Intelligent spindle machining process monitoring system and method based on edge calculation

Abstract

The invention discloses an intelligent spindle machining process monitoring system and method based on edge calculation. The system comprises a vibration exciter, a force sensor, a connecting rod, a bearing and an acceleration sensor. The acceleration sensor is arranged on a box body of the intelligent main shaft, the acceleration sensor is connected with the input end of the embedded controller, the acceleration sensor is used for collecting vibration acceleration signals in the machining process of the intelligent main shaft, and the embedded controller is used for executing a frequency response function calculation and cutting force online identification method; the bearing is arranged at the output end of the intelligent main shaft, one end of the connecting rod is connected with the outer ring of the bearing, and the other end is connected with the vibration exciter; the embedded controller is connected with the wireless transmission module through an I / O port; according to the intelligent spindle, the structure of the spindle does not need to be changed, the spindle can be installed on the spindle box body in a pasting / adsorption mode and the like, and an existing machine tool / industrial robot spindle can be upgraded into the intelligent spindle with the machining process monitoring function.

Description

Technical Field

[0001] The present invention belongs to the field of intelligent manufacturing, and specifically relates to an intelligent spindle machining process monitoring system based on edge computing. Background Art

[0002] The intelligent spindle is a core functional component of intelligent machine tools and industrial robots. Its performance directly determines the overall performance of intelligent CNC machine tools and industrial robots, and it represents the development direction for the next generation of spindles supporting these machines. The development of intelligent spindles by integrating multiple sensors into traditional spindles has become an industry trend and a key future research direction. The key difference between intelligent spindles and traditional spindles lies in their sensing, decision-making, and execution capabilities. These capabilities enable intelligent spindles to achieve higher machining accuracy and efficiency by monitoring and controlling multiple signals during operation, such as vibration, current, heat, and force. This also places higher demands on dynamic performance analysis of the spindle during design, as well as on state monitoring and sensing decisions during operation.

[0003] For intelligent machine tools and industrial robotic machining systems, cutting force is the system input, frequency response function (FRF) is the inherent characteristic of the machining system, and vibration is the output. Therefore, cutting force and FRF are crucial for monitoring the machining process. Currently, the most commonly used method for measuring cutting force is using a piezoelectric tabletop dynamometer produced by Kistler. The workpiece and dynamometer are connected via bolts at the base of the workpiece, requiring the workpiece base to be machined into a specific shape in order to directly measure milling forces during machining. This measurement method limits the size and shape of the workpiece. To overcome this problem, Kistler has developed and manufactured a rotary dynamometer, which requires no contact with the workpiece and has no specific requirements for workpiece shape. However, this device has a complex internal structure and is expensive. Furthermore, it must be installed between the tool and the spindle, which can affect the spindle's dynamic performance and machining stability. FRFs are often measured using modal hammer testing to obtain the static FRF of the tool tip. However, the FRF of the tool tip varies at different rotational speeds during actual machining. This invention proposes an intelligent spindle machining process monitoring system and method based on edge computing, achieving intelligent spindle intelligence for machine tools / industrial robots. Specifically, it incorporates functions such as acceleration acquisition, frequency response function acquisition, online cutting force identification, data visualization, data storage, and wireless data transmission. Solutions already exist for spindle monitoring system installation (Spindle Operating Parameter Monitoring Device, Monitoring System, and Processing Machine Tool CN 218504032 U) and self-powering (System and Method for Monitoring Spindle Operating Parameters, System, and Processing Machine Tool CN 115990789 A). However, the core of an intelligent spindle lies in its acceleration acquisition, frequency response function acquisition, online cutting force identification, data visualization, data storage, and wireless data transmission. Summary of the Invention

[0004] In order to solve the technical problems raised in the above background technology, the present invention aims to design an intelligent spindle processing process monitoring system based on edge computing, providing a solution for realizing the intelligentization of the spindle of machine tools / industrial robots.

[0005] To achieve the above technical objectives, in the first aspect, the present invention provides an intelligent spindle processing process monitoring system based on edge computing, including an exciter, a force sensor, a connecting rod, a bearing, and an acceleration sensor; the acceleration sensor is arranged on the intelligent spindle housing, and the acceleration sensor is connected to the input end of the embedded controller, wherein the acceleration sensor is used to collect the vibration acceleration signal during the intelligent spindle processing process, and the embedded controller is used to execute the frequency response function calculation and the cutting force online identification method; the bearing is arranged at the output end of the intelligent spindle, one end of the connecting rod is connected to the outer ring of the bearing, and the other end is connected to the exciter, the force sensor is arranged on the connecting rod, and the force sensor is used to monitor the force on the connecting rod; the embedded controller is connected to the wireless transmission module through the I / O port, and the wireless transmission module is used to transmit the monitoring data or the data processed by the embedded controller in real time.

[0006] Furthermore, the acceleration sensor is set on the intelligent spindle box near the front end bearing of the spindle by adsorption or pasting. This sensor installation position minimizes the cutting force transmission path, minimizes energy reduction, and increases sensitivity.

[0007] Furthermore, the embedded controller is also connected to a storage device and a display device through an I / O port. The storage device is used to store the monitored data in real time, and the display device is used to display the real-time monitored data or the data processed by the embedded controller.

[0008] Furthermore, a battery is provided to power the force sensor, acceleration sensor, storage device, display device and wireless transmission module, and the embedded controller, wireless communication device and battery are integrated into a module, which is installed on the intelligent spindle box by pasting or adsorption; the acceleration sensor and the embedded controller are connected via a serial port or a network cable.

[0009] Furthermore, the wireless communication module is a Lora, Zigbee, Bluetooth module, 4G or 5G communication module.

[0010] In a second aspect, the present invention provides an operating method of the intelligent spindle machining process monitoring system based on edge computing, comprising the following steps: By sweeping the frequency, we obtain the acceleration frequency response function from the tool tip of the intelligent spindle to the location where the acceleration sensor is installed in the box. Then, we obtain the acceleration frequency response function of the tool tip of the intelligent spindle at different speeds. Furthermore, obtaining the acceleration frequency response function of the tool tip of the intelligent spindle at different speeds includes: the exciter outputs vibrations from low to high frequencies, the connecting rod transmits the vibration output by the exciter to the bearing installed on the tool tip, the force sensor measures the force output by the exciter to the bearing, the acceleration sensor measures the acceleration signal corresponding to the force signal of different frequencies, the intelligent spindle adjusts the speed, and the acceleration frequency response function of the tool tip at the processing point of the intelligent spindle at different speeds can be obtained by sweeping the frequency of the exciter.

[0011] The convolution relationship between the acceleration (output response) and cutting force (input) of the machining system is shown below:

[0012] in, is the acceleration, is the impulse response function matrix, is the cutting force, It's noise.

[0013] The cutting force identification is converted into an ill-posed problem, and the regularization operator and the acceleration frequency response function are used to solve the ill-posed problem to obtain the cutting force identification result. The cutting force identified by the measured acceleration consists of two parts: acceleration and noise interference. When the impulse response function matrix When there are singular values close to zero, the noise that is difficult to avoid in the actual measurement process It will cause the identification results to be extremely inaccurate. The existence of items close to zero in the singular values will cause the Picard condition to be not satisfied, resulting in the cutting force identification problem not meeting the stability condition. From the analysis, it can be obtained that the main reason for the inaccurate cutting force identification is the impulse response function matrix The combined effect of the corresponding smaller singular values and the system measurement noise is unavoidable. Therefore, the intuitive idea to solve this problem is to correct the smaller singular values to reduce the identification error.

[0014] Furthermore, the rapid identification of cutting force is converted into an ill-posed problem, and the regularization operator and acceleration frequency response function are used to solve the ill-posed problem. The cutting force identification results obtained include: Considering the cutting force identification as an ill-posed problem, regularization is used to transform the transfer matrix and regularization operator into the following form:

[0015] in, is the regularization parameter, is the regularization operator ; Using a row of effective values in the transfer function matrix and the acceleration signal, the acceleration is first input into the circular queue data structure. The acceleration data is continuously input during the process of using the acceleration to inversely calculate the cutting force. The circular queue data structure contains half of the identified cutting force and half of the acceleration. When the cutting force is identified and output, the cutting force is inversely calculated using the acceleration, and the newly measured acceleration is input into the circular queue data structure. After the identified cutting force is output, the cutting force calculation obtained by the acceleration inverse is completed, that is, the circular queue data structure still contains half of the identified cutting force and half of the acceleration.

[0016] In a third aspect, the present invention can also provide an intelligent spindle equipped with the above-mentioned edge computing-based intelligent spindle processing monitoring system.

[0017] An industrial robot or industrial machine tool using the intelligent spindle may also be provided.

[0018] Compared with the existing technology, the present invention has at least the following advantages: the present invention designs an intelligent spindle processing process monitoring system based on edge computing. To address the dilemma that spindle intelligence strategies that require changing the original spindle structure and installing a large number of sensors are not suitable for actual industrial use, the present invention proposes an integrated intelligent spindle processing process monitoring system based on edge computing. Without changing any spindle structure, it can upgrade the supporting spindle of machine tools / industrial robots to an intelligent spindle, enabling it to have the capabilities of frequency response function acquisition, online cutting force identification, data visualization, data storage, and wireless data transmission; The accelerometer collects acceleration signals, the embedded controller processes, stores and displays the signals, the wireless transmission module is responsible for transmitting the signals to the cloud, and the battery is responsible for powering the system.

[0019] Furthermore, based on modal hammer tests and a novel frequency sweep device, the acceleration frequency response function of the tool tip of the intelligent spindle at rest and at different speeds was obtained. The transfer matrix and regularization operator properties were utilized to shorten the cutting force identification time to milliseconds. A circular queue data structure was used to enable online identification of cutting forces. Data from multiple intelligent spindles was collected, visualized, and shared, enabling the edge computing-based intelligent spindle process monitoring system to not only optimize maintenance strategies and parts inventory scheduling but also provide excellent human-computer interaction. The recursive nature of the regularization operator significantly reduced the time required for cutting force identification. Finally, based on the circular queue data structure, cutting forces could be acquired online using accelerometers. Replacing expensive dynamometers with accelerometers significantly reduced the cost of the intelligent spindle. The intelligent spindle collects a large amount of data during processing. The edge computing-based intelligent spindle process monitoring system, integrated with wireless communication devices, facilitates post-analysis and the transmission of important data to the cloud for long-term storage.

[0020] Furthermore, each intelligent spindle can communicate with each other and share information. Based on the data shared by multiple intelligent spindles, the scheduling and optimization of maintenance and parts inventory can be achieved. The data of any other intelligent spindle's processing process can be viewed through the display device of each intelligent spindle, allowing the operator to intuitively understand the processing status of any intelligent spindle, facilitating human-computer interaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 Schematic diagram of the frequency sweeping device of the present invention.

[0022] Figure 2 is the regularization operator feature map.

[0023] Figure 3 It is a schematic diagram of the method for rapid identification of cutting forces.

[0024] Figure 4 It is a schematic diagram of the data structure of the cutting force online identification method.

[0025] Figure 5 This is a schematic diagram of an intelligent spindle described in this application.

[0026] Among them: 1. Vibrator; 2. Force sensor; 3. Connecting rod; 4. Bearing; 5. Acceleration sensor; 6. Intelligent spindle. DETAILED DESCRIPTION

[0027] The following is a further description of specific embodiments of the present invention in conjunction with the accompanying drawings. It should be noted that the description of these embodiments is intended to facilitate understanding of the present invention and does not constitute a limitation of the present invention. In addition, the technical features involved in the various embodiments of the present invention described below may be combined with each other as long as they do not conflict with each other.

[0028] The present invention designs an integrated edge computing device, which is installed on the intelligent spindle housing by means of magnetism / adhesion. The edge computing device does not change any structure of the spindle. It can upgrade the spindle of machine tools / industrial robots to an intelligent spindle, enabling it to have the capabilities of frequency response function acquisition, online identification of cutting force, data visualization, data storage, and wireless data transmission.

[0029] The present invention designs an intelligent spindle machining process monitoring system based on edge computing to realize the intelligent upgrade of the existing machine tool / industrial robot supporting spindle.

[0030] For embedded controller, wireless communication device and battery integration, the acceleration sensor communicates with the embedded controller via serial port / network cable; As an optional embodiment, the wireless communication can be based on local area network communication such as Lora, Zigbee, Bluetooth, etc., or it can be based on 4G or 5G communication.

[0031] The edge computing system uploads the collected signals to the cloud as the intelligent spindle operation log through a wireless communication device; the intelligent spindle based on edge computing has a wireless communication function, and can upload the data that needs to be analyzed or important data to the cloud through the wireless communication device; each intelligent spindle can communicate with each other and share information, and based on the data shared by multiple intelligent spindles, the scheduling and optimization of maintenance and parts inventory can be realized; the data of any other intelligent spindle during the processing process can be viewed through the display device of each intelligent spindle, so that the operator can intuitively understand the processing status of any intelligent spindle, facilitating human-computer interaction.

[0032] The display device makes the intelligent spindle processing data visual, realizing good human-computer interaction.

[0033] The acceleration frequency response functions of the tool tip of a machine tool / industrial robot at rest or at different speeds are obtained through modal hammer testing or the system described in the present invention. Once the acceleration frequency response functions of the tool tip at all processing points are known, the transfer matrix of that processing point is also known, regardless of where the machine tool / industrial robot is processing. The cutting force identification problem is transformed into an ill-posed problem through convolution discretization, and then solved using a regularization algorithm. The characteristics of the transfer matrix and regularization operator are utilized to shorten the cutting force identification time to milliseconds. The cutting force identification method is embedded in the embedded controller, and a circular queue data structure is used to realize online identification of the cutting force; the processing data is stored to prepare for future optimization of the intelligent spindle processing process monitoring system based on edge computing.

[0034] This embodiment takes the commonly used spindle milling of machine tools / industrial robots as an example to illustrate the intelligent spindle processing monitoring system based on edge computing designed by the present invention, as follows.

[0035] The edge computing-based intelligent spindle processing monitoring system is installed on the intelligent spindle box by adsorption or pasting.

[0036] When a machine tool / industrial robot performs rough machining, the acceleration frequency response function of the tool tip at the machining point of the intelligent spindle under static conditions is obtained through modal hammer testing.

[0037] Figure 1 and Figure 5 As shown, the frequency sweeping device includes an exciter 1, a force sensor 2, a connecting rod 3, a bearing 4, and an acceleration sensor 5. The acceleration sensor 5 is arranged on the intelligent spindle housing, and the acceleration sensor is connected to the input end of the embedded controller. The bearing 4 is arranged at the output end of the intelligent spindle. One end of the connecting rod 3 is connected to the outer ring of the bearing 4, and the other end is connected to the exciter 1. The force sensor 2 is arranged on the connecting rod 3, and the force sensor 2 is used to monitor the force on the connecting rod 3.

[0038] During finishing of machine tools / industrial robots, Figure 1 The frequency sweeping device shown here obtains the acceleration frequency response function of the tool tip at the processing point of the intelligent spindle at different speeds. Specifically, the exciter 1 outputs vibrations from low to high frequencies. The connecting rod 3 transmits the vibrations output by the exciter 1 to the bearing 4 mounted on the tool tip. The force sensor 2 measures the force output by the exciter 1 on the bearing 4. The acceleration sensor 5 measures the acceleration signals corresponding to the force signals of different frequencies. The intelligent spindle 6 adjusts the speed. By sweeping the frequency of the exciter 1, the acceleration frequency response function of the tool tip at the processing point of the intelligent spindle at different speeds can be obtained.

[0039] The relationship between cutting force and acceleration is shown below.

[0040] (3) in, is the cutting force, is the transfer function matrix, is the acceleration.

[0041] The convolution relationship between the acceleration (output response) and cutting force (input) of the machining system is shown below:

[0042] in, is the acceleration, is the impulse response function matrix, is the cutting force, It's noise.

[0043] The cutting force identification is converted into an ill-posed problem, and the regularization operator and the acceleration frequency response function are used to solve the ill-posed problem to obtain the cutting force identification result. The cutting force identified by the measured acceleration consists of two parts: acceleration and noise interference. When the impulse response function matrix When there are singular values close to zero, the noise that is difficult to avoid in the actual measurement process The identification results will be extremely inaccurate. The existence of items close to zero in the singular values will cause the Picard condition to be not satisfied, resulting in the cutting force identification problem not meeting the stability condition. From the analysis, it can be concluded that the main reason for the inaccurate cutting force identification is the impulse response function matrix The combined effect of the corresponding small singular values and the system measurement noise. System measurement noise is unavoidable, so the intuitive way to solve this problem is to correct the small singular values to reduce the identification error.

[0044] In practice, the eigenvalues of the transfer function matrix may be close to zero, that is, the transfer function matrix may not be directly inverted. Therefore, the present invention considers cutting force identification as an ill-posed problem. Since the problem to be solved is ill-posed, regularization is used to transform the transfer matrix and regularization operator into the following form: (4) in, is the regularization parameter, , regularization operator The feature map of Figure 2 As shown. Regularization parameter The larger the value, the smaller the effect of measurement noise, the worse the approximation of the solution, and the regularization parameter The smaller the value, the opposite effect. Therefore, how to balance the relationship between the two and select the appropriate regularization parameter The value is the key. How to determine the regularization parameter There are many methods to calculate the value of , and the generalized cross-validation criterion (GCV) is used here.

[0045] refer to Figure 2 and Figure 3 Experiments have shown that each row of the regularization operator is recursive and repeatable. This property can significantly reduce the complexity of cutting force calculations. In other words, when solving for cutting forces, instead of using the transfer function matrix and the acceleration signal, the effective values (values significantly greater than zero) of a row in the transfer function matrix and the acceleration signal are used. This significantly reduces the time required to identify the cutting force.

[0046] (5) in is a row of valid values in the transfer function matrix, is the acceleration, is the column translation matrix.

[0047] Then, using the first-in-first-out circular queue data structure, refer to Figure 4 , realizing online identification of cutting force. First, input acceleration into the circular queue data structure, and continuously input acceleration data while using acceleration to inversely calculate cutting force. At this time, the circular queue data structure contains half of the identified cutting force and half of the acceleration. When the cutting force is identified and output, the cutting force is inversely calculated using acceleration, and the newly measured acceleration is input into the circular queue data structure. After the identified cutting force is output, the calculation of the cutting force inversely calculated using acceleration is completed, that is, the circular queue data structure again contains half of the identified cutting force and half of the acceleration. The use of the circular queue data structure can realize indirect identification of cutting force when the cutting force identification time is relatively short.

[0048] The data collected and processed by the intelligent spindle can be transmitted to the cloud via a wireless transmission device, serving as the basis for logs, maintenance, and parts inventory scheduling, providing a reference for future optimization of the intelligent spindle machining process monitoring system based on edge computing.

[0049] Example 3: The present invention can also provide an intelligent spindle equipped with the above-mentioned edge computing-based intelligent spindle processing monitoring system.

[0050] In Example 4, an industrial robot or industrial machine tool using the above-mentioned intelligent spindle can also be provided; and the intelligent spindle can obtain cutting force data according to the above-mentioned operating method, perform edge computing, and save the data.

[0051] In summary, the present invention provides an edge computing-based intelligent spindle machining process monitoring system. The functions of this edge computing-based intelligent spindle machining process monitoring system include: acceleration acquisition, frequency response function acquisition, online cutting force identification, data visualization, data storage, and wireless data transmission. An accelerometer, embedded controller, wireless communication device, and battery are integrated and mounted on the intelligent spindle housing via adhesive bonding or adhesive bonding. Before machining, a hammer test or a novel frequency sweep device is used to obtain the acceleration frequency response function at the tool tip of the intelligent spindle at rest and at different speeds. During machining, the intelligent spindle only requires real-time acceleration signals for online cutting force identification. Data visualization facilitates human-machine interaction, and stored data is uploaded to the cloud via wireless transmission for use as a reference for machining logs, maintenance, and parts inventory scheduling. This invention does not require any changes to the spindle structure; simply by attaching the edge computing-based intelligent spindle machining process monitoring system to the spindle housing via adhesive bonding or adhesive bonding, it can upgrade existing spindles used in machine tools or industrial robots to intelligent spindles with machining process monitoring capabilities.

[0052] The above content is only for explaining the technical idea of the present invention and cannot be used to limit the protection scope of the present invention. Any changes made on the basis of the technical solution in accordance with the technical idea proposed by the present invention shall fall within the protection scope of the claims of the present invention.

Claims

1. The intelligent spindle machining process monitoring system based on edge computing is characterized by: The invention comprises an exciter (1), a force sensor (2), a connecting rod (3), a bearing (4), and an acceleration sensor (5); the acceleration sensor (5) is arranged on the intelligent spindle housing, and the acceleration sensor is connected to the input end of the embedded controller, wherein the acceleration sensor is used to collect vibration acceleration signals during the intelligent spindle processing, and the embedded controller is used to execute frequency response function calculation and cutting force online identification method; the bearing (4) is arranged at the output end of the intelligent spindle, one end of the connecting rod (3) is connected to the outer ring of the bearing (4), and the other end is connected to the exciter (1), the force sensor (2) is arranged on the connecting rod (3), and the force sensor (2) is used to monitor the force of the connecting rod (3); the embedded controller is connected to the wireless transmission module through the I / O port, and the wireless transmission module is used to transmit the monitoring data or the data processed by the embedded controller in real time.

2. The intelligent spindle machining process monitoring system based on edge computing according to claim 1 is characterized in that: The acceleration sensor (5) is arranged on the intelligent spindle housing by adsorption or adhesion.

3. The intelligent spindle machining process monitoring system based on edge computing according to claim 1 is characterized in that: The embedded controller is also connected to a storage device and a display device through an I / O port. The storage device is used to store the monitored data in real time, and the display device is used to display the monitored data in real time or the data processed by the embedded controller.

4. The intelligent spindle machining process monitoring system based on edge computing according to claim 3 is characterized in that: A battery is provided to power the force sensor (2), the acceleration sensor (5), the storage device, the display device and the wireless transmission module; the embedded controller, the wireless communication device and the battery are integrated into a module, and the module is installed on the intelligent spindle housing by pasting or adsorption; the acceleration sensor (5) is connected to the embedded controller by a serial port or a network cable.

5. The intelligent spindle machining process monitoring system based on edge computing according to claim 1 is characterized in that: The wireless communication module is Lora, Zigbee, Bluetooth module, 4G or 5G communication module.

6. The method for operating the intelligent spindle machining process monitoring system based on edge computing according to any one of claims 1 to 5, characterized in that: The following steps are involved: Obtain the frequency response function of tool tip acceleration of the intelligent spindle at different speeds; The rapid identification of cutting force is converted into an ill-posed problem, which is solved by using a regularization operator and an acceleration frequency response function to obtain the cutting force identification result.

7. The operating method according to claim 1, characterized in that: The method for obtaining the acceleration frequency response function of the tool tip of the intelligent spindle at different speeds includes: the vibrator (1) outputs vibrations from low to high frequencies, the connecting rod (3) transmits the vibration output by the vibrator (1) to the bearing (4) installed at the tool tip, the force sensor (2) measures the force output by the vibrator (1) to the bearing (4), the acceleration sensor (5) measures the acceleration signal corresponding to the force signal of different frequencies, and the intelligent spindle (6) adjusts the speed. By sweeping the frequency of the vibrator (1), the acceleration frequency response function of the tool tip at the processing point of the intelligent spindle at different speeds can be obtained.

8. The operating method according to claim 6, characterized in that: The cutting force rapid identification is converted into an ill-posed problem, which is solved using the regularization operator and the acceleration frequency response function. The cutting force identification results include: Considering the cutting force identification as an ill-posed problem, regularization is used to transform the transfer matrix and regularization operator into the following form: in, is the regularization parameter, is the regularization operator ; Using a row of effective values in the transfer function matrix and the acceleration signal, the acceleration is first input into the circular queue data structure. The acceleration data is continuously input during the process of using the acceleration to inversely calculate the cutting force. The circular queue data structure contains half of the identified cutting force and half of the acceleration. When the cutting force is identified and output, the cutting force is inversely calculated using the acceleration, and the newly measured acceleration is input into the circular queue data structure. After the identified cutting force is output, the cutting force calculation obtained by the acceleration inverse is completed, that is, the circular queue data structure still contains half of the identified cutting force and half of the acceleration.

9. An intelligent spindle, characterized in that: Equipped with an intelligent spindle machining process monitoring system based on edge computing as described in any one of claims 1 to 5.

10. The application of the intelligent spindle according to claim 9, characterized in that: For use in industrial robots or industrial machine tools.

Images