Self-adaptive flutter control robot impeller milling device and flutter detection method

Abstract

The invention discloses a self-adaptive flutter control robot impeller milling device, and belongs to the technical field of robot milling. The device is composed of a milling mechanical arm, an end effector assembly, an auxiliary mechanism assembly and a clamping disc device. When the self-adaptive flutter control robot carries out impeller milling, the robot end effector assembly is provided with a voice coil motor, a hydraulic damper and an eddy current damper, vibration and deformation of a milling cutter in the machining process can be effectively restrained, meanwhile, the supporting head is further provided with an auxiliary mechanism assembly, and it is ensured that the auxiliary supporting effect is achieved in the workpiece milling process. In the milling process, vibration signals are obtained through a sensor in the clamping disc device, variational mode decomposition, flutter sensitive feature extraction and milling flutter detection of the support vector machine are carried out on the signals, and high classification performance and high detection speed are achieved.

Description

Technical Field

[0001] The invention relates to the technical field of robot machining and milling, and in particular to an adaptive flutter control robot impeller milling machining device, and an adaptive flutter detection method based on milling operation. Background Art

[0002] As a core component of fluid machinery, impellers are used in aircraft engines, automotive turbochargers, fuel cell high-speed air compressors and other equipment to convert fluid energy into mechanical energy through rotation. They are key components for realizing the mutual conversion between fluid kinetic energy and mechanical energy. At present, the main milling method for such parts in China is five-axis linkage CNC milling equipment, which can reduce tool costs and clamping times, and is suitable for mass production of impellers with simple shapes. However, the ability to process complex shapes is limited, and large deformation and vibration will occur during the processing, which will reduce the surface processing quality of the workpiece and make it difficult to ensure the accuracy of the remaining wall thickness. Summary of the invention

[0003] The purpose of the present invention is to overcome the defects of the prior art and provide an adaptive vibration control robot impeller milling processing device mainly from two aspects of flexible clamping of the impeller and reducing vibration during the robot milling process.

[0004] To solve the above problems, the technical solution adopted by the present invention is an adaptive flutter control robot impeller milling processing device, which is composed of a milling robot arm, an end effector assembly, an auxiliary mechanism assembly, and a clamping disk device. The milling robot arm is fixedly installed on the ground by anchor bolts, the end effector assembly is installed at the end of the milling robot arm by bolts, the auxiliary mechanism assembly is installed on the side end of the end effector assembly by bolts, and the clamping disk device is fixedly installed on the ground by anchor bolts.

[0005] Preferably, the end effector assembly is composed of a support head, a support end cover, a fixed plate, a second fixed plate, a rib plate, a voice coil motor, a hydraulic damper, an eddy current damper, a milling drill bit mounting seat, and a milling cutter. The support head is fixedly mounted on the end of the milling robot arm by bolts, the fixed plate is fixedly mounted on the left and right side ends of the support head by bolts, the rib plate is mounted on the left and right side ends of the support head, and is mounted on the same side as the fixed plate and is located at the lower part of the fixed plate, the second fixed plate is fixedly mounted on the side end of the support head by bolts together with the rib plate, the voice coil motor is mounted between the fixed plate and the second fixed plate by bolts, and is fixedly supported by the fixed plate and the second fixed plate, the support end cover is mounted on the lower end of the second fixed plate and is coaxial with the support head, the hydraulic damper is mounted on the lower end of the support head through the hole of the support end cover, the milling drill bit mounting seat is mounted on the lower end of the support head through the hole of the support end cover, the milling cutter is mounted on the milling drill bit mounting seat, and the eddy current damper is fixedly mounted on the front end of the support head by bolts.

[0006] Preferably, the auxiliary mechanism assembly is composed of an auxiliary support plate, an auxiliary main board, an auxiliary movable plate, and an auxiliary support head. The auxiliary support plate is installed at the rear end of the support head, the auxiliary main board is fixedly installed at the rear end of the support head through bolts and the auxiliary support plate, the auxiliary movable plate is installed on the outer wall of the auxiliary main board through bolts and nuts, the auxiliary movable plate has an oblong hole, and the auxiliary support head is installed on the bolts of the auxiliary movable plate through threaded cooperation.

[0007] Preferably, the clamping disc device is composed of a base, a rotating support, a driving motor, a rotating disc, a ball screw mechanism, a sensor mounting plate, and a sensor. The base is fixedly mounted on the ground by anchor bolts, the rotating support and the stepping motor are mounted on the base, the rotating disc is mounted on the rotating support by key matching, the ball screw mechanism is fixedly mounted in a fixing groove of the rotating disc, the sensor mounting plate is fixedly mounted on the rotating disc by bolts, and the sensor is clamped and mounted between two sensor mounting plates.

[0008] On the other hand, the present invention also provides an adaptive chatter detection method based on milling operation, characterized in that it comprises the following steps: S1. Obtaining a vibration signal from a sensor in the clamping disk device; S2. Perform variational mode decomposition on the signal and extract flutter sensitive features; S3. Milling chatter detection using support vector machine.

[0009] This method uses an optimized version of variational mode decomposition (VMD) to make it fully adaptable to different machining scenarios. The vibration signal is decomposed by VMD, and the modes with rich chatter information are selected for further analysis. Features in the time and frequency domains are extracted from these modes and used to train a support vector machine classifier to predict the stable state of the process.

[0010] The benefit of this invention is that it uses an adaptive flutter control robot to perform impeller milling and detect the vibration of the workpiece during the milling process. The end effector assembly of this invention is provided with a voice coil motor, which is a direct-drive motor based on the Lorentz force principle. It uses a permanent magnetic field and a coil winding to generate a force proportional to the current applied to the coil. It can achieve rapid push-pull conversion by changing the direction of the current, thereby achieving the suppression of vibration during the milling process. The end effector of this invention is provided with a hydraulic damper to protect the milling cutter during the milling process and prevent the equipment from being damaged by impact loads. At the same time, the hydraulic damper can freely adjust the speed of the piston to adapt to different working conditions. At the same time, the end effector is also provided with an eddy current damper, which can generate eddy currents in the end effector, thereby generating resistance, consuming the vibration energy of the end effector, and achieving a shock absorbing effect. Due to the characteristics of its structure and working principle, the maintenance cost of the eddy current damper is relatively low. During the milling process, the vibration signal is obtained through the sensor in the clamping disk device, and the signal is subjected to variational mode decomposition, chatter-sensitive feature extraction and milling chatter detection by support vector machine, which has high classification performance and fast detection speed. According to the SVM classification results, the PID controller is used to adjust the machining parameters of the robot arm and the damping force of the eddy current damper to perform real-time feedback regulation of the milling process.

[0011] The auxiliary mechanism assembly of the invention is installed on the support head. The auxiliary movable plate and the auxiliary main plate are provided with oblong holes. The auxiliary support head is installed by bolts, so that the auxiliary support head can be moved within a certain range, thereby adjusting the position relationship with the workpiece, ensuring that it plays an auxiliary support role in the milling process of the workpiece, thereby reducing vibration.

[0012] The clamping disk device of this invention adopts a ball screw to clamp the workpiece. The installation of four ball screw mechanisms can realize effective clamping of workpieces of different shapes to ensure processing accuracy. At the same time, a sensor is installed on the rotating disk to detect the vibration of the workpiece in real time during the milling process.

[0013] In a word, the device can effectively suppress the vibration in the milling process, complete the rapid milling of the impeller workpiece, improve the impeller milling accuracy, and can be widely used in impeller milling work. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 It is a structural schematic diagram of the present invention; Figure 2 for Figure 1 Schematic diagram of the end effector assembly in FIG. Figure 3 for Figure 1 Schematic diagram of auxiliary mechanism components; Figure 4 for Figure 1 Schematic diagram of the clamping disc device; Figure 5 for Figure 4 Schematic diagram of the ball screw mechanism in FIG. Figure 6 The flowchart of the adaptive flutter detection method of the device is as follows; In the figure, 1, milling robot arm, 2, end effector assembly, 3, auxiliary mechanism assembly, 4, clamping disc device; 201, support head, 202, eddy current damper, 203, support end cover, 204, milling drill bit mounting seat, 205, milling cutter, 206, hydraulic damper, 207, second fixing plate, 208, rib plate, 209, voice coil motor, 210, fixing plate; 301, auxiliary main board, 302, auxiliary movable board, 303, auxiliary support board, 304, auxiliary support head; 401, rotating disc, 402, ball screw mechanism, 403, rotating support, 404, base, 405, sensor, 406, sensor mounting plate; 4021, ball screw, 4022, clamping plate, 4033, stepping motor; DETAILED DESCRIPTION

[0015] Embodiment 1

[0016] An adaptive flutter control robot impeller milling processing device is composed of a milling mechanical arm 1, an end effector assembly 2, an auxiliary mechanism assembly 3, and a clamping disc device 4. The milling mechanical arm 1 is fixedly installed on the ground by anchor bolts, the end effector assembly 2 is installed at the end of the milling mechanical arm 1 by bolts, the auxiliary mechanism assembly 3 is installed on the side end of the end effector assembly 2 by bolts, and the clamping disc device 4 is fixedly installed on the ground by anchor bolts.

[0017] The end effector assembly 2 is composed of a support head 201, a support end cover 203, a fixing plate 209, a second fixing plate 207, a voice coil motor 208, a hydraulic damper 206, an eddy current damper 202, a milling drill bit mounting seat 204, a milling cutter 205, and a rib plate 208. The support head 201 is fixedly mounted on the end of the milling robot arm 1 by bolts, the fixing plate 209 is fixedly mounted on the left and right sides of the support head 201 by bolts, the rib plate 208 is mounted on the left and right sides of the support head 201, and is mounted on the same side as the fixing plate 210 and is located at the bottom of the fixing plate 210, the second fixing plate 207 is fixedly mounted on the side end of the support head 201 by bolts together with the rib plate 208, the voice coil motor 209 is fixedly mounted between the fixing plate 210 and the second fixing plate 207 by bolts, and is connected to the fixing plate 210 by bolts. The support is fixedly supported by the fixing plate 210 and the second fixing plate 207, the support end cover 203 is installed at the lower end of the second fixing plate 207 and is coaxial with the support head 201, the hydraulic damper 206 is installed at the lower end of the support head 201 through the hole of the support end cover 203, the milling drill bit mounting seat 204 is installed at the lower end of the support head 201 through the hole of the support end cover 203, the milling cutter 205 is installed on the milling drill bit mounting seat 204, and the eddy current damper 202 is fixedly installed at the front end of the support head 201 by bolts.

[0018] The auxiliary mechanism assembly 3 is composed of an auxiliary support plate 303, an auxiliary main plate 301, an auxiliary movable plate 302, and an auxiliary support head 304. The auxiliary support plate 303 is installed at the rear end of the support head 201, the auxiliary main plate 301 is fixedly installed at the rear end of the support head 201 through bolts and the auxiliary support plate 303, the auxiliary movable plate 302 is installed on the outer wall of the auxiliary main plate 301 through bolts and nuts, the auxiliary movable plate 302 has an oblong hole, and the auxiliary support head 304 is installed on the bolts of the auxiliary movable plate 302 through threaded cooperation.

[0019] The clamping disk device 4 is composed of a base 404, a rotating support 403, a driving motor 407, a rotating disk 401, a ball screw mechanism 402, a sensor mounting plate 406, and a sensor 405. The base 404 is fixedly mounted on the ground by anchor bolts, the rotating support 403 and the stepping motor 407 are mounted on the base 404, the rotating disk 401 is mounted on the rotating support 403 by key matching, the ball screw mechanism 402 is fixedly mounted in the fixing groove of the rotating disk 401, the sensor mounting plate 406 is fixedly mounted on the rotating disk 401 by bolts, and the sensor 405 is clamped and mounted between the two sensor mounting plates 406.

[0020] When the device is in use, the milling impeller is placed on the rotating disc 401, and the ball screw mechanism 402 clamps the impeller. After the impeller is clamped, the milling robot arm 1 performs milling. During the milling process, the auxiliary support head 304 in the auxiliary mechanism 3 can adjust the distance to provide auxiliary support for the impeller milling. The voice coil motor 208 generates a force proportional to the current applied to the coil by utilizing the permanent magnetic field and the coil winding. It can achieve rapid push-pull conversion by changing the direction of the current, thereby achieving vibration suppression during the milling process. The eddy current damper 202 can generate eddy currents in the end effector 2, thereby generating resistance, consuming the vibration energy of the end effector 2, and achieving a shock absorbing effect. At the same time, the hydraulic damper 206 plays the role of buffering and shock absorbing and protecting the milling cutter 205. Finally, the sensor 405 on the clamping disc device 4 obtains the vibration signal. The device is in the milling process.

[0021] Embodiment 2

[0022] Based on the processing device provided in the first embodiment, this embodiment further proposes an adaptive vibration detection method based on milling operation, which performs vibration detection on the milling workpiece in real time: Firstly, the vibration signal is subjected to variational mode decomposition, chatter-sensitive feature extraction and milling chatter detection using support vector machine. The specific expressions are as follows: Variational mode decomposition (VMD) aims to decompose a real-valued input signal into a discrete number of sub-signals (modes), decomposing the original signal s(t) into a set of K modes.

[0023] ; in Represents the signal to be decomposed, usually a time series, such as the vibration signal in the milling process; represents the kth intrinsic mode function (IMF), which is obtained from the signal The components decomposed in ; K represents the number of mode functions decomposed.

[0024] In order to control the width and frequency concentration of the mode, VMD introduces a penalty term, and its optimization objective can be expressed as.

[0025] ; ; in represents the center frequency of the kth mode. It is used for modulation The frequency can reflect the main frequency components of the mode; represents Dirac distribution; It represents the partial derivative with respect to time t, reflecting the rate of change of the signal over time. s represents the signal to be decomposed; t represents time.

[0026] In order to solve the constraint of accurate reconstruction, VMD introduces a quadratic penalty term.

[0027] ; in represents convolution; represents a secondary penalty; represents the Lagrange multiplier.

[0028] It assumes that the solution to this optimization problem can be achieved through a series of iterative sub-optimizations called the alternating direction method of multipliers (ADMM), and the formula for each mode is as follows.

[0029] ; Similarly, the optimization for each center frequency can be solved as follows.

[0030] ; in represents the center frequency of the signal to be decomposed; The superscripts denote the Fourier transform of the signal, Represents the penalty factor, which is used to balance the bandwidth of the mode and the accuracy of signal reconstruction.

[0031] Extracting features from the signal that retains chatter-related information can improve the prediction performance of the algorithm and reduce the computation time. A new feature, the energy ratio, is developed. The energy ratio is the ratio of the energy of the chatter-related mode selected according to the kurtosis criterion to the energy of the entire signal, and the formula is as follows.

[0032] Continuous signal The calculation formula of energy E is as follows.

[0033] ; And for a discrete signal with n samples (sampled in the time domain): ; Then, the energy ratio of the kth mode will be:

[0034] Where E is the energy of the entire signal. These features are calculated from the decomposed vibration signals of the x-axis and y-axis. A total of 6 features will be used as input to the classification algorithm. Vibration detection is performed using a support vector machine (SVM) based on the features calculated from the decomposed signal. The implementation of the SVM algorithm is executed in Python. According to the SVM classification results, a PID controller is used to perform real-time feedback adjustment on the milling process. When the support vector machine (SVM) classifier detects that the vibration is classified as a stable state, the PID controller maintains the current machining parameters without adjustment. When the SVM detects mild chatter, the PID controller calculates the error between the current vibration amplitude and the target vibration amplitude, and outputs an adjustment signal to appropriately reduce the feed rate, cutting depth, or adjust the tool path to reduce the impact of chatter. When the SVM detects severe chatter, the PID controller adjusts the feed rate, cutting depth, or even suspends the machining operation by a large margin. At the same time, it quickly suppresses the vibration by adjusting the damping force of the eddy current damper to avoid further expansion of the chatter. The conceptual flow of the adaptive chatter detection method is as follows Figure 6 shown.

[0035] For those skilled in the art, according to the teachings of the present invention, without departing from the principles and spirit of the present invention, changes, modifications, substitutions and variations made to the implementation methods are still within the protection scope of the present invention.

Claims

1. An adaptive flutter control robot impeller milling processing device, characterized in that: It includes a milling robot arm, an end effector assembly, an auxiliary mechanism assembly, and a clamping disc device; The milling robot arm is fixedly mounted on the ground by anchor bolts, the end effector assembly is mounted at the end of the milling robot arm, the auxiliary mechanism assembly is mounted on the side end of the end effector assembly, and the clamping disk device is fixedly mounted on the ground by anchor bolts.

2. The end effector assembly comprises a support head, a support end cover, a fixed plate, a second fixed plate, a rib plate, a voice coil motor, a hydraulic damper, an eddy current damper, a milling drill bit mounting seat, and a milling cutter. The support head is fixedly mounted at the end of the milling robot arm by bolts, the fixed plate is fixedly mounted at the left and right side ends of the support head by bolts, the rib plate is mounted at the left and right side ends of the support head, is mounted on the same side as the fixed plate and is located at the lower part of the fixed plate, the second fixed plate is fixedly mounted on the side end of the support head by bolts together with the rib plate, the voice coil motor is mounted between the fixed plate and the second fixed plate by bolts, and is fixedly supported by the fixed plate and the second fixed plate, the support end cover is mounted at the lower end of the second fixed plate and is coaxial with the support head, the hydraulic damper is mounted at the lower end of the support head through the hole of the support end cover, the milling drill bit mounting seat is mounted at the lower end of the support head through the hole of the support end cover, the milling cutter is mounted on the milling drill bit mounting seat, and the eddy current damper is fixedly mounted at the front end of the support head by bolts.

3. The adaptive flutter control robot impeller milling processing device according to claim 1, characterized in that: The auxiliary mechanism assembly includes an auxiliary support plate, an auxiliary main board, an auxiliary movable plate, and an auxiliary support head. The auxiliary support plate is installed at the rear end of the support head. The auxiliary main board is fixed to the rear end of the support head through bolts and the auxiliary support plate. The auxiliary movable plate is installed on the outer wall of the auxiliary main board through bolts and nuts. The auxiliary movable plate has an oblong hole, and the auxiliary support head is installed on the bolts of the auxiliary movable plate through threaded cooperation.

4. The adaptive flutter control robot impeller milling processing device according to claim 1, characterized in that: The clamping disk device includes a base, a rotating support, a driving motor, a rotating disk, a ball screw device, a sensor mounting plate, and a sensor. The base is fixedly mounted on the ground by anchor bolts, the rotating support and the stepper motor are mounted on the base, the rotating disk is mounted on the rotating support by key cooperation, the ball screw device is fixedly mounted in a fixing groove of the rotating disk, the sensor mounting plate is fixedly mounted on the rotating disk by bolts, and the sensor is clamped and mounted between two sensor mounting plates.

5. The adaptive flutter control robot impeller milling processing device according to claim 1, characterized in that: A ball screw mechanism is arranged on the surface of the clamping disc device, which can clamp workpieces of different shapes.

6. The adaptive flutter control robot impeller milling processing device according to claim 1, characterized in that: The surface of the clamping disc device is provided with a sensor, which can detect the vibration of the workpiece during the milling process.

7. An adaptive chatter detection method based on milling operation, characterized in that: The following steps are involved: S1. Obtaining a vibration signal from a sensor in the clamping disk device; S2. Perform variational mode decomposition on the signal and extract flutter sensitive features; S3. Milling chatter detection using support vector machine.

8. The adaptive chatter detection method based on milling operation according to claim 6, characterized in that: In step S2, the original vibration signal s(t) is first decomposed into a set of K modes: ； In order to control the width and frequency concentration of the mode, VMD introduces a penalty term, and its optimization objective can be expressed as: ； ； In order to solve the constraint of accurate reconstruction, VMD introduces a quadratic penalty term: ； The solution to the optimization problem can be achieved through a series of iterative sub-optimizations called the alternating direction method of multipliers (ADMM), where the formula for each mode is as follows: ； The optimization for each center frequency can be solved as: ； Extracting features from the signal that retains chatter-related information can improve the prediction performance of the algorithm and reduce the computation time. A new feature, the energy ratio, is developed. The energy ratio is the ratio of the energy of the chatter-related mode selected according to the kurtosis criterion to the energy of the entire signal. The formula is as follows: Continuous signal The energy E is calculated as follows: ； For a discrete signal with n samples (sampled in the time domain): ； The energy ratio of the kth mode will be: ； in E is the energy of the entire signal, and there are a total of 6 features that will be used as input to the classification algorithm.

9. The adaptive chatter detection method based on milling operation according to claim 7, characterized in that: The chatter detection is performed using support vector machine (SVM) based on the features calculated from the decomposed signal. The implementation of the SVM algorithm is executed in Python. According to the SVM classification results, the PID controller is used to adjust the machining parameters of the robot arm and the damping force of the eddy current damper to perform real-time feedback regulation of the milling process.

Images