides a new feed rate profile that is much more time-optimal than the original constant feed rate, while ensuring that the commanded spindle torque does not exceed the target maximum value. At the point of cutting tool 220 immediately after arrow 270, the maximum amount of material is removed from the workpiece 210. Thus, at that point (SPTCMD) i The spindle torque command value in ) is the spindle torque command target value SPTCMD TARG It should be very close to this, which means that the feed rate at this point in the new speed profile will be very close to the original constant feed rate of the reference machining cycle.

[0048] Conversely, at the position of arrow 280, the amount of material removed from workpiece 210 is minimal. Therefore, at that point (SPTCMD) i The spindle torque command value in ) is the spindle torque command target value SPTCMD TARG This is significantly lower, meaning that the feed rate at this point in the new speed profile is much faster than the original constant feed rate of the reference machining cycle. The increased feed rate in the low material removal region allows machining operations to be completed more quickly using the new optimal feed rate profile.

[0049] When calculating a new speed profile using equation (3), control parameters may be defined to limit the increase or decrease in feed rate. For example, the feed rate at each point i in the new speed profile is limited to a range of 0.8 to 3.0 times the original feed rate. In other words, the new feed rate cannot be slower than 0.8 times the original feed rate, nor can it be faster than 3.0 times the original feed rate. These values ​​are merely examples, and the feed rate limits can be arbitrarily selected to suit the specific application.

[0050] It is emphasized that the spindle torque command data recorded in the reference machining cycle does not require a torque sensor inside the machine tool or on the cutting tool. The spindle torque command data is already known to the machine control device. The reason for this is that the control device monitors the spindle rotation speed and sends a torque command (via motor current) to the spindle motor configured to maintain the target spindle speed. Therefore, the above technology is sensorless, and uses the functions built into the machine tool (position and speed measurement) and the parameter data recognized by the machine control device.

[0051] In some embodiments, the spindle torque command time series data of the reference machining cycle can be preprocessed to remove air cut torque. This is because the total spindle torque is equal to the sum of the cutting torque and the friction torque of the machine tool itself. That is, T total =T cut +T friction holds. The friction torque T friction is the amount of torque required to rotate the spindle at the cutting speed without contact between the tool and the workpiece. Therefore, the friction torque can be measured during the air cut step.

[0052] In the embodiment where the air cut torque is removed, the air cut (friction) torque is also removed from the target spindle torque command SPTCMD TARG . Although the air cut (friction) torque is small compared with the actual cutting torque, removing the friction torque from the calculation of the optimal speed profile may provide slightly better results in some applications.

[0053] After the new feed rate profile is calculated as described above, a new motion program for the machining operation can be created by using the geometric shape of the original tool path and inserting new feed rate commands at each point on the tool path. In this resulting motion program, the feed rate along the tool path approximates the optimal feed rate profile. The new motion program acquires the feed rate profile (containing [position, feed rate] data pairs) and converts this into time-series data points for the new motion program using methods known in the art. This new motion program is then used by the machine control unit for the machining operation.

[0054] Figure 3 includes a spindle torque versus time graph 300 and a feed rate versus time graph 350, each graph including a data trace of a conventional constant feed rate and a data trace where the feed rate changes based on the amount of material removed according to one embodiment of the present disclosure. Graphs 300 and 350 represent experimental results of sinusoidal cutting path operations of the type shown in Figure 2 using the feed rate optimization technique described above.

[0055] In graph 300, data trace 310 plots spindle torque data for a conventional constant feed rate machining cycle, such as the one used for the reference machining cycle described above. Data trace 320 plots spindle torque data for machining cycles in which the operation program was created using the feed rate optimization technique described above. Line 330 represents the target spindle torque command SPTCMD TARG It represents.

[0056] Graph 300 shows that the spindle torque in the constant feed rate machining cycle (data trace 310) oscillates between a maximum and minimum value as the amount of material removed from the workpiece changes. The maximum value of the spindle torque in data trace 310 matches the target spindle torque, as expected. This is the behavior described above, as shown in the graph in Figure 2A. In contrast, the spindle torque in the optimal feed rate machining cycle (data trace 320) includes peaks and valleys, but the peaks last longer at the target spindle torque value, and the valleys do not decrease as sharply as in the constant feed rate data trace.

[0057] In Graph 350, a constant feed rate is shown in data trace 360. In contrast, the feed rate of the optimal feed rate machining cycle (data trace 370) increases sharply in the portion of the machining cycle where the amount of material removed is small, and decreases to the feed rate of the conventional constant feed rate machining cycle in the portion of the machining cycle where the amount of material removed is large. This is expected behavior given the feed rate optimization calculation method described above.

[0058] In the experimental implementations shown in Graphs 300 and 350, the maximum feed rate change in the optimized machining cycle was set to three times the constant feed rate of the reference machining cycle. This was achieved by applying a feed rate boundary limit, as described above. Comparing it to data trace 360 ​​in Graph 350, we can see that the feed rate has increased threefold in data trace 370. If the feed rate could be increased further (more than three times), the magnitude of the dip in data trace 320 in Graph 320 would decrease. However, at some point, if the feed rate increases excessively, the tool movement may become unacceptably jerky (the peak in data trace 370 becomes steeper), potentially exceeding the machine's acceleration and jerk limits.

[0059] The results shown in Figure 3 clearly demonstrate that the feed rate optimization technique of this disclosure enables machining cycles with significantly less spindle torque fluctuation than conventional constant feed rate machining cycles, while simultaneously guaranteeing that the maximum spindle torque does not exceed the target spindle torque value. Most importantly, machining cycles with feed rate optimization are completed in a much shorter time than conventional constant feed rate machining cycles. In this experimental implementation, the machining cycle time was reduced by more than 40% by using the feed rate optimization method of this disclosure. This reduction in cycle time can be confirmed in both graphs of Figure 3.

[0060] In another experimental implementation, the machining operation involved machining a workpiece from a solid block of material. The workpiece had numerous bosses, ribs, and cavities of varying heights and thicknesses, resulting in complex tool-workpiece engagement both on the inside and outside of the workpiece. In this experimental implementation, the maximum feed rate increase of the optimized machining cycle was set to twice the original feed rate. In this example, using the sensorless feed rate optimization method of this disclosure, the machining cycle time was reduced by approximately 20% compared to a constant feed rate machining cycle.

[0061] The feed rate optimization technique described above can be considered a type of feedforward control. In feedforward control, the feed rate is changed in anticipation of an imminent change in the amount of material removed from the workpiece. In another embodiment, feedforward control using feed rate optimization can be combined with real-time feedback control to achieve more robust and effective feed rate management.

[0062] In known feedback control systems for machine tools, spindle torque data, along with other parameters such as spindle temperature, is monitored in real time by the machine control unit. When the spindle torque value reaches a predetermined threshold (similar to the target spindle torque command mentioned above), the machine control unit reduces the feed rate as needed to return the spindle torque value to below the threshold. Due to the nature of feedback control, overshoot of the spindle torque threshold may be unavoidable, especially in high-speed machining. Combining feedforward control and feedback control of the machining operation allows for the best characteristics of both control strategies to be achieved.

[0063] Figure 4 is a block diagram of a sensorless feed optimization system 400 with auxiliary feedback control, according to one embodiment of the present invention, which includes providing a reference feed rate from an optimized feed rate profile and adjusting the reference feed rate using feedback control. In system 400, a machine control device 410 controls the machine tool 470 in the manner described above as shown in Figure 1.

[0064] In the feed rate optimization technique of the present disclosure, the control device 410 includes an operation program that includes the optimal feed rate for all position points along the tool path, in accordance with the sensorless feed rate optimization of the present disclosure detailed above. Next, at each time step during the actual execution of the machining cycle for the production part, the position of the tool tip from block 420 is provided to the operation program block 430, and the operation program block 430 determines the optimal feed rate based on the position of the tool tip along the tool path. In a basic implementation of the feed rate optimization method, the optimal feed rate from block 430 is provided to the machine tool 470, and the machine tool 470 uses the commanded feed rate.

[0065] In more advanced implementations, feedback control can be employed by adding elements within the dashed box 412. In this embodiment, the optimal feed rate from block 430 is provided as the reference feed rate to the feedback control block 440. The feedback control block 440 calculates the adjusted feed rate based on the difference between the reference spindle torque command value at the current time step (the torque value targeted by the feedback control, provided from block 450) and the actual torque command value at the previous time step (provided as feedback from the machine tool 470 via line 460). The feedback control block 440 may employ PID control or other feedback control algorithms well known in the art.

[0066] If the actual spindle torque command is greater than the reference torque command, the feedback control block 440 calculates an adjusted feed rate lower than the optimal feed rate from block 430. Conversely, if the actual spindle torque command is less than the reference torque command, the feedback control block 440 calculates an adjusted feed rate higher than the optimal feed rate from block 430. In the PID control system, the feed rate adjustment calculation is performed using proportional, integral, and differential logic, rather than based on simple differences. The output from the feedback control block 440 is the adjusted feed rate, which is provided as a command to the machine tool 470. The spindle torque command, designed to keep the spindle rotating at the target rotational speed, is also provided to the machine tool 470.

[0067] As shown in Figure 4, combining feedback control of the cutting tool feed rate with pre-preparation of an optimal feed rate profile provides the advantages of both feedforward and feedback control. The feedforward reference rate provides the theoretically optimal feed rate at every point along the tool path. Real-time feedback control, on the other hand, can make minor adjustments as needed to prevent cutting torque overshoot and, where possible, increase the feed rate. Adjustments by the feedback control module 440 are typically very minor, but may be necessary due to reasons such as the discrepancy between the nominal workpiece position specified in the machine tool's operating program and the actual workpiece position. Experimental implementations of this combined control strategy have demonstrated that it can very effectively reduce spindle torque command overshoot.

[0068] Figure 5 is a flowchart 500 of a sensorless feed rate optimization method according to one embodiment of the present disclosure, which includes calculating an optimal feed rate profile based on spindle torque data from a reference machining cycle.

[0069] In box 502, time-series data of a reference machining cycle is collected from the machine tool's control unit. The time-series data includes the position of the tool tip along a predetermined tool path and the corresponding spindle torque command value at each time step. In a preferred embodiment, the reference machining cycle is performed using an operation program (e.g., from NC or CNC) that maintains a constant feed rate for the cutting tool along the tool path. If variable feed rates are used in the reference machining cycle, the feed rate at each time-series data point is recorded along with the tool tip position and spindle torque command.

[0070] In box 504, the spindle torque command target value (SPTCMD) TARGThe spindle torque command target value is determined. The spindle torque command target value can be set as the maximum value of the spindle torque command over the entire reference machining cycle. Alternatively, the spindle torque command target value can be set based on the machine requirements, for example, to a value equal to 20% of the machine's maximum spindle torque. In either case, SPTCMD TARG The values ​​will be similar.

[0071] In box 506, the optimal feed rate profile for the tool path is calculated. As described above, for each time step i in the reference machining cycle dataset, a new optimal feed rate is calculated using equation (3). Here, the new feed rate is the original feed rate for the time step multiplied by PTCMD. i PTCMD TARG It is equal to the ratio multiplied by PTCMD i This is the spindle torque command value recorded for a specific time-series data point i. In the new optimal feed rate profile, the tool tip position (pos) corresponding to the time step i is used. i ) in a new feed rate optim,i This is used. As a result, a new feed rate profile is obtained in which the optimal feed rate is determined at each tool tip position along the tool path.

[0072] As described above, in calculating the optimal feed rate profile in box 506, limitations can be placed on the increase or decrease in feed rate calculated from equation (3). For example, the increase or decrease in the maximum feed rate can be limited to a 2x or 3x. The speed increase coefficient and speed decrease coefficient may be different, and these coefficients can be selected according to the requirements of the application.

[0073] Furthermore, as mentioned above, in calculating the optimal feed rate profile in box 506, the spindle torque value (PTCMD) is used. TARG and PTCMD i By pre-treating both of these factors, the air-cut (friction) torque can be subtracted. This allows the optimal feed rate to be calculated directly based on the torque related to material cutting, without considering parasitic friction torque.

[0074] Box 508 generates an operating program with an optimal feed rate profile, which is used by the machine control unit for production processing operations. The operating program used in Box 508 includes the optimal feed rates at each position along the tool path, calculated in Box 506. As a result, an operating program is obtained that follows the original tool path, with the feed rates along the tool path approximating the optimal feed rate profile.

[0075] In another embodiment, the machine control device performing the production work in box 508 uses a new operating program with an optimal feed rate profile, and also uses a feedback control block to adjust the commanded feed rate based on the actual spindle torque, as shown in Figure 4. By using feedback control in addition to the feedforward command for the optimal feed rate, it becomes possible to automatically fine-tune the feed rate in real time and address unexpected overshoot or undershoot of the target spindle torque.

[0076] The sensorless feed rate optimization technique described above has been demonstrated to be highly effective in shortening cycle times while maintaining the cutting tool load below target values, compared to constant-speed machining. Sensorless technology is applicable to machining operations where the amount of material removed from the workpiece is sufficient to cause large fluctuations in spindle torque along the machining trajectory of the tool path. However, in other applications, such as machining small components for electronic devices, the cutting tool diameter is small and the material removal rate is low, resulting in small fluctuations in spindle torque during the machining process, and sensorless spindle torque-based technology is not effective. In such applications, an acceleration sensor can be added to the machine tool, and the acceleration signal can be analyzed for feed rate optimization. This technology will be discussed later.

[0077] The fundamental premise of sensor-based feed rate optimization technology is that both the lateral and axial forces acting on the cutting tool are roughly proportional to the tool's feed rate. Even when the cutting tool is moving perpendicular to the tool axis, the cutting tool has a helical cutting edge shape (see cutting tool 120 in the inset in Figure 1), so axial forces (in addition to lateral forces) are generated by the cutting action of the tool. Furthermore, if the cutting tool has an axial motion component (for example, when it gradually bites into the workpiece), axial forces that depend on the feed rate are also generated.

[0078] JPEG2026142547000005.jpg50170

[0079] Instead of directly measuring the axial force acting on the cutting tool, it has been demonstrated that changes in the axial force acting on the cutting tool can be detected by processing and analyzing time-series axial acceleration data. Therefore, the Z-axis acceleration signal can be used as an indicator of the axial force acting on the tool, and this is related to the cutting force in the lateral direction (X-axis and Y-axis directions). The optimal feed rate profile can be calculated using the change in cutting force obtained from the axial acceleration signal.

[0080] Returning to Figure 1, the sensor 150 is mounted on the machine tool 110. In a preferred embodiment, the sensor 150 is an accelerometer, more specifically, an accelerometer configured to measure time-series acceleration data in the axial direction (parallel to the spindle, i.e., the Z direction indicated by the arrow). The sensor 150 collects time-series data during machining and provides the acceleration data to the control device 140. In the control device 140, the collected data (e.g., Z-axis acceleration data) is associated with the position of the tool tip along the tool path during machining, and the sensor data is then analyzed to calculate an optimal feed rate profile in a manner similar to the sensorless technology described above.

[0081] For the purposes of the following explanation, we define the acceleration parameter X used in feed rate optimization calculations. In one embodiment, the time series data X origTo generate the X-axis acceleration data, the time-series data from the axial acceleration sensor is double-integrated and high-pass filtered. Double integration converts the axial acceleration data into a value corresponding to the axial displacement, which changes according to the axial force. High-pass filtering removes drift from the time-series data. Double integration and high-pass filtering are performed from the raw time-series axial acceleration data. orig This is merely one example of a data processing technique for generating [the desired result]. Other techniques can be used as needed.

[0082] In some preferred embodiments, the acceleration sensor 150 provides data measurements at a faster rate than the time step increments used by the control device 140 to control the machine tool 110. Therefore, multiple axial acceleration data points are available at all time steps of the control device, preventing sudden increases and decreases in acceleration data from artificially influencing the acceleration parameter X.

[0083] JPEG2026142547000006.jpg41170

[0084] Similar to the sensorless feed rate optimization technique described above, the sensor-based method also begins with recording time-series data in a reference machining cycle, where axial acceleration data is recorded at each point along the tool path. In a preferred embodiment, the machine tool is typically set up by a pre-programmed NC tool path at a constant feed rate. During the reference machining cycle, the axial acceleration is recorded along with the tool tip position at each incremental position of the tool tip along the tool path 230 (Figure 2). This results in a time-series dataset in which the axial acceleration and tool tip position at each time step are recorded. The feed rate along the tool path 230 in the reference machining cycle is also known. The time step increments (i.e., time and distance between each time-series data point) can be selected according to the application requirements.

[0085] After the reference machining cycle is completed, the time-series data can be evaluated to determine the maximum value of the acceleration parameter X (calculated from axial acceleration data) throughout the entire reference machining cycle. This maximum value is the target acceleration amplitude value X. TARG It can be defined as follows.

[0086] JPEG2026142547000007.jpg75170

[0087] X i <X TARG In terms of the tool path, the optimal feed rate is increased from the original feed rate to improve the cycle time. i >X TARG Along the tool path, the optimal feed rate is reduced from the original feed rate to prevent overloading the cutting tool. The allowable increase or decrease in feed rate within the optimal feed rate profile may be limited, as described above in relation to equation (3).

[0088] Sensor-based feed rate optimization works in the same way and for the same reasons as described above with respect to Figure 2. That is, in areas of the tool path with low material removal, the feed rate increases above the nominal value, and conversely, in areas with high material removal, it decreases. When the sensor-based solution is applied to a sinusoidal tool path as shown in Figure 2, it yields an optimal feed rate profile similar to that shown and described above in Figure 3, in which much of the machining cycle operates at a feed rate significantly higher than the nominal constant feed rate. Again, this feed rate optimization is achieved using sensor-based technology even when spindle torque command data is unavailable because the machine control system cannot detect spindle torque fluctuations.

[0089] As shown in the graphs in Figures 2A and 2B, and as described above, it was confirmed once again that the new feed rate profile completes the machining cycle in a shorter time than the standard machining cycle. In the example above, the feed rate profile generated by the sensor-based method completed machining in approximately 32% less time than the machining cycle at a constant feed rate. The new feed rate profile includes sets of positions (along the tool path) and their corresponding feed rates, which the control device can use appropriately to control the operation of the machine tool according to the feed rate profile.

[0090] The advantages of combining feed rate feedback control with feedforward control were discussed above in relation to sensorless feed rate optimization techniques, and a block diagram of such a system is shown in Figure 4. The same approach can also be applied to sensor-based feed rate optimization. In that case, it should be noted that instead of using the actual torque command (known data to the machine control device) as feedback for line 460, the real-time axial acceleration signal from sensor 150 must be used as feedback.

[0091] In some embodiments of sensor-based feed rate optimization techniques, additional functionality may be added to address the problem of directional-dependent vibration cross-coupling in machine tools. As described above, consider a coordinate system for a machine tool where the Z-axis is parallel to the spindle (e.g., vertical) and the X and Y axes are oriented in the "forward" and "lateral" cutting directions (e.g., in the horizontal plane). Due to design considerations of the machine tool, the structure and mechanism for tool movement in one direction (e.g., X) differs from the structure and mechanism for tool movement in an orthogonal direction (e.g., Y). This means that the stiffness / flexibility characteristics differ between the X and Y directions, resulting in directional-dependent vibrations in the machine tool. Such vibrations can lead to cross-coupling effects that also depend on the structure of the machine tool. In other words, vibrations in the X direction are felt to some extent in the Y and Z directions, and vice versa. Furthermore, vibrations in the X direction result in different vibration characteristics in the Z direction than vibrations in the Y direction.

[0092] In other words, the acceleration in the Z direction resulting from cutting in the Y direction can be significantly different from the acceleration in the Z direction resulting from cutting in the X direction. Since sensor-based feed rate optimization methods are based on the magnitude of the acceleration in the Z direction, it can be advantageous to consider the cutting direction when processing acceleration data. This involves dividing the entire tool path into feed angle groups and assigning an X direction specific to each feed angle group. TARG This can be achieved by using values.

[0093] Figure 6 is a plan view of the feed angle workspace of a machine tool in the implementation of sensor-based feed rate optimization according to an embodiment of the present disclosure, where multiple sectors are defined, and a target value is selected at each time step based on which sector the tool feed angle vector belongs to. Figure 6 shows the X-axis 610 and Y-axis 620 of the machine tool coordinate system. In the machine tool coordinate system, the Z-axis extends downward parallel to the axis of the cutting tool. These definitions are consistent with the coordinate directions used throughout the present disclosure. The X and Y directions in Figure 6 relate to the velocity of the tool, not its position, as will be discussed later.

[0094] JPEG2026142547000008.jpg65170

[0095] JPEG2026142547000009.jpg41170

[0096] The approach to defining feed angle groups described above for quadrant 630 can, of course, also be applied to feed angles belonging to other quadrants (negative X and / or Y velocity directions). These other quadrants may be symmetrical to quadrant 630 or may be independently defined. Furthermore, the 30° sector is merely one non-limiting example; each quadrant may contain more or fewer than three sectors, and the sector sizes do not all need to be the same. The sector sizes of feed angle groups can be defined based on the characteristics of the machine tool to meet the application requirements.

[0097] Figure 7 is a flowchart 700 of a sensor-based feed rate optimization method according to one embodiment of the present disclosure, which includes calculating an optimal feed rate profile based on axial acceleration data from a reference machining cycle.

[0098] In box 702, time-series data of a reference machining cycle is collected from the machine tool control unit, and this time-series data includes axial acceleration data from sensors attached to the machine tool. The time-series data includes the position of the tool tip along a predetermined tool path and the corresponding axial acceleration value at each time step. In a preferred embodiment, the reference machining cycle is performed using an operation program (e.g., from NC or CNC) that maintains a constant feed rate of the cutting tool along the tool path. If a variable feed rate is used in the reference machining cycle, the feed rate at each time-series data point is recorded along with the position of the tool tip and the axial acceleration value.

[0099] Axial acceleration data from the sensor may be provided at a frequency higher than the control cycle frequency of the control device. In other words, multiple acceleration data points are received and recorded from the sensor for every time step of the control device that controls the machine tool along the tool path, increasing the amount of data that can be considered for averaging and other data processing techniques, and making it less susceptible to the influence of individual spikes in acceleration data.

[0100] In box 704, acceleration-related parameters X are calculated from raw time-series data of axial acceleration. As described above, the value of acceleration parameter X at each time-series data point can be the peak value of axial acceleration or the double integral of the acceleration signal. Other data processing techniques can also be applied to the original time-series acceleration data to provide acceleration parameter X, which is intended to represent the amplitude of axial acceleration (and thus the force acting on the cutting tool) at each time step of the machining cycle.

[0101] In box 706, the target value of the acceleration parameter X is determined. The target value is X TARGThis is expressed as and can be defined as the maximum value of the acceleration parameter X over the entire reference machining cycle. In some embodiments, each path point in the machining cycle is assigned to a feed angle group based on the direction of the tool velocity vector, and a target value X is set for each feed angle group. TARG This is calculated.

[0102] In box 708, the optimal feed rate profile for the tool path is calculated. As described above, for each time step i in the reference machining cycle dataset, a new optimal feed rate is calculated using equation (6), and this new feed rate is calculated by multiplying the original feed rate at that time step by X. i X for TARG It is equal to the value obtained by multiplying by the ratio of X, where X i This is the value of the acceleration parameter X at a specific time-series data point i. In the new optimal feed rate profile, the tool tip position (pos) corresponding to the time step i is used. i ) in a new feed rate optim,i This is used. As a result, a new feed rate profile is obtained in which the optimal feed rate is determined at each tool tip position along the tool path.

[0103] As described above, when calculating the optimal feed rate profile in Box 708, you can limit the increase or decrease in feed rate calculated from equation (6). For example, you can specify that the maximum feed rate can be increased or decreased by two or three times. The rate increase factor and rate decrease factor may be different, and these factors can be selected according to the application requirements.

[0104] Furthermore, as mentioned above, when calculating the optimal feed rate profile in box 708, each time-series data point can first be assigned to a feed angle group, and each feed angle group has a unique X used in equation (6). TARGIt has the value of . The feed angle group processing function cancels out the effects of cross-coupling of direction-dependent vibrations in machine tools and can yield better results in some applications (such as when the machining cycle includes a wide range of feed angles or when the stiffness characteristics of the machine tool differ significantly between the longitudinal and lateral directions).

[0105] In Box 710, an operating program is created using the optimal feed rate profile, and this program is used by the machine control unit for production machining operations. The operating program used in Box 710 includes the optimal feed rate at each position along the tool path, calculated in Box 708. As a result, an operating program is obtained that follows the original tool path, with the feed rate along the tool path approximating the optimal feed rate profile. Sensor-based solutions generate the optimal feed rate profile without processing spindle torque command data, which is essential in applications such as electronics machining where spindle torque fluctuations are too small to be detected.

[0106] In another embodiment, the machine control device performing production processing in box 708 uses a new operating program with an optimal feed rate profile, as well as a feedback control block that adjusts the commanded feed rate based on the actual axial acceleration value measured in real time, as described above. By using feedback control in addition to the feedforward command for the optimal feed rate, it is possible to automatically fine-tune the feed rate in real time to address unexpected overshoot or undershoot of the cutting force.

[0107] The sensorless and sensor-based feed rate optimization technologies described above provide powerful capabilities for creating feed rate profiles that minimize machining cycle time while avoiding excessive spindle torque and material removal rates. Another important factor in machining operations is the surface roughness of the workpiece after machining. Techniques for predicting nominal surface roughness in finishing operations are described below.

[0108] Figure 8 is a diagram of a machining operation, showing how a wavy shape is formed on the surface of a workpiece by the movement of a tool, and is provided to illustrate the concepts used in the surface roughness prediction technique of the present disclosure. Figure 8 shows the same type of machining operation as shown in Figures 1 and 2. A machine tool (not shown) moves in the direction indicated by the feed arrow 810. The machine tool has a cutting tool 820 (e.g., an end mill) on its spindle, which rotates at a spindle speed S and moves at a linear feed speed F. The spindle rotates about the Z axis, and the machine tool moves in a direction perpendicular to the spindle, in this case the X direction as shown in the coordinate system in Figure 8.

[0109] The workpiece 830 is fixed in place and machined by the cutting tool 820. With each rotation of the cutting tool 820, the flutes 822 cut and remove the cutting portion 832 from the workpiece 830. The size and shape of the cutting portion 832 depend on the spindle rotation speed S, the feed rate F, the diameter of the cutting tool 820, and the number of flutes on the cutting tool 820 (described later). When the feed rate F is very small, the cutting tool 820 makes a smooth surface on the workpiece 830. However, since the flutes 822 move in the X direction with each rotation, the cutting tool 820 actually makes the machined surface of the workpiece 830 "wavy," and this shape is particularly noticeable in the lower part of the machined surface shown in 834.

[0110] Figure 9 shows a cutting tool with an ideal shape and a real cutting tool with runout, illustrating how tool runout affects the wavy shape of the workpiece after machining, and how this effect can be simulated in the embodiments of this disclosure. The cutting tool 900 on the left has a rotating axis 910. The cutting tool 900 has a perfectly axially symmetric shape, and the tip diameters of each flute 920 are equal to each other. That is, under constant spindle speed and feed rate conditions, each flute 920 cuts the same amount of material from the workpiece with each rotation of the tool. Although the cutting tool 900 embodies an ideal tool, in reality, there is always some asymmetry in cutting tools.

[0111] The cutting tool 930 on the right has a rotation axis 940. Cutting tool 940 is an example of a cutting tool with runout. Runout is a term that refers to a state in which the pattern of cutting flutes is deviated from the rotation axis of the cutting tool. Cutting tool 930 has flutes 950 (shown by solid lines) that are laterally deviated to the right from the ideal flute pattern 960 (shown by dotted lines). This state is independent of the feed rate and feed direction and is a characteristic of the cutting tool 930 itself. The effect of tool runout is that a specific flute (952) that coincides with the direction of runout will cut most of the material from the workpiece with each rotation of the cutting tool 930. Below, we describe techniques for simulating the surface roughness of a workpiece based on a set of machining parameters. These techniques can optionally be configured to take into account the effect of tool runout.

[0112] The surface roughness of the workpiece during the lowering operation can be simulated based on parameters including the cutting tool diameter, feed rate, and spindle speed. This simulation will be explained with reference to the following figure.

[0113] Figure 10 shows the trajectories of the flute tips of a cutting tool on the plane of movement and the corresponding surface profiles, illustrating how the surface roughness of the workpiece is simulated in the embodiments of this disclosure. A cutting tool having two flutes, as shown in Figure 8, is modeled in Graph 1000. Following the convention described above, the cutting tool moves in the XY plane (the plane of movement of the cutting tool) perpendicular to the Z-axis, which is the axis of rotation of the spindle. The trajectory of the tip of one flute is represented by point 1010. As the cutting tool rotates and moves in the illustrated feed direction, point 1010 traces the trajectory 1020. The flute on the opposite side of the cutting tool traces a similar trajectory to trajectory 1020, but with a half-rotation phase shift.

[0114] Graph 1000 was created using the following technique. Each flute tip of the cutting tool is given an initial coordinate in the XY plane. For example, the flute tip of the cutting tool represented by point 1010 is given an initial coordinate corresponding to its position on the cutting tool's "equator" and its position backward in the feed direction. The flute on the opposite side is given an initial coordinate corresponding to its position on the cutting tool's equator and its position forward in the feed direction. Based on the feed rate and spindle speed, it is possible to determine how much each flute tip moves in the feed direction for each rotation of the cutting tool. For example, if the spindle speed is 1000 rpm and the feed rate is 100 mm / min, the cutting tool moves 0.1 mm / rev. Next, the spatial motion of point 1010 can be calculated at discrete points in each simulation step. In other words, when the spindle rotation is stepped by 1 degree in the simulation, point 1010 rotates 1 degree around the spindle axis at each step (new X and Y coordinates are obtained), and is also translated by an amount calculated by ΔX = (0.1 mm / rev) × (1 / 360 rev) in the X direction. Using the new X and Y coordinates resulting from the tool rotation and translation ΔX, the position of point 1010 at each simulation step can be calculated. Plotting the X and Y coordinates of point 1010 at all simulation steps yields the path 1020.

[0115] Graph 1030 shows the surface profile of the workpiece after cutting by the flutes of the cutting tool moving along path 1020 and the flutes on the opposite side. Graph 1030 is highly magnified, with the vertical axis (Y-axis) in micrometers, compared to tool path trace graph 1000, where the vertical axis units are millimeters. In Graph 1030, the surface shape generated by path 1020 is shown as profile 1040 (dark line font), and the surface shape generated by the flutes on the opposite side is shown as profile 1050 (light line font). From profiles 1040 and 1050, the surface roughness of the workpiece can be calculated in any appropriate way. Embodiments of surface roughness metrics obtained from the simulation include the absolute mean of the surface profile, the root mean square (RMS) of the surface profile, and the peak-to-valley height. Other metrics can also be used as needed.

[0116] The surface roughness simulation technique described above can be modified to simulate cutting tools with two or more (e.g., four) flutes, and / or to accommodate asymmetric cutting tool geometries (e.g., unequal pitch (flutes are not evenly spaced) or runout (as described above, one flute performs most of the cutting)). In either case, the simulation provides an index of the predicted surface roughness for a given set of feed rate and spindle speed conditions.

[0117] Parts machined using the machining operations described in this disclosure typically have surface roughness tolerances that must be met. Therefore, by comparing the calculated surface roughness prediction values, as shown in Figure 10, with the part tolerances, it is possible to determine whether the machining parameters will result in an acceptable workpiece surface finish / roughness. The objective of the sensorless and sensor-based feed rate optimization techniques described above is to increase the feed rate as much as possible where the material removal rate is low in the machining cycle. The speed increase obtained by feed rate optimization may result in an unacceptably high surface roughness. Therefore, it is desirable to evaluate the surface roughness after performing the feed rate optimization calculation, rather than before executing the optimized feed rate profile.

[0118] Figure 11 is a flowchart 1100 of a method for calculating a machining feed rate profile according to an embodiment of the present disclosure, which includes first performing sensorless or sensor-based feed rate optimization, then evaluating surface roughness, and adjusting the feed rate profile as necessary.

[0119] In box 1102, the optimal feed rate profile is calculated using either the sensorless or sensor-based feed rate optimization techniques described in detail above. As mentioned above, this results in a feed rate profile in which the optimal feed rate is determined at each tool tip position along the tool path. In many cases, the optimal feed rate profile includes higher feed rates than the original (e.g., constant rate) profile.

[0120] In box 1104, a surface roughness prediction simulation is performed using the techniques shown in Figure 10 and described above. Specifically, at least the fastest feed rate within the optimal feed rate profile is identified, and the surface roughness is predicted using that fastest feed rate and other relevant machining parameters (e.g., spindle speed, tool diameter). In the judgment diamond 1106, it is determined whether the surface roughness predicted in box 1104 is within an acceptable range. This determination may be made by a human evaluating the roughness prediction, or it may be made programmatically by comparing the predicted roughness with a threshold or specific value.

[0121] If the predicted surface roughness is unacceptable, the process proceeds from the judgment diamond 1106 to box 1108, where a modified feed rate profile is calculated. This involves reducing the portion of the feed rate profile that has the highest feed rate. The reduction amount can be determined in any appropriate way, such as a fixed percentage (e.g., 5%) or a percentage selected based on the required reduction amount of the surface roughness index. By blending the feed rate changes in a portion of the profile using smoothing or blending techniques, the feed rate in that portion can be reduced to a new maximum value. In the modified feed rate profile resulting from box 1108, the maximum feed rate is lower than that in the optimal feed rate profile from box 1102.

[0122] The process returns to box 1104, where the surface roughness is predicted using the modified feed rate profile. In judgment diamond 1106, it is determined again whether the surface roughness (this time based on the modified feed rate profile) is within the acceptable range. If it is outside the acceptable range, another modified feed rate profile with a further reduced maximum feed rate is calculated in box 1108. The loop process in boxes 1108 and 1104 can be repeated multiple times to achieve the desired surface roughness index. If the surface roughness predicted by judgment diamond 1106 is within the acceptable range, the process proceeds to box 1110, where the operation program using the final modified feed rate profile is used by the machine control unit for actual workpiece machining.

[0123] The above discloses techniques for optimizing machining feed rates using sensorless and sensor-based methods, as well as for evaluating the workpiece surface roughness of the optimal feed rate profile before applying it to production operations. All of these calculations and evaluations can be advantageously incorporated into software applications equipped with a graphical user interface (GUI) that allows users to configure machining operations.

[0124] Figure 12 shows an example of a graphical user interface (GUI) screen 1200 of a software application configured to perform feed rate optimization of a machining operation according to an embodiment of the present disclosure. The GUI screen 1200 is configured to guide the user through the feed rate optimization process in a highly automated manner.

[0125] In section 1210 at the top of GUI screen 1200, the user selects two input files. The first input file contains an NC program that defines the machining operation of the workpiece / part. The second input file contains data from the aforementioned reference machining cycle, i.e., tool position and motor torque data (in the case of a sensorless embodiment) collected from the reference machining cycle and used to calculate the optimal feed rate profile.

[0126] The optimization settings and configuration options are defined in section 1220 on the left side of screen 1200. The settings include specifying the time range for the optimization to be performed, which is typically the entire machining cycle. The settings also include selecting a target load level, which, as mentioned above, is a defined percentage of the maximum machine torque used to scale the optimal feed rate profile. The target load level can be set to 100% of the default value, but it can also be set to a higher or lower value; for example, a lower target value may be selected to extend tool life.

[0127] Increasing or improving the maximum feed rate is also selectable using the slider bar in section 1220. The selection of increasing the maximum feed rate (for example, three times the constant feed rate used in the machining program) is as described above. In the lower left, shown in 1230, there is a set of checkboxes for defining other configuration settings. The first two, removing friction torque during feed rate optimization calculations and enabling both increases and decreases in feed rate, have been described above. The third option, "Advanced Fine-Tuning," relates to a function described in relation to the diagram below.

[0128] After the settings are determined as desired, the user can click the optimization execution button 1240 to run the feed rate optimization calculation. To the right of the GUI screen 1200 is the graph section 1250. Before clicking button 1240 to run the feed rate optimization, the graph section 1250 contains a graph showing spindle torque data from a reference machining cycle as a curve 1260 and the target spindle torque as a straight line 1270. To the left of the Y axis, the percentage relative to the maximum spindle torque is labeled.

[0129] Figure 13 shows the graphical user interface (GUI) screen 1200 of Figure 12 after the feed rate optimization calculation has been performed according to an embodiment of the present disclosure. As described above, when the user clicks the "Execute Optimization" button 1240, the software application performs the feed rate optimization calculation and displays the results (the calculation is performed almost instantaneously). At this point, graph section 1250 is reformatted as graph section 1250A, with curve 1260 still containing spindle torque data for the reference machining cycle, line 1270 containing the target spindle torque, and an additional feed rate curve 1330. The feed rate curve 1330 plots the optimal feed rate as a percentage (%) relative to the original constant feed rate, as indicated by the label on the right side of the Y axis. It can be seen that the feed rate curve 1330 is higher as the spindle torque of the reference machining cycle decreases, and vice versa. This is the effect shown in the figure above and described in detail above.

[0130] Once the feed rate optimization calculation is complete, the program save button 1350 becomes active, allowing the user to save a new machining program containing the optimized feed rate profile. The saved new machining program can then be used for production machining operations. If the user desires further calibration settings, they can adjust the buttons and other controls in section 1220, and the optimize execution button 1240 will become active again.

[0131] The above software application and GUI screen 1200 are specifically configured to optimize the feed rate of the machining operation using sensorless technology based on spindle torque data in a reference machining cycle. The same application and a similar GUI screen can be provided to optimize the feed rate of the machining operation using sensor-based technology that utilizes vibration data in a reference machining cycle.

[0132] Another feature of the disclosed feed rate optimization technique is shown in the lower left of Figure 12 and is related to the "Advanced Fine-Tuning" checkbox mentioned above. This feature allows not only the optimization of the feed rate for a given machining program, but also the addition of new command lines to the program for finer feed rate adjustments. This feature will be discussed later.

[0133] Figure 14 shows a workpiece machined by a toolpath defined by a program including a command line with sparse control point spacing, and a workpiece with a redefined toolpath with a high control point density, according to an embodiment of the present disclosure.

[0134] As shown in Figure 1400 on the left, the workpiece 1410 has a wavy shape along its left edge. The workpiece 1410 is machined by the cutting tool 1420 along the tool path 1430 (dashed line) to have a flat surface along its left edge. The tool path 1430 is defined in the machining program by a set of command line control points 1432, 1434, 1436, 1438, etc. These control points are sufficient to define the tool path (a straight line in this case), but are spaced quite far apart. In the machining program, a unique feed rate can be defined for each command line containing each control point. However, even if a feed rate optimization calculation is performed on the tool path 1430, little improvement in feed rate is achieved. This is because each segment of the tool path (for example, from control point 1432 to control point 1434) contains sections with high material removal rates and sections with low (or zero) material removal rates, but only a single feed rate can be used in each segment.

[0135] The solution to the above problem is shown in Figure 1450 on the right. The workpiece 1410 is the same as described above. However, by using the technology of this disclosure, a new machining program is defined that includes additional command-line control points, allowing for a more appropriate application of feed rate optimization calculations. The cutting tool 1420 follows the same trajectory as the toolpath 1430, but is defined by a different set of command-line control points and is therefore identified as toolpath 1430A. Toolpath 1430A includes additional control points 1462, 1464, 1466, etc., in addition to the original control points 1432, 1434, etc. Performing a feed rate optimization calculation on toolpath 1430A allows for a more appropriate adjustment of the feed rate to the actual cutting conditions in each section of the toolpath. For example, in the toolpath section from control point 1462 to 1464, the amount of material removed from the workpiece 1410 is small or no, so the feed rate can be set to the maximum. Conversely, in the tool path section from control point 1434 to 1466, the amount of material removed is large, so the feed rate is maintained at a value close to the original feed rate to avoid exceeding the target spindle load.

[0136] The functions described above (adding command lines to the machining program by making the control point interval finer and optimizing the feed rate in the new machining program) in relation to Figure 14 are provided using the "Advanced Fine Tuning" checkbox shown in Figures 12 and 13. When this box is checked, the user enters the number of control points to add to the machining program. After performing feed rate optimization using the first few additional control points, the number of additional control points can be increased as needed, and the optimization calculation can be rerun until a suitable result is obtained. Fine tuning of the control point interval can also be automated. For example, when "Advanced Fine Tuning" is selected, the program can automatically insert additional command points if the spindle load changes by a predetermined rate between existing command points, and / or if the optimal feed rate changes by a predetermined rate between command points. This control point interval function can be configured by programming the GUI and its underlying algorithm, and / or by user-entered parameters.

[0137] In other embodiments, by adding the surface roughness prediction calculation and flowchart shown in Figure 11 to the software application, all functions can be provided to the user through a single interface. This will be discussed later.

[0138] Figure 15 shows a GUI screen 1500 of a software application configured to perform feed rate optimization and surface roughness prediction in a machining operation according to an embodiment of the present disclosure. GUI screen 1500 includes all the functions of the feed rate optimization GUI screen 1200 described above, including a configuration setting section 1220 and a graph section 1250A. Furthermore, GUI screen 1500 includes a surface roughness prediction section 1510, which allows the user to view the surface roughness prediction results of the optimized feed rate profile and modify the feed rate profile as needed to obtain the desired surface roughness characteristics of the workpiece.

[0139] After optimizing the feed rate and reviewing the results in sections 1220 and 1250A, the user can run a surface roughness prediction simulation on the optimized feed rate profile by clicking button 1520. The results of this surface roughness prediction simulation are displayed in Table 1530, which includes various surface roughness metrics. These metrics include, in non-limiting embodiments, maximum surface roughness, average surface roughness, RMS surface roughness, etc.

[0140] If the user wishes to improve the surface roughness from the optimized feed profile, this can be done on the right side of section 1510. In box 1540, the user can enter a target surface roughness index or a percentage reduction in the maximum feed rate in the optimized profile. In either case, when the user clicks button 1550, the software application executes the steps in the flowchart of Figure 11 to calculate a new feed rate profile with a lower maximum feed rate (for example, a rate lower than the 300% rate increase described in Figure 12) and simulates the surface roughness for the new feed rate profile. The predicted surface roughness results for the modified feed rate profile are displayed in Table 1560. If the user is satisfied with the results, they can save the machining program with the modified feed rate profile by clicking button 1570.

[0141] The software application, equipped with GUI screen 1500, provides users with all the functions of the feed rate optimization technology of this disclosure, including feed rate optimization for both sensorless and sensor-based use cases, flexibility in defining the amount of feed rate improvement, convenient configuration options, the ability to add command-line control points to the machining program to further optimize speed in localized sections of the machining operation, and integrated surface roughness prediction with unique configurability and target matching. The optimized operating profile using these functions can dramatically improve the work efficiency of the machine tool.

[0142] The preceding explanation described and suggested various computers and control devices. It should be understood that the software applications and modules for these computers and control devices run on one or more electronic computing devices having processors and memory modules. In particular, this includes the machine control device 140 in Figure 1 and the control device 410 in Figure 4. Some or all of the feed rate optimization calculations and surface roughness predictions can also be performed by another computing device communicating with the machine control device. Specifically, the processors of the control devices 140 / 410 and the other computing device are configured to perform the sensorless and / or sensor-based feed rate optimization described above, which includes the method steps in Figures 5, 7 and / or 11, calculations using equations (1)-(6) and other techniques described above, execution of the software applications and GUIs in Figures 12, 13 and 15, and control of the machine tool itself.

[0143] Having described several exemplary embodiments and models of sensorless and sensor-based feed rate optimization methods, those skilled in the art will recognize their modifications, permutations, additions, and subcombinations. Accordingly, the appended and subsequent claims are construed to include all modifications, permutations, additions, and subcombinations that fall within their true spirit and scope.

Claims

1. This is a method for optimizing feed rate without sensors. A machine control device collects time-series data of a reference machining cycle performed by a machine tool, wherein the time-series data includes the tool tip position along the tool path and spindle torque commands at multiple time steps. The steps include determining the target value of the spindle torque command, A step of calculating an optimal feed rate profile, comprising: for each time step of the time series data, calculating a new feed rate by multiplying the original feed rate from the reference machining cycle by the ratio of the spindle torque command target value to the spindle torque command for the time step; and pairing the new feed rate with the tool tip position in the optimal feed rate profile; The steps include controlling the machine tool that performs production processing work using the machine control device with the operation program created using the optimal feed rate profile, Methods that include...

2. The method according to claim 1, wherein the spindle torque command recorded in the reference machining cycle has a value calculated by the machine control device so that the spindle of the machine tool maintains a predetermined rotational speed.

3. The method according to claim 1, wherein the step of determining the spindle torque command target value includes selecting a maximum spindle torque command value from the reference machining cycle, or selecting a spindle torque command value based on the limitations of the machine tool.

4. The method according to claim 1, wherein the standard machining cycle is performed using a predetermined operation program that includes a constant feed rate.

5. The method according to claim 1, wherein calculating a new feed rate includes limiting the new feed rate by a maximum increase rate and a maximum decrease rate relative to the original feed rate.

6. The method according to claim 1, wherein the operation program is created by pairing each tool tip position in the optimal feed rate profile with a corresponding new feed rate, and defining an operation command for the machine tool that includes all the pairings.

7. The method according to claim 1, wherein, before calculating the optimal feed rate profile, the air cut torque value is subtracted from the spindle torque command target value, and further subtracted from the spindle torque command in the time-series data of the reference machining cycle.

8. The method according to claim 1, further comprising controlling the machine tool that performs the production processing operation by the machine control device using a feedback control algorithm and the optimal feed rate profile.

9. The method according to claim 8, wherein for each step of the operation program, an optimal feed rate is determined based on the tool tip position, the feedback control algorithm calculates an adjusted feed rate from the optimal feed rate and the difference between the reference torque command and the actual torque command of the previous step, and the adjusted feed rate and the new torque command are provided to the machine tool.

10. The method according to claim 8, wherein the feedback control algorithm uses proportional, integral, and derivative control.

11. The method according to claim 1, further comprising performing a workpiece surface roughness prediction simulation using the feed rate data, spindle rotation speed, cutting tool radius, and number of flutes in the optimal feed rate profile.

12. The method according to claim 11, wherein the surface roughness prediction simulation of the workpiece is performed to calculate the shape of the machined surface of the workpiece based on the distance traveled by the cutting tool in the feed direction of cutting by one of the flutes of the cutting tool.

13. The method according to claim 1, further comprising using a software application having a graphical user interface (GUI), wherein the user uses the GUI to define configuration settings for feed rate optimization and to view the graphical results of the feed rate optimization.

14. The method according to claim 13, wherein the GUI and the software application include user-selectable options, the options which include a plurality of additional command-line control points in the operation program, and an optimal feed rate profile is calculated for the operation program including the additional command-line control points.

15. The method according to claim 13, wherein the GUI and the software application include a surface roughness prediction simulation performed on the optimal feed rate profile, the simulation including the calculation of a modified feed rate profile that satisfies a user-defined surface roughness specification.

16. The method according to claim 1, wherein the machine tool is a multi-axis machine tool or an industrial robot in which a machining tool head is attached as an arm end tool.

17. This is a method for optimizing feed rate without sensors. A machine control device collects time-series data of a reference machining cycle performed by a machine tool, wherein the time-series data includes the tool tip position along the tool path and spindle torque commands at multiple time steps, and the spindle torque commands have values ​​calculated by the machine control device so that the spindle of the machine tool maintains a predetermined rotational speed. The steps include determining the target value of the spindle torque command, The steps include subtracting the air cut torque value from the spindle torque command target value, and further subtracting it from the spindle torque command in the time-series data of the reference machining cycle, A step of calculating an optimal feed rate profile, comprising: for each time step of the time series data, calculating a new feed rate by multiplying the original feed rate from the reference machining cycle by the ratio of the spindle torque command target value to the spindle torque command for the time step; and pairing the new feed rate with the tool tip position in the optimal feed rate profile; The steps include controlling the machine tool that performs production processing work using the machine control device with the operation program created using the optimal feed rate profile, Methods that include...

18. This is a sensorless system for optimizing the feed rate of machine tools. A machine tool configured to perform operations on a workpiece, The system includes a computing device configured to communicate with the machine tool and calculate and use the optimal feed rate by performing multiple steps, The aforementioned steps are: A step of collecting time-series data of a reference machining cycle performed by the machine tool, wherein the time-series data includes the tool tip position along the tool path and spindle torque commands at multiple time steps, The steps include determining the target value of the spindle torque command, A step of calculating an optimal feed rate profile, comprising: for each time step of the time series data, calculating a new feed rate by multiplying the original feed rate from the reference machining cycle by the ratio of the spindle torque command target value to the spindle torque command for the time step; and pairing the new feed rate with the tool tip position in the optimal feed rate profile; The steps include controlling the machine tool that performs the production processing work using an operation program created using the optimal feed rate profile, A system that includes this.

19. The system according to claim 18, wherein the spindle torque command recorded in the reference machining cycle has a value calculated by the calculation device such that the spindle of the machine tool maintains a predetermined rotational speed, and before calculating the optimal feed rate profile, the air cut torque value is subtracted from the spindle torque command target value, and further subtracted from the spindle torque command in the time series data of the reference machining cycle.

20. The system according to claim 18, wherein the step of determining the spindle torque command target value includes selecting a maximum spindle torque command value from the reference machining cycle, or selecting a spindle torque command value based on the limitations of the machine tool.

21. The system according to claim 18, wherein calculating a new feed rate includes limiting the new feed rate by a maximum increase rate and a maximum decrease rate relative to the original feed rate.

22. The system according to claim 18, wherein the operation program is created by pairing each tool tip position in the optimal feed rate profile with a corresponding new feed rate, and defining an operation command for the machine tool that includes all the pairings.

23. The system according to claim 18, further comprising using a feedback control algorithm and the optimal feed rate profile to control the machine tool that performs the production processing operation using the computing device.

24. The system according to claim 23, wherein for each step of the operation program, an optimal feed rate is determined based on the tool tip position, the feedback control algorithm calculates an adjusted feed rate from the optimal feed rate and the difference between a reference torque command and the actual torque command of the previous step, and the adjusted feed rate and the new torque command are provided to the machine tool.

25. The system according to claim 18, further comprising performing a workpiece surface roughness prediction simulation using the feed rate data, spindle rotation speed, cutting tool radius, and number of flutes in the optimal feed rate profile.

26. The system according to claim 25, wherein the surface roughness prediction simulation of the workpiece calculates the shape of the machined surface of the workpiece based on the distance traveled by the cutting tool in the feed direction of cutting by one of the flutes of the cutting tool.

27. The system according to claim 18, wherein the computing device executes a software application having a graphical user interface (GUI), and the user uses the GUI to define configuration settings for feed rate optimization and visually confirms the graphical results of the feed rate optimization.

28. The system according to claim 27, wherein the GUI and the software application include user-selectable options, the options of which include a plurality of additional command-line control points in the operation program, and an optimal feed rate profile is calculated for the operation program including the additional command-line control points.

29. The system according to claim 27, wherein the GUI and the software application include a surface roughness prediction simulation performed on the optimal feed rate profile, the simulation including the calculation of a modified feed rate profile that satisfies a user-defined surface roughness specification.

30. The system according to claim 18, wherein the machine tool is a multi-axis machine tool, or an industrial robot in which a machining tool head is attached as an arm end tool.

Images