Method and system for predicting maintenance time of machine tool and electronic equipment

Abstract

The invention belongs to the field of machine tools, and discloses a machine tool maintenance time prediction method and system and electronic equipment, and the method comprises the steps: obtaining a quadrant sharp angle error value time sequence of a machine tool; and predicting the maintenance time of the machine tool according to the quadrant sharp angle error value time sequence. According to the method, the simulation circle following track is generated by using the digital twin model of the machine tool, and the quadrant sharp angle error value is determined according to the polar coordinate deviation of the simulation circle following track and the theoretical circle following track, so that the quadrant sharp angle error value can accurately reflect the actual health condition of the machine tool. And then the maintenance time is predicted according to the quadrant sharp angle error value time sequence, so that the determination of the maintenance time is associated with the change condition of the actual health condition of the machine tool, the prediction accuracy of the maintenance time is improved, the scientificity and rationality of machine tool maintenance are effectively improved, the production loss and unnecessary production line shutdown are reduced, and the production efficiency is improved. And idle and waste of resources are avoided.

Description

Technical Field

[0001] The present invention relates to the field of machine tools, and in particular to a method, system, electronic device and computer-readable storage medium for predicting machine tool maintenance time. Background Art

[0002] Machine tools are controlled by pre-programmed computer software to perform precise machining and manufacturing operations. Program control systems are able to process programs specified by control codes or other symbolic instructions, controlling machine tools to perform various complex machining operations.

[0003] Existing machine tool maintenance is based on "preventive maintenance," meaning it's performed on a regular, routine basis, without fully understanding the machine's health. However, this approach has the following drawbacks: if the machine is overloaded or frequently operated, it could exceed precision standards before scheduled maintenance, resulting in production losses due to non-standard product dimensions. Alternatively, maintenance may be performed while the machine is still healthy, leading to unnecessary downtime on the production line.

[0004] Therefore, how to determine the precise maintenance time of a machine tool based on its health status is a technical problem to be solved by those skilled in the art. Summary of the Invention

[0005] In order to solve the problem in the prior art that it is impossible to determine the accurate maintenance time of a machine tool according to the health status of the machine tool, the present invention provides a method, system, electronic device and computer-readable storage medium for predicting the maintenance time of a machine tool.

[0006] A method for predicting machine tool maintenance time, comprising:

[0007] Obtaining a time series of quadrant cusp error values of a machine tool; wherein the quadrant cusp error value is determined by generating a simulated circular trajectory using a digital twin model of the machine tool and based on a polar coordinate deviation between the simulated circular trajectory and a theoretical circular trajectory;

[0008] The maintenance time of the machine tool is predicted according to the quadrant cusp error value time series.

[0009] Optionally, predicting the maintenance time of the machine tool according to the quadrant cusp error value time series includes:

[0010] Generate a state indicator time series according to the quadrant cusp error value time series; wherein the state indicator is the difference between a preset threshold and the quadrant cusp error value;

[0011] Using the state indicator time series to train a decay model, and using the trained decay model to predict the time when the state indicator becomes zero;

[0012] The time when the status indicator becomes zero is determined as the maintenance time of the machine tool.

[0013] Optionally, the method further includes:

[0014] In the decay model, predicting a confidence interval for the time at which the state indicator becomes zero;

[0015] When the quadrant cusp error value time series is updated, the state indicator time series is updated, and the parameter estimation deviation of the decay model is corrected according to the updated state indicator time series to compress the time confidence interval.

[0016] Optionally, obtaining a time series of quadrant cusp error values of a machine tool includes:

[0017] Constructing a digital twin model of the machine tool; the digital twin model integrates the LuGre friction model and the servo control loops of each axis of the machine tool;

[0018] The quadrant cusp error value acquisition process is triggered according to a preset period, and the obtained quadrant cusp error value and its corresponding acquisition time are stored in the quadrant cusp error value time series.

[0019] Optionally, the process of obtaining the quadrant sharp angle error value includes:

[0020] Obtaining parameters of the machine tool and assigning values to the digital twin model based on the parameters;

[0021] The assigned digital twin model is used to generate a simulated circular trajectory, and the polar coordinate deviation between the simulated circular trajectory and the theoretical circular trajectory is determined as the quadrant cusp error value.

[0022] Optionally, obtaining parameters of the machine tool and assigning values to the digital twin model based on the parameters includes:

[0023] Sending segmented constant speed and acceleration motion instructions to the machine tool, and collecting speed, acceleration and torque data of each axis motor;

[0024] Based on the speed, acceleration and torque data of the motors of each axis, the parameters of the LuGre friction model in the digital twin model are determined by the least squares method.

[0025] Optionally, obtaining parameters of the machine tool and assigning values to the digital twin model based on the parameters includes:

[0026] Sending segmented constant speed and acceleration motion instructions to the machine tool, and collecting speed, acceleration and torque data of each axis motor;

[0027] The motor parameters of the machine tool are obtained, and the servo loop parameters in the digital twin model are determined based on the speed, acceleration, torque data of the motors of each axis and the motor parameters.

[0028] A system for predicting machine tool maintenance time, comprising:

[0029] an acquisition module for acquiring a time series of quadrant cusp error values of a machine tool; wherein the quadrant cusp error values are determined by generating a simulated circular trajectory using a digital twin model of the machine tool and based on a polar coordinate deviation between the simulated circular trajectory and a theoretical circular trajectory;

[0030] A prediction module is used to predict the maintenance time of the machine tool according to the quadrant cusp error value time series.

[0031] An electronic device, comprising:

[0032] A processor and a memory, wherein the memory is used to store at least one instruction, and when the instruction is loaded and executed by the processor, the method for predicting machine tool maintenance time as described in any one of the above is implemented.

[0033] A computer-readable storage medium stores a computer program, wherein when the computer program is executed by a processor, the method for predicting machine tool maintenance time as described in any one of the above is implemented.

[0034] The method for predicting the maintenance time of a machine tool provided by an embodiment of the present invention obtains the quadrant cusp error value time series of the machine tool; and predicts the maintenance time of the machine tool based on the quadrant cusp error value time series. The present invention first generates a simulated circular trajectory using a digital twin model of the machine tool, and determines the quadrant cusp error value based on the polar coordinate deviation between the simulated circular trajectory and the theoretical circular trajectory, so that the quadrant cusp error value can accurately reflect the actual health status of the machine tool. The maintenance time is then predicted based on the quadrant cusp error value time series, so that the determination of the maintenance time is associated with the changes in the actual health status of the machine tool, thereby improving the accuracy of the maintenance time prediction, avoiding the problem of premature or late maintenance that may occur in the regular maintenance method, effectively improving the scientificity and rationality of machine tool maintenance, reducing production losses and unnecessary production line shutdowns, and avoiding idle and wasteful resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0036] Figure 1A flowchart of a method for predicting machine tool maintenance time provided by an embodiment of the present invention;

[0037] Figure 2 A schematic diagram of a machine tool transmission structure provided by an embodiment of the present invention;

[0038] Figure 3 A schematic diagram of a quadrant sharp angle provided by an embodiment of the present invention;

[0039] Figure 4 A schematic diagram of a true circular trajectory provided by an embodiment of the present invention;

[0040] Figure 5 A schematic diagram of a simulated circular trajectory provided by an embodiment of the present invention;

[0041] Figure 6 for Figure 1 A flowchart of a practical expression of S01 in a method for predicting machine tool maintenance time is provided;

[0042] Figure 7 A schematic diagram of a LuGre friction model provided in an embodiment of the present invention;

[0043] Figure 8 A schematic diagram of a model architecture of a servo control loop for each axis of a machine tool provided by an embodiment of the present invention;

[0044] Figure 9 A schematic diagram of a segmented constant speed and acceleration motion instruction provided by an embodiment of the present invention;

[0045] Figure 10 for Figure 1 A flowchart of a practical expression of S02 in a method for predicting machine tool maintenance time is provided;

[0046] Figure 11 A schematic diagram of a method for predicting the maintenance time of a machine tool using a decay model provided by an embodiment of the present invention;

[0047] Figure 12 This is a structural diagram of a machine tool maintenance time prediction system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0048] In order to better understand the technical solution of the present invention, the embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0049] It should be understood that the embodiments described are only a portion of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without creative work are within the scope of protection of the present invention.

[0050] The terms used in this embodiment are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The singular forms "a", "an", "the" and "the" used in the embodiments of the present invention and the appended claims are also intended to include plural forms unless the context clearly indicates otherwise.

[0051] It should be understood that the term "and / or" as used herein is merely a description of the relationship between associated objects, indicating that three possible relationships exist. For example, "A and / or B" can represent: A exists alone, A and B exist simultaneously, or B exists alone. Furthermore, the character " / " in this document generally indicates that the associated objects are in an "or" relationship.

[0052] Please refer to Figure 1 , which is a flow chart of a method for predicting machine tool maintenance time provided by an embodiment of the present invention, comprising the following steps:

[0053] Step S01, obtaining a time series of quadrant angular error values of a machine tool.

[0054] The quadrant cusp error value is determined based on the polar coordinate deviation between the simulated circular trajectory and the theoretical circular trajectory generated using the machine tool's digital twin model. The quadrant cusp error value time series contains the quadrant cusp error value and the corresponding acquisition time point for each quadrant cusp error value. The acquisition time point is used to identify the generation time of the error value, thus forming a complete time series dataset.

[0055] Please refer to Figure 2 and Figure 3 , Figure 2 A schematic diagram of a machine tool transmission structure provided by an embodiment of the present invention. Figure 3 A schematic diagram of a quadrant sharp corner following a circular trajectory provided by an embodiment of the present invention.

[0056] like Figure 2 As shown in the figure, the XYZ axis motor of the machine tool drives the platform to move through the gearbox and lead screw to realize the processing of the product on the platform. The machine tool will produce platform accuracy errors due to various factors, resulting in product size differences. One of the error sources is the friction between the lead screw and the platform. This friction will cause the motor torque to increase but the platform position to remain unchanged when the lead screw changes its rotation direction (rotating in one direction and then gradually stops and then rotates in the other direction), which will cause the platform error to occur when the motor starts to rotate. Figure 3 As shown, this platform error will appear in the form of quadrant protrusion when the platform moves to produce a circular trajectory.

[0057] As the machine tool is used for a longer time, wear and tear will cause the platform error to become larger, which in turn causes the quadrant sharp angle of the circular trajectory to become larger when the machine tool platform moves in a circular motion.

[0058] Please refer to Figure 4 and Figure 5 , Figure 4 A schematic diagram of a true circular trajectory provided by an embodiment of the present invention is shown in FIG. Figure 5 A schematic diagram of a simulated circular trajectory provided by an embodiment of the present invention.

[0059] like Figure 4 As shown in FIG, after a circular motion instruction is issued to the machine tool, the machine tool will execute the circular motion instruction to generate a true circular trajectory. The true circular trajectory is a circular trajectory with multiple protrusions ( Figure 4 The theoretical circular trajectory is a smooth circular trajectory ( Figure 4 middle green track).

[0060] The value obtained by subtracting the actual circular trajectory from the theoretical circular trajectory is the platform error generated by the machine tool. Excessive platform error will cause the dimensions of the products processed by the machine tool during circular motion to exceed the standard specifications. Therefore, the need for machine maintenance is determined by whether the platform error exceeds the threshold. However, the platform error requires the machine tool to stop production and perform circular motion, while also measuring data using measuring instruments. This makes the acquisition of platform error overly cumbersome and causes unnecessary production line downtime.

[0061] Based on this, this embodiment calculates the platform error through the digital twin model, that is, the value obtained by subtracting the simulated circular trajectory from the theoretical circular trajectory is determined as the platform error generated by the machine tool, so as to simplify the method of obtaining the platform error.

[0062] like Figure 5 As shown in Figure 2, after a circular motion instruction is issued to the digital twin model of the machine tool, the digital twin model will execute the circular motion instruction to generate a simulated circular trajectory. The simulated circular trajectory is a circular trajectory with a few protrusions ( Figure 5 The theoretical circular trajectory is a smooth circular trajectory ( Figure 5 middle green track).

[0063] like Figure 4 and Figure 5 As shown in Figure 2, both the real circular trajectory and the simulated circular trajectory produce quadrant sharp corners caused by friction at the two-axis reversal points (i.e., 0 degrees, 90 degrees, 180 degrees, and 270 degrees).

[0064] On this basis, in some embodiments, the simulated circular trajectory may include an XY plane simulated circular trajectory, a YZ plane simulated circular trajectory, and an XZ plane simulated circular trajectory;

[0065] Each simulated circular trajectory contains quadrant sharp angle error values in four directions: 0°, 90°, 180°, and 270°.

[0066] This embodiment uses a digital twin model of a machine tool to generate a simulated circular trajectory. The quadrant cusp error is determined based on the polar coordinate deviation between the simulated and theoretical circular trajectories. This quadrant cusp error is then used to characterize the magnitude of the platform error. If the quadrant cusp error exceeds a threshold, the machine tool is shut down for maintenance or error compensation to ensure product quality.

[0067] In this embodiment, the quadrant cusp error value time series refers to a series of data points of the quadrant cusp error values that are continuously obtained at certain time intervals during the operation of the machine tool, showing the changes over time. These data points can clearly show the fluctuation of the quadrant cusp error of the machine tool over time, and can provide a key basis for subsequent analysis of the accuracy change trend of the machine tool, judgment of the health status of the machine tool, and accurate prediction of the maintenance time of the machine tool.

[0068] Step S02 : predicting the maintenance time of the machine tool according to the quadrant cusp error value time series.

[0069] Since the quadrant cusp error value can reflect the change in the accuracy of the machine tool at the point where the quadrant passes through the quadrant, and this change is closely related to the health of the machine tool, by analyzing the quadrant cusp error value time series, the fluctuation trend of the machine tool accuracy over time can be captured. Therefore, in this embodiment, the change trend of the quadrant cusp error value can be predicted based on the quadrant cusp error value time series, and then the time when the quadrant cusp error value exceeds the preset threshold can be predicted, which indicates that the machine tool may be worn or malfunction at this time and requires maintenance. Compared with the traditional regular maintenance method, this embodiment can more accurately determine the actual maintenance needs of the machine tool, avoid production losses caused by untimely maintenance or waste of resources caused by excessive maintenance, thereby effectively reducing maintenance costs and improving production efficiency and equipment utilization.

[0070] In some embodiments, a suitable prediction model, such as a linear regression or machine learning model, can be established to predict when the quadrant cusp error value will exceed a preset threshold, i.e., when the machine tool requires maintenance. The specific process may be:

[0071] First, the acquired quadrant cusp error value time series needs to be preprocessed to remove noise and outliers and improve data quality; then, key features that can characterize the machine tool status, such as mean, variance, trend terms, and periodic terms, need to be extracted; then, a prediction model is constructed based on these features, using methods such as machine learning or deep learning; finally, the trained model is used in combination with the latest quadrant cusp error value data to predict the maintenance time of the machine tool, thereby achieving accurate maintenance, reducing maintenance costs and downtime, and improving production efficiency and equipment utilization.

[0072] Based on the above technical solution, the method for predicting the maintenance time of a machine tool provided by an embodiment of the present invention obtains the quadrant cusp error value time series of the machine tool; and predicts the maintenance time of the machine tool according to the quadrant cusp error value time series. The present invention first uses the digital twin model of the machine tool to generate a simulated circular trajectory, and determines the quadrant cusp error value according to the polar coordinate deviation between the simulated circular trajectory and the theoretical circular trajectory, so that the quadrant cusp error value can accurately reflect the actual health status of the machine tool. Then, the maintenance time is predicted based on the quadrant cusp error value time series, so that the determination of the maintenance time is associated with the changes in the actual health status of the machine tool, thereby improving the prediction accuracy of the maintenance time, avoiding the problem of premature or late maintenance that may occur in the regular maintenance method, effectively improving the scientificity and rationality of machine tool maintenance, reducing production losses and unnecessary production line shutdowns, and avoiding idleness and waste of resources.

[0073] Please refer to Figure 6 ,for Figure 1 A flowchart of a practical embodiment of S01 in a method for predicting machine tool maintenance time is provided. In some embodiments, step S01 mentioned in obtaining a time series of quadrant cusp error values of a machine tool may specifically include the following steps:

[0074] Step S11: construct a digital twin model of the machine tool.

[0075] The digital twin model integrates the LuGre friction model and the servo control loops of each axis of the machine tool.

[0076] In this embodiment, the digital twin model is a virtual model corresponding to the actual machine tool. Not only is its geometry identical to the real machine tool, but it also integrates the LuGre friction model and the servo control loops for each axis of the machine tool, enabling it to simulate the dynamic behavior and control characteristics of the machine tool during actual operation. The LuGre friction model accurately describes friction characteristics, while the servo control loops for each axis control the motion of each machine tool axis, ensuring it follows the predetermined trajectory and speed.

[0077] In some embodiments, please refer to Figure 7, which is a schematic diagram of a LuGre friction model provided in an embodiment of the present invention.

[0078] like Figure 7 As shown in Figure 2, the mathematical expression of the LuGre friction model is:

[0079]

[0080] Among them, F LuGre is the friction force, σ0 is the stiffness of the average burr, σ1 is the damping of the average burr, σ2 is the viscosity coefficient of the two sliding surfaces, z is the deformation of the average burr, is the differential of the average burr deformation, v is the relative velocity of the two sliding surfaces, F S is the static friction, F C is the Coulomb friction force, v s is the velocity when the Stribeck effect occurs, δ is the parameter used to modify the profile of the Stribeck curve, g(v) is the deformation, and sign(v) is the information function used to determine the sign of the relative velocity v.

[0081] In this example, the LuGre friction model is used to describe friction phenomena in the digital twin model of a machine tool. By integrating the LuGre friction model into the digital twin model, the errors and vibrations caused by friction during actual machining can be more accurately simulated, thereby improving the model's accuracy.

[0082] In some embodiments, please refer to Figure 8 , is a schematic diagram of a model architecture of a servo control loop for each axis of a machine tool provided by an embodiment of the present invention, wherein:

[0083] 1) POSC (position control command):

[0084] This is the input to the system and represents the desired position.

[0085] 2) Err (error calculation):

[0086] Calculate the error between the desired position (POSC) and the actual position (POSF).

[0087] 3) Kpp (position proportional gain):

[0088] Position proportional gain is a key parameter of the proportional controller, which is used to adjust the response speed and stability of the position control loop so that the proportional controller can generate a control signal based on the position error.

[0089] 4) Vcmd (speed command):

[0090] The speed command output by the position controller is used for the speed control loop.

[0091] 5) Speed loop (blue dotted box):

[0092] This is an inner loop that controls the speed of the motor. The output of this loop is the current command (Tcmd), which is used to control the current of the motor and thus the speed.

[0093] ①Kvp (speed proportional gain) and Kvi (speed integral gain):

[0094] These gains are used to adjust the response of the speed controller.

[0095] 6) Current control loop (purple dotted box):

[0096] This is the inner loop of the current control, which is used to accurately control the current of the motor.

[0097] ①Tcmd (torque):

[0098] The torque output by the speed control loop is used to control the speed of the motor.

[0099] ②1 / Js:

[0100] Indicates the motor's moment of inertia, which is used to calculate the motor's speed.

[0101] ③V fb (speed):

[0102] The motor's speed, used to calculate the output position.

[0103] 7) Friction:

[0104] represents the friction in the system, which affects the operation and control accuracy of the motor. In this embodiment, the friction is determined according to the above-mentioned LuGre friction model. Wherein, τf is the torque generated by the friction.

[0105] 8) Position control loop (green dotted box):

[0106] The position control loop is the outer loop in the servo system and is primarily responsible for position control. It continuously adjusts the output through a proportional controller to ensure that the actual position quickly and accurately follows the desired position.

[0107] 9) POSF (actual position feedback):

[0108] This is the output of the system and represents the actual position of the motor or platform.

[0109] 10) 1 / s (integrator):

[0110] Used to convert speed signal into position signal.

[0111] In this embodiment, by incorporating the servo control loop of each axis into the digital twin model, the impact of the control characteristics of the machine tool on the processing accuracy can be fully considered, and accurate simulation and prediction of the machine tool's operating status can be achieved.

[0112] In some embodiments, the process of building a digital twin model of a machine tool mentioned in step S11 may specifically be:

[0113] 1) Collect detailed information such as the machine tool's geometric structure, material properties, and dynamic parameters, and use professional modeling software (such as SolidWorks, UG, etc.) to create a three-dimensional geometric model of the machine tool.

[0114] 2) The LuGre friction model is integrated into the geometric model, and the model parameters are determined according to the actual friction characteristics of the machine tool.

[0115] 3) Establish a servo control loop model for each axis of the machine tool, including modeling and parameter setting of components such as controllers, drivers, and motors, to ensure that it can accurately reflect the actual control behavior of the machine tool.

[0116] 4) Verify and optimize the constructed digital twin model through experimental data, and adjust the model parameters to improve its consistency with the actual machine tool.

[0117] Step S12: triggering the process of obtaining the quadrant cusp error value according to a preset period, and storing the obtained quadrant cusp error value and its corresponding acquisition time in the quadrant cusp error value time series.

[0118] In this embodiment, the measurement and calculation program of the quadrant cusp error value is automatically started at fixed time intervals during the operation of the machine tool, and the error value obtained each time is associated with a specific time to form an ordered data pair. These data pairs are then added to the quadrant cusp error value time series in sequence, thereby realizing dynamic tracking of changes in machine tool performance.

[0119] This embodiment binds error values to acquisition times to form a time series, providing comprehensive historical data and powerful support for subsequent analysis of machine tool performance trends and maintenance time prediction. Furthermore, the preset period can be flexibly adjusted based on the actual frequency and importance of the machine tool, ensuring data timeliness without excessively increasing the burden of data storage and processing.

[0120] In some embodiments, the process of obtaining the sharp angle limit error value mentioned in step S12 may specifically include:

[0121] Step S21: Obtain the parameters of the machine tool and assign values to the digital twin model based on the parameters.

[0122] Different machine tools have unique parameters during their design, manufacturing, and use, which directly impact their performance and machining accuracy. Therefore, in this embodiment, by collecting machine tool parameters and applying them to a constructed digital twin model of the machine tool, the digital twin model accurately reflects the characteristics and behavior of the actual machine tool. This process ensures consistency in performance and functionality between the digital twin model and the actual machine tool, providing a foundation for subsequent simulation analysis and prediction.

[0123] Parameter assignment enables the digital twin model to be targeted and personalized, reflecting the characteristics of a specific machine tool. This helps provide more practical results and recommendations in subsequent applications such as maintenance prediction and performance optimization. Furthermore, as machine tools are used and wear out, their parameters may change. Regularly updating parameter assignments keeps the digital twin model synchronized with the actual machine tool, ensuring the model's timeliness and effectiveness.

[0124] In some embodiments, the step S21 mentioned above, the step of obtaining the parameters of the machine tool and assigning values to the digital twin model based on the parameters may specifically include:

[0125] Step S31: Send segmented constant speed and acceleration motion instructions to the machine tool, and collect speed, acceleration and torque data of each axis motor.

[0126] For example, see Figure 9 , is a schematic diagram of a segmented constant speed and acceleration motion instruction provided by an embodiment of the present invention. Figure 9 As shown, this embodiment inputs specially designed motion commands into the machine tool's control system, causing each axis of the machine tool to move according to predetermined constant velocity and acceleration patterns. During this process, sensors installed on each axis' motor collect real-time operating data such as motor speed, acceleration, and torque. This data can reflect the dynamic performance and load conditions of the machine tool under different motion states.

[0127] Step S32: Based on the speed, acceleration and torque data of each axis motor, various parameters in the LuGre friction model in the digital twin model are determined by the least squares method.

[0128] For example, after collecting the speed, acceleration, and torque data for each axis motor, a model equation containing unknown parameters can be established based on the mathematical expression of the LuGre friction model. An error function is then defined to measure the difference between the model output and the actual collected data. A least squares algorithm is then used to adjust the unknown parameters in the model to minimize the error function. This process can be implemented using optimization libraries in mathematical software such as MATLAB or programming languages such as Python.

[0129] In this embodiment, friction is one of the key factors affecting the machining accuracy and dynamic performance of machine tools. Accurate friction model parameters enable the digital twin model to more realistically reflect the actual operation of the machine tool, improving the model's simulation accuracy. This embodiment uses the speed, acceleration, and torque data of each axis motor, combined with the mathematical optimization method of least squares, to solve for the unknown parameter values in the LuGre friction model.

[0130] As a classic parameter estimation method, the least squares method can effectively process multiple sets of data and solve for parameters by minimizing the sum of squared errors. It has the advantages of computational stability and wide applicability. By substituting the collected motor data into the model and applying the least squares method, an optimal set of parameter values can be found that minimizes the error between the model output and the actual data. Furthermore, using motor data collected during actual operation to determine parameters avoids the limitations of subjective assumptions and theoretical derivations, making the model parameters in the LuGre friction model more realistic.

[0131] This embodiment can achieve accurate identification of the friction characteristic parameters of machine tools, which not only improves the simulation accuracy of the digital twin model for the actual operating status of the machine tool, but also enhances the model's adaptability to different working conditions, providing a reliable basis for subsequent model-based performance analysis, fault prediction and maintenance strategy optimization, and effectively improving the maintenance efficiency and production reliability of the machine tool.

[0132] In some embodiments, the step S21 mentioned above, the step of obtaining the parameters of the machine tool and assigning values to the digital twin model based on the parameters may specifically include:

[0133] Step S41: Send segmented constant speed and acceleration motion instructions to the machine tool, and collect speed, acceleration and torque data of each axis motor.

[0134] The specific implementation of this step is the same as that of step S31 and will not be repeated here.

[0135] Step S42: Obtain the motor parameters of the machine tool, and determine the servo loop parameters in the digital twin model based on the speed, acceleration, torque data and motor parameters of each axis motor.

[0136] In this embodiment, the machine tool's motor parameters include position proportional gain, velocity proportional gain, and velocity integral gain, which can be obtained by querying the machine tool's manufacturer documentation or motor nameplate. After obtaining the motor parameters, mathematical modeling and parameter identification methods are used to calculate key servo loop parameters in the digital twin model, such as the filter constants for the position, velocity, and current loops, combined with dynamic data such as the speed, acceleration, and torque of each axis motor. This allows the digital twin model to accurately simulate the machine tool's actual servo control performance.

[0137] Servo loop parameters directly impact the machining accuracy and dynamic performance of machine tools. Accurate parameters ensure that each axis of the machine tool behaves as expected during operation. By combining motor parameters with actual operating data to determine servo loop parameters, we can bridge the gap between theoretical models and actual machines, improving the accuracy and reliability of digital twin models.

[0138] In some embodiments, the collected operating data and motor parameters can be used to calculate unknown parameters in the servo loop model through a parameter identification algorithm (such as the least squares method, extended Kalman filtering, etc.).

[0139] In step S22, a simulated circular trajectory is generated using the assigned digital twin model, and the polar coordinate deviation between the simulated circular trajectory and the theoretical circular trajectory is determined as the quadrant cusp error value.

[0140] In this embodiment, after assigning parameters to the machine tool's digital twin model, the model is simulated to execute circular motion, generating a simulated circular trajectory. Simultaneously, the ideal circular trajectory of the machine tool, calculated theoretically, is compared with the simulated trajectory. The deviation between the two trajectories in polar coordinates is used as the quadrant angular error. This process quantifies the difference between the machine tool's actual machining accuracy and its ideal state, providing a basis for subsequent maintenance and optimization.

[0141] Based on the above technical solution, this embodiment constructs a digital twin model of the machine tool that integrates the LuGre friction model and the servo control loops of each axis. Quadrant cusp error values are acquired at a preset period and stored in a time series. The error value is then determined using the polar coordinate deviation between the simulated and theoretical circular trajectories. This model accurately and in real time reflects the actual operating status and performance changes of the machine tool. This provides reliable data support for subsequent maintenance time predictions, effectively improving prediction accuracy and enabling refined management and efficient maintenance of the machine tool.

[0142] Please refer to Figure 10 ,for Figure 1A flowchart of an actual embodiment of step S02 in a method for predicting machine tool maintenance time is provided. In some embodiments, step S02, as mentioned above, predicting the maintenance time of the machine tool based on the quadrant cusp error value time series, can specifically include the following steps:

[0143] Step S51: Generate a state indicator time series according to the quadrant cusp error value time series.

[0144] Among them, the state indicator is the difference between the preset threshold and the quadrant sharp angle error value;

[0145] In this embodiment, generating a status indicator time series from a quadrant cusp error time series involves comparing the quadrant cusp error data series generated during machine tool operation with a preset threshold and calculating the difference between the two, thereby forming a new status indicator time series. Specifically, the status indicator is defined as the difference between the preset threshold and the quadrant cusp error value. This difference can intuitively reflect the operating status and health of the machine tool at different moments. By analyzing the status indicator time series, changing trends in machine tool performance can be more clearly captured, providing more direct and effective data support for subsequent maintenance time prediction.

[0146] Because the quadrant cusp error value reflects changes in machine tool accuracy, which is closely related to the machine tool's health, the difference between the preset threshold and the quadrant cusp error value can be calculated to quantify the deviation between the machine tool's current state and its ideal state, thereby forming a state indicator time series. This series can more intuitively demonstrate the changing trends in machine tool performance, providing more direct and effective data support for subsequent maintenance time prediction, helping to achieve accurate prediction of machine tool maintenance time.

[0147] In some embodiments, as mentioned in step S51, the state indicator time series is generated according to the quadrant cusp error value time series, which may specifically be:

[0148] A preset threshold corresponding to the normal operation of the machine tool is set. This threshold can be determined based on the machine tool's design parameters, historical operating data, or industry standards. Then, for each data point in the quadrant cusp error time series, the difference between the data point and the preset threshold is calculated to obtain the status indicator. The formula for calculating the status indicator is: Status indicator = Preset threshold - Quadrant cusp error. Finally, each calculated status indicator is arranged in chronological order to form a status indicator time series.

[0149] This sequence can intuitively reflect the degree of deviation between the machine tool's operating status at different time points and the ideal state, providing basic data for subsequent maintenance time prediction.

[0150] Step S52: using the state indicator time series to train a decay model, and using the trained decay model to predict the time when the state indicator becomes zero.

[0151] In this embodiment, a decay model is a mathematical model used to describe the gradual degradation or deterioration of system, device, or component performance over time. Because the time series of status indicators can intuitively reflect the changing trends in machine tool performance, the decay model can effectively describe and predict these trends. By training the decay model, the temporal changes in the status indicator can be captured, allowing predictions to be made regarding when the status indicator will reach zero, indicating when the machine tool requires maintenance.

[0152] Therefore, this embodiment establishes a mathematical model describing the performance degradation patterns based on the status indicator data generated in step S51, and uses this model to predict the specific time when the machine tool requires maintenance. The status indicator time series reflects the changing trend of machine tool performance over time, and the decay model can capture the inherent patterns of this change. By training the decay model, the degradation trend of the status indicator over time can be accurately described, and then the time when the status indicator will become zero can be predicted, that is, the time when the machine tool requires maintenance. This process can achieve accurate prediction of machine tool maintenance time, avoid the blindness of traditional regular maintenance methods, reduce production losses caused by untimely maintenance or resource waste caused by excessive maintenance, thereby effectively reducing maintenance costs and improving production efficiency and equipment utilization.

[0153] In some embodiments, the decline model may include at least one of a similarity model, a survival model, and a degradation model.

[0154] In some embodiments, the process of using the state indicator time series to train the decay model mentioned in step S52 may specifically be:

[0155] First, the state indicator time series is preprocessed to remove noise and outliers and improve data quality. Then, features that can characterize the machine tool state, such as mean, variance, trend term, and periodic term, are extracted from the state indicator time series. Finally, a recession model is constructed based on these features, and the recession model is fitted to the state indicator time series to complete the training of the recession model.

[0156] Step S53: The time when the status indicator becomes zero is determined as the maintenance time of the machine tool.

[0157] For example, see Figure 11 , is a schematic diagram of a method for predicting the maintenance time of a machine tool using a decay model provided by an embodiment of the present invention. Figure 11As shown in Figure 2, the decay model trained using the state indicator time series can capture the changing trends in machine tool performance over time. This decay model can also predict the time when the state indicator reaches zero, which is the time when the machine tool requires maintenance. Furthermore, the remaining useful life of the machine tool is calculated by subtracting the time when the most recent state indicator data was collected from the maintenance time.

[0158] In this embodiment, the machine tool's operating status is continuously monitored and its status indicator is calculated. Since the status indicator reflects the deviation of the machine tool's performance from a preset threshold, when the status indicator reaches zero, it indicates that the machine tool's performance has degraded to a critical point, at which point maintenance should be scheduled. This process accurately determines maintenance timing based on actual machine tool performance changes, effectively avoiding over- or under-maintenance, ensuring reliable equipment operation, and reducing maintenance costs and downtime.

[0159] Based on the above technical solution, this embodiment of the present invention generates a time series of status indicators by monitoring quadrant cusp error values and trains a decay model to predict when the status indicators return to zero, which serves as the maintenance time. This allows accurate determination of maintenance timing based on actual machine tool performance changes. Compared to traditional scheduled maintenance, this embodiment avoids production losses caused by untimely maintenance and resource waste caused by excessive maintenance, achieving precise, efficient, and economical equipment maintenance, effectively extending machine tool service life and reducing maintenance costs.

[0160] Based on the above embodiment, in some embodiments, the method may further include the following steps:

[0161] Step S54: In the decay model, predict the confidence interval of the time when the state indicator becomes zero.

[0162] like Figure 11 As shown, this embodiment can predict the time confidence interval when the state indicator becomes zero through the decay model, wherein the time confidence interval is used to estimate the range of the true value of the time when the state indicator becomes zero.

[0163] In this embodiment, two important parameter estimation methods in statistics (i.e., maximum likelihood estimation method or Bayesian method) may be used to determine the predicted time range (i.e., time confidence interval) for the state indicator to change over time until reaching the maintenance threshold, where:

[0164] The maximum likelihood estimation method constructs a prediction model by finding the model parameter values that maximize the probability of the observed data;

[0165] The Bayesian method combines prior knowledge and observation data to update the probability distribution of parameters, thereby obtaining a more comprehensive description of uncertainty.

[0166] Both methods provide a scientific basis for predicting time confidence intervals. Maximum likelihood estimation is suitable for parameter estimation of large numbers of independent and identically distributed samples, while the Bayesian method can effectively integrate prior information and data, making it particularly advantageous when the sample size is limited or when parameter uncertainty needs to be considered. This embodiment uses maximum likelihood estimation or the Bayesian method to estimate the model parameters of the decay model and, combined with the uncertainty quantification results of the decay model, determines the time confidence interval, providing a reliable basis for machine tool maintenance time decisions.

[0167] Step S55 , when the quadrant cusp error value time series is updated, the state indicator time series is updated, and the parameter estimation deviation of the decay model is corrected according to the updated state indicator time series to compress the time confidence interval.

[0168] In this embodiment, as the machine tool continues to operate, its quadrant cusp error values continuously generate new data points, forming a dynamic update of the time series. Accordingly, the status indicator time series also needs to be updated synchronously to reflect the latest changes in machine tool performance. Based on the updated status indicator time series, the parameters of the decay model are re-estimated and corrected to reduce the deviation of the model parameters, thereby making the predicted time confidence interval more accurate and reliable.

[0169] As a machine tool operates, its performance gradually changes. New quadrant cusp error data provides the latest performance information. Without timely updates to the status indicator time series and corrections to the decay model parameters, the model's predictions may lag behind the actual machine tool performance, leading to inaccurate maintenance time predictions. Dynamic updates and corrections can better capture changing machine tool performance trends, improve the adaptability and accuracy of the decay model, and ultimately narrow the temporal confidence interval, enhancing the credibility and practicality of the predictions.

[0170] Based on the above technical solution, the present invention employs maximum likelihood estimation or Bayesian methods in the decay model to predict the confidence interval of the time when the state indicator reaches zero. This method then corrects the state indicator time series in real time as the quadrant cusp error time series is updated, optimizing the decay model's parameter estimates and effectively compressing the time confidence interval. This not only improves the scientific nature and reliability of maintenance time predictions but also enables more accurate determination of machine tool maintenance opportunities, reducing production losses and maintenance costs caused by prediction errors.

[0171] Based on the above embodiment, in some embodiments, before executing step S12 to trigger the process of obtaining the quadrant cusp error value according to a preset period, the method may further include:

[0172] Step S61: Obtain the usage frequency of the machine tool.

[0173] In this embodiment, acquiring machine tool usage frequency refers to collecting data related to the number of times a machine tool is used, the duration of use, or the intensity of use within a certain period of time. This data can reflect the actual workload and operating status of the machine tool, providing basic information for subsequent maintenance planning, performance evaluation, and production scheduling.

[0174] The frequency of machine tool usage directly reflects its workload. Frequently used machines tend to experience greater wear and fatigue, requiring more frequent maintenance and inspection. Frequency analysis can also help rationally schedule production tasks, optimize production processes, and improve equipment utilization and productivity. Frequency data can also provide a reference for machine tool performance evaluation and lifespan prediction, helping users promptly identify potential issues and develop appropriate equipment upgrade plans.

[0175] Step S62: adjusting the preset period according to the frequency of use.

[0176] In this embodiment, the originally set fixed period is dynamically modified based on relevant data such as the number of times the machine tool is actually used, the duration of use, or the intensity of use. This adjustment allows the process of obtaining quadrant cusp error values to more flexibly adapt to the actual operating conditions of the machine tool, ensuring timely and accurate acquisition of machine tool performance data under different usage conditions.

[0177] The frequency of machine tool usage directly affects the rate of wear and performance change. Frequently used machines may experience faster performance degradation, necessitating more frequent acquisition of quadrant cusp error values to capture performance changes promptly. In contrast, infrequently used machines maintain relatively stable performance, allowing for longer acquisition cycles to reduce unnecessary data collection and processing. Furthermore, by adjusting the preset cycle based on usage frequency, maintenance resource allocation can be optimized, avoiding over-maintenance during infrequent machine use and under-maintenance during frequent use. This improves maintenance efficiency, reduces maintenance costs, and extends the life of the machine.

[0178] Based on the above technical solution, an embodiment of the present invention adjusts the preset cycle of the quadrant cusp error value acquisition process based on the frequency of machine tool use. High frequency of use shortens the cycle, while low frequency increases it. This adjustment ensures timely acquisition of error values when the machine tool is frequently used, preventing the impact of error accumulation on machining accuracy. Furthermore, when the machine tool is used less frequently, the cycle is extended, minimizing unnecessary data collection and processing and reducing computing resource consumption. Furthermore, by dynamically adjusting the preset cycle, changes in machine tool performance can be more accurately captured, improving the accuracy of maintenance time predictions, enabling on-demand maintenance, extending the machine tool's service life, and reducing maintenance costs.

[0179] Please refer to Figure 12 , is a structural diagram of a machine tool maintenance time prediction system provided by an embodiment of the present invention. Figure 12As shown, the machine tool maintenance time prediction system may include:

[0180] An acquisition module 100 is configured to acquire a time series of quadrant cusp error values of a machine tool. The quadrant cusp error values are determined by generating a simulated circular trajectory using a digital twin model of the machine tool and based on the polar coordinate deviation between the simulated circular trajectory and a theoretical circular trajectory.

[0181] The prediction module 200 is used to predict the maintenance time of the machine tool according to the quadrant cusp error value time series.

[0182] Based on the above embodiments, in some embodiments, the prediction module 200 may be specifically used for:

[0183] Generate a state indicator time series based on the quadrant cusp error value time series; wherein the state indicator is the difference between the preset threshold and the quadrant cusp error value;

[0184] Use the state indicator time series to train a decay model, and use the trained decay model to predict the time when the state indicator becomes zero;

[0185] The time when the status indicator becomes zero is determined as the maintenance time of the machine tool.

[0186] Based on the above embodiment, in some embodiments, the prediction module 200 may also be used to:

[0187] In the recession model, the confidence interval of the time when the state indicator becomes zero is predicted by maximum likelihood estimation or Bayesian method;

[0188] When the quadrant cusp error value time series is updated, the state indicator time series is updated, and the parameter estimation deviation of the recession model is corrected according to the updated state indicator time series to compress the time confidence interval.

[0189] Based on the above embodiments, in some embodiments, the acquisition module 100 may be specifically used to:

[0190] Build a digital twin model of the machine tool; the digital twin model integrates the LuGre friction model and the servo control loops of each axis of the machine tool;

[0191] Triggering the acquisition process of the quadrant cusp error value according to a preset period, and storing the obtained quadrant cusp error value and its corresponding acquisition time in the quadrant cusp error value time series;

[0192] The process of obtaining the quadrant sharp angle error value includes:

[0193] Obtain the parameters of the machine tool and assign values to the digital twin model based on the parameters;

[0194] The assigned digital twin model is used to generate a simulated circular trajectory, and the polar coordinate deviation between the simulated circular trajectory and the theoretical circular trajectory is determined as the quadrant cusp error value.

[0195] Based on the above embodiments, in some embodiments, the acquisition module 100 may be specifically used to:

[0196] Send segmented constant speed and acceleration motion instructions to the machine tool, and collect speed, acceleration and torque data of each axis motor;

[0197] Based on the speed, acceleration and torque data of each axis motor, the parameters of the LuGre friction model in the digital twin model are determined by the least squares method.

[0198] Based on the above embodiments, in some embodiments, the acquisition module 100 may be specifically used to:

[0199] Send segmented constant speed and acceleration motion instructions to the machine tool, and collect speed, acceleration and torque data of each axis motor;

[0200] Obtain the motor parameters of the machine tool and determine the servo loop parameters in the digital twin model based on the speed, acceleration, torque data and motor parameters of each axis motor.

[0201] Based on the above embodiment, in some embodiments, the acquisition module 100 may also be used to:

[0202] Get the usage frequency of the machine tool;

[0203] Adjust the preset cycle according to the frequency of use.

[0204] This embodiment provides an electronic device, including a processor and a memory, the memory being used to store at least one instruction, which is loaded and executed by the processor to implement the above-mentioned method for predicting machine tool maintenance time. Its execution method and beneficial effects are similar and will not be repeated here.

[0205] An embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the method for predicting machine tool maintenance time described above is implemented. The execution method and beneficial effects thereof are similar and will not be described in detail here.

[0206] It should be noted that although the above describes the various steps in a specific order, it does not mean that the steps must be performed in the above specific order. In fact, some of these steps can be executed concurrently or even in a different order as long as the required functions can be achieved.

[0207] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for predicting machine tool maintenance time, characterized in that: include: Obtaining a time series of quadrant cusp error values of a machine tool; wherein the quadrant cusp error value is determined by generating a simulated circular trajectory using a digital twin model of the machine tool and based on a polar coordinate deviation between the simulated circular trajectory and a theoretical circular trajectory; The maintenance time of the machine tool is predicted according to the quadrant cusp error value time series.

2. The method according to claim 1, characterized in that The predicting of the maintenance time of the machine tool according to the quadrant cusp error value time series includes: Generate a state indicator time series according to the quadrant cusp error value time series; wherein the state indicator is the difference between a preset threshold and the quadrant cusp error value; Using the state indicator time series to train a decay model, and using the trained decay model to predict the time when the state indicator becomes zero; The time when the status indicator becomes zero is determined as the maintenance time of the machine tool.

3. The method according to claim 2, characterized in that The method further comprises: In the decay model, predicting a confidence interval for the time at which the state indicator becomes zero; When the quadrant cusp error value time series is updated, the state indicator time series is updated, and the parameter estimation deviation of the decay model is corrected according to the updated state indicator time series to compress the time confidence interval.

4. The method according to claim 1, wherein The method of obtaining a time series of quadrant angular error values of a machine tool includes: Constructing a digital twin model of the machine tool; the digital twin model integrates the LuGre friction model and the servo control loops of each axis of the machine tool; The quadrant cusp error value acquisition process is triggered according to a preset period, and the obtained quadrant cusp error value and its corresponding acquisition time are stored in the quadrant cusp error value time series.

5. The method according to claim 4, characterized in that The process of obtaining the quadrant sharp angle error value includes: Obtaining parameters of the machine tool and assigning values to the digital twin model based on the parameters; The assigned digital twin model is used to generate a simulated circular trajectory, and the polar coordinate deviation between the simulated circular trajectory and the theoretical circular trajectory is determined as the quadrant cusp error value.

6. The method according to claim 4, characterized in that The obtaining of the parameters of the machine tool and assigning values to the digital twin model based on the parameters includes: Sending segmented constant speed and acceleration motion instructions to the machine tool, and collecting speed, acceleration and torque data of each axis motor; Based on the speed, acceleration and torque data of the motors of each axis, the parameters of the LuGre friction model in the digital twin model are determined by the least squares method.

7. The method according to claim 4, characterized in that The obtaining of the parameters of the machine tool and assigning values to the digital twin model based on the parameters includes: Sending segmented constant speed and acceleration motion instructions to the machine tool, and collecting speed, acceleration and torque data of each axis motor; The motor parameters of the machine tool are obtained, and the servo loop parameters in the digital twin model are determined based on the speed, acceleration, torque data of the motors of each axis and the motor parameters.

8. A system for predicting machine tool maintenance time, characterized in that: include: an acquisition module for acquiring a time series of quadrant cusp error values of a machine tool; wherein the quadrant cusp error values are determined by generating a simulated circular trajectory using a digital twin model of the machine tool and based on a polar coordinate deviation between the simulated circular trajectory and a theoretical circular trajectory; A prediction module is used to predict the maintenance time of the machine tool according to the quadrant cusp error value time series.

9. An electronic device, characterized in that: include: A processor and a memory, wherein the memory is used to store at least one instruction, and when the instruction is loaded and executed by the processor, the method for predicting machine tool maintenance time according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for predicting machine tool maintenance time according to any one of claims 1 to 7 is implemented.

Images