Milling wireless vibration measurement cutter handle manufacturing system based on FPC flexible packaging

Abstract

A milling wireless vibration measurement cutter handle manufacturing system based on FPC flexible packaging comprises a milling machine, a wireless intelligent cutter handle arranged on the milling machine, a cutter located at the tail end of the wireless intelligent cutter handle and a workpiece arranged opposite to the cutter, the wireless intelligent cutter handle is connected with an upper computer in a wireless mode and transmits three-way vibration signals of the cutter handle position in the cutting process, the three-way vibration signals are processed and analyzed through the upper computer through an artificial intelligence algorithm, the cutter abrasion value is monitored in real time, the remaining life of a cutter is predicted, and cutting parameters are optimized. And more reasonable cutting parameters are provided for cutting machining of the workpiece, and the service life of the cutter is prolonged. The reduction of strength and rigidity caused by the damage of the original cutter handle structure is made up through the composite material, and the feasibility in a high-speed and high-precision cutting processing task is ensured; meanwhile, the signal acquisition module and the charging module are flexibly distributed on the outer cylindrical surface of the knife handle based on the FPC, the overall size of the intelligent knife handle is minimized, and the applicability in the complex path working condition is guaranteed.

Description

Technical Field

[0001] The present invention relates to a technology in the field of intelligent manufacturing, in particular to a milling wireless vibration measuring tool handle manufacturing system based on FPC flexible packaging. Background Art

[0002] Cutting vibration data is a signal in the cutting process. Monitoring and analyzing it can realize real-time monitoring of the tool status in the process, avoid waste caused by early replacement in order to ensure the processing quality in the actual processing process, and achieve cost reduction and efficiency improvement in the processing process. With the development of sensor integration technology, smart tool handles that integrate sensors on the tool handle have been widely studied. Compared with traditional wired vibration sensors, the sensors in smart tool handles are closer to the cutting area and can collect more realistic vibration signals. Moreover, the position of the sensor in the smart tool handle is fixed, avoiding the problem of poor repeatability caused by the non-fixed installation position of the sensor each time. However, there are many problems with existing smart tool handles. The overall size of some smart tool handles is too large, which has limitations in complex processing path tasks; some smart tool handles have destroyed the original structure of the tool handle in order to reduce the size, resulting in a decrease in the strength and stiffness of the tool handle, which has limitations in high-speed and high-precision processing tasks; and the acquisition module is deployed in the metal tool handle. There is an electromagnetic shielding problem, which causes the acquisition signal to be distorted. Summary of the invention

[0003] In view of the fact that the existing technology is difficult to meet the needs of engineering inspection in terms of strength and rigidity, and is prone to generate greater vibration during the processing, especially when the rigidity is insufficient, the tool holder is prone to offset under the action of cutting force, causing the cutting path to deviate from the predetermined trajectory, seriously affecting the processing accuracy and quality. The present invention proposes a milling wireless vibration measurement tool holder manufacturing system based on FPC flexible packaging, which compensates for the decrease in strength and rigidity caused by the destruction of the original tool holder structure through composite materials, ensuring the feasibility of high-speed and high-precision cutting tasks; at the same time, the signal acquisition module and the charging module are flexibly distributed on the outer cylindrical surface of the tool holder based on FPC, minimizing the overall size of the intelligent tool holder and ensuring the applicability in complex path conditions. At the same time, the deployment of the acquisition module on the composite material can avoid electromagnetic shielding and ensure the quality of the collected signal.

[0004] The present invention is achieved through the following technical solutions:

[0005] The present invention relates to a milling wireless vibration measuring toolholder manufacturing system based on FPC flexible packaging, comprising: a milling machine, a wireless intelligent toolholder arranged thereon, a tool located at the end thereof, and a workpiece arranged relative to the tool, wherein: the workpiece is arranged on a workbench of the milling machine through a vise, the wireless intelligent toolholder is connected to a host computer in a wireless manner and transmits a three-dimensional vibration signal of the toolholder position during the cutting process, the host computer processes and analyzes the three-dimensional vibration signal through an artificial intelligence algorithm, realizes real-time monitoring of the tool wear value, predicts the remaining life of the tool and optimizes the cutting parameters, provides more reasonable cutting parameters for the cutting processing of the workpiece, and improves the service life of the tool.

[0006] The artificial intelligence algorithm refers to: using Kolmogorov-Arnold neural network (KAN) to monitor tool wear status, decomposing complex high-dimensional functions into simple one-dimensional function combinations to handle complex data fitting problems, and realizing accurate mapping between cutting vibration signals and tool wear, specifically: Where: Φ q,p is a spline curve, Φ q,p and Φ q is a univariate continuous function, X is the feature set from the vibration signal, and f is the true mapping between the feature set and the tool wear value.

[0007] The wireless intelligent tool handle comprises: a tool handle base and a turning part connected in sequence and a composite material reinforcement layer arranged outside the turning part, wherein: an annular groove and two module grooves are arranged on the composite material reinforcement layer.

[0008] The composite material reinforcement layer is realized by the following method:

[0009] i) Establish different toolholder models for simulation, determine the maximum thickness of the turning part and the direction of laying while ensuring the overall strength and stiffness of the toolholder, and determine the number of layers of composite material laid according to the wall thickness of the turning part of the toolholder;

[0010] ii) Apply the resin evenly on the turning part of the tool handle, wrap the composite material layer by layer on the cylindrical surface, use rollers to remove bubbles, ensure that the composite material fits tightly, and use a vacuum pump to extract air to create a negative pressure environment to remove internal bubbles and improve adhesion. After the composite material is wrapped for the second time, vacuum is drawn again, and finally high-temperature curing is performed. FPC flexible module slots and soft cable layout slots are processed through micro-machining technology.

[0011] iii) The static structural strength, dynamic structural strength and natural frequency of the prepared composite material reinforced tool handle are verified, and the FPC flexible module is assembled and a signal acquisition test is performed.

[0012] A lithium battery is arranged in the annular groove.

[0013] The two module slots are respectively provided with an FPC-packaged signal acquisition module and an FPC-packaged charging module, and the charging module is connected to a lithium battery to stabilize the input voltage between the signal acquisition module and the wireless transmission module.

[0014] The FPC-packaged signal acquisition module comprises: an acceleration sensor, a microprocessor and a Wi-Fi transmission module, wherein: the acceleration sensor acquires the vibration signal of the tool handle during the cutting process and transmits it to the microprocessor through serial communication; the microprocessor transmits the processed signal to the Wi-Fi module as the server based on the serial communication; the Wi-Fi module communicates with the host computer as the client based on the TCP / IP protocol.

[0015] The charging module includes: a power management chip, a voltage stabilizing chip and a ring-shaped lithium battery, wherein: the power management chip realizes the conversion, distribution and monitoring of electric energy to ensure the power supply requirements of the acquisition module; the voltage stabilizing chip provides a stable 3.3V voltage for the acquisition module to limit voltage fluctuations; the ring-shaped lithium battery is the power supply of the acquisition module.

[0016] The FPC packaging is to select flexible printed materials as the substrate of the flexible collection module and the charging module, transfer the circuit pattern to the substrate through photolithography technology, install the electronic components on the FPC flexible circuit board using surface mounting technology, and use a protective coating to package the entire circuit to enhance durability.

[0017] A protective shell is arranged outside the composite material reinforcement layer. Technical Effects

[0018] The present invention designs and manufactures a vibration signal acquisition module and a charging module based on the FPC flexible packaging process and applies them to a wireless intelligent tool handle to realize the collection of vibration signals during the cutting process, uses KAN to monitor the wear state of the tool, and constructs a KAN-Wear wear monitoring model. Compared with the prior art, the signal acquisition module and charging module designed and manufactured based on the FPC flexible packaging process of the present invention can make full use of the outer cylindrical surface size of the tool handle, reduce the structural damage to the tool handle, and avoid excessive reduction in the stiffness and strength of the tool handle; at the same time, the overall size of the wireless intelligent tool handle can be reduced, which is completely consistent with the size of the traditional commercial tool handle of the same model. When it is used in the actual processing process later, the user does not need to modify the process. The traditional square hard signal acquisition module and charging module designed and manufactured based on the PCB board are difficult to simultaneously achieve the requirements of small structural damage and small overall size. The flexible signal acquisition module and the flexible charging module of the present invention are deployed in the composite material reinforcement layer. Compared with the existing wireless intelligent tool handle deployed inside the metal tool handle, the wireless intelligent tool handle of the present invention can avoid electromagnetic shielding and effectively improve the signal transmission quality. Compared with traditional CNN, LSTM, MLP and other models, the KNN tool wear state monitoring model KAN-Wear based on KAN can converge faster, and the RMSE and MAE values ​​of tool wear monitoring on the test set are greatly improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 This is a working condition diagram of the wireless intelligent knife handle of the present invention;

[0020] Figure 2 This is a composite material reinforcement structure diagram of the wireless intelligent knife handle of the present invention;

[0021] Figure 3 This is an overall assembly diagram of the wireless intelligent knife handle of the present invention;

[0022] Figure 4 It is the KAN-Wear tool status monitoring model structure embedded in the wireless intelligent tool handle software system of the present invention;

[0023] Figure 5 Schematic diagram of the test conditions of the PHM2010 public cutting data set used in the present invention to verify the effectiveness of the KAN-Wear model;

[0024] Figure 6 This is the full life cycle wear curve of three tools in the PHM2010 public cutting data set. DETAILED DESCRIPTION

[0025] like Figure 1As shown, this embodiment involves a manufacturing system of the above-mentioned wireless intelligent tool handle system, including: a milling machine 10, a wireless intelligent tool handle 20 arranged thereon, a tool 30 located at the end thereof, and a workpiece 40 arranged relative to the tool 30, wherein: the workpiece 40 is set on the workbench of the milling machine 10 through a vise 50, the wireless intelligent tool handle 20 is connected to the host computer 60 by wireless means and transmits a three-way vibration signal of the tool handle position during the cutting process, and the host computer 60 processes and analyzes the three-way vibration signal through an artificial intelligence algorithm to realize real-time monitoring of the wear value of the tool 30, predict the remaining life of the tool and optimize the cutting parameters, so as to provide more reasonable cutting parameters for the cutting processing of the workpiece 40 and improve the service life of the tool 30.

[0026] like Figure 2 As shown, the wireless intelligent tool handle 20 includes: a turning part 21 on the tool handle base connected in sequence and a composite material reinforcement layer 22 arranged outside the turning part 21, wherein: an annular groove 23, a signal acquisition module groove 24 and a charging module groove 25 are provided on the composite material reinforcement layer 22.

[0027] The handle base 21 is based on the mainstream commercial HSK-63A handle as an example; in order to avoid the electromagnetic shielding problem caused by deploying the acquisition module on the metal material and to avoid the overall size of the smart handle being too large, the handle base 21 is turned to form a turned part 22 at the end.

[0028] The built-in algorithms in the life prediction module 53 and the wear monitoring module 54 are based on deep learning. After the three-way raw data is input into the algorithm, the life prediction and wear monitoring of the tool are realized through KAN. The operator does not need to extract features during the whole process, which is an end-to-end monitoring algorithm. Figure 4 The KAN-Wear structure is described. The input is the time domain features (maximum value, mean value and variance) of the cutting vibration data, and the number of nodes in the three layers of KAN is set to 3, 7 and 1 respectively.

[0029] The PHM2010 public cutting dataset (https: / / phmsociety.org / phm_competition / 2010-phm-society-conference-data-challenge / ) is used to verify the KAN-Wear tool condition monitoring algorithm. The test conditions of the PHM2010 public milling dataset are shown in the figure below. Figure 5As shown. The experimental machine tool uses Roder Tech RFM769 high-speed CNC machine tool 61. The workpiece 65 is clamped on the workbench of the high-speed CNC machine tool 61. The tool used is a 3-tooth ball head milling cutter 62. The cutting method adopts dry milling and side milling. The length of each pass is 108mm. During the cutting process, the cutting force in three directions was measured using a Kistler three-component platform dynamometer 66. In addition, a Kistler piezoelectric accelerometer 63 was deployed to measure the three-axis vibration of the workpiece. A Kistler acoustic emission sensor 64 was used to measure the high-frequency stress waves generated during the cutting process. The data collected by these three sensors were uniformly collected through the NI data acquisition card 68 to form a 7-channel signal, which is the cutting force in the x-direction, the cutting force in the y-direction, the cutting force in the z-direction, the vibration in the x-direction, the vibration in the y-direction, the vibration in the z-direction and the acoustic emission signal. After each pass, the average wear bandwidth VB of the back face of each blade was measured using a LEICMZ12 microscope 67 as the wear value of the blade. The average wear value of the three blades is taken as the wear value of the entire tool. The tool life cycle wear curves of the three groups of experiments are as follows Figure 6 shown.

[0030] Each tool pass will generate a large amount of data. Directly using the data of the entire tool pass as sample input will result in too much calculation of the model and may even cause overfitting of the model. This embodiment divides the data of each tool pass into 100 subsequences, and extracts the maximum value, mean and variance of the subsequences respectively, thereby forming a shorter time series (100×3) as the input of the KAN-Wear tool monitoring model. In the tool wear value monitoring task, the VB value is directly used as the label of each input feature; in the tool remaining life prediction task, VB=150μm is used as the threshold, and the usage time when the back face wear of the three milling cutters reaches 150μm is calculated as the tool life, and the remaining life of each tool pass is calculated as the label of each input feature, specifically R L =TT i , where: R L is the remaining life, T is the tool life, T i is the cumulative usage time of the tool for the i-th cutting pass.

[0031] Table 1 Specific settings of the cross dataset

[0032] This embodiment specifically adopts the KAN-Wear model and uses GPU for training under the Windows 10 platform. The graphics card model is RTX 3090, and the CPU model is AMD EPYC 7624. The KAN-Wear network is built with the Pytorch 1.11.0 framework. The Python version is 3.8, and the Cuda version is 11.3. Adam is used as the training optimizer during training, and the generalization of the model is improved by setting weight decay to prevent overfitting in model training. The loss function is calculated using MAE, and the hyperparameter settings during model training are shown in Table 2. The tool grinding monitoring model and the tool remaining life prediction model are trained in data sets T1, T2 and T3 respectively, and the performance of the model in the test set is shown in Table 3.

[0033] Table 2 Model training hyperparameters

[0034] like Figure 3 As shown, the overall assembly of the wireless intelligent knife handle involved in this embodiment includes: a knife handle base 31, three annular lithium batteries 32, a composite material reinforcement layer 33, an FPC-packaged signal acquisition module 34, an FPC-packaged charging module 34, a protective shell 36 and a wireless intelligent knife handle 37.

[0035] The annular lithium battery 32 is arranged in the annular groove of the composite material reinforcement layer 33; the FPC-packaged signal acquisition module 34 is assembled in the left groove of the composite material reinforcement layer 33; the FPC-packaged signal acquisition module 35 is assembled in the right groove of the composite material reinforcement layer 33; the composite material reinforcement layer 33, the signal acquisition module 34, the charging module 35 and the three annular lithium batteries 32 are all encapsulated in the protective shell 36 to obtain an assembled wireless intelligent knife handle 37.

[0036] The above-mentioned specific implementation can be partially adjusted in different ways by those skilled in the art without departing from the principle and purpose of the present invention. The protection scope of the present invention shall be based on the claims and shall not be limited by the above-mentioned specific implementation. Each implementation scheme within its scope shall be subject to the constraints of the present invention.

Claims

1. A manufacturing system for a wireless intelligent tool handle based on FPC flexible packaging, characterized in that: include: A milling machine and a wireless intelligent tool holder arranged thereon, a tool located at the end thereof, and a workpiece arranged relative to the tool, wherein: the workpiece is set on a workbench of the milling machine through a vise, the wireless intelligent tool holder is connected to a host computer wirelessly and transmits a three-dimensional vibration signal of the tool holder position during the cutting process, the host computer processes and analyzes the three-dimensional vibration signal through an artificial intelligence algorithm, realizes real-time monitoring of the tool wear value, predicts the remaining life of the tool and optimizes the cutting parameters, provides more reasonable cutting parameters for the cutting processing of the workpiece, and improves the service life of the tool.

2. The manufacturing system of the wireless intelligent tool handle based on FPC flexible packaging according to claim 1 is characterized in that: The artificial intelligence algorithm refers to: using Kolmogorov-Arnold neural network (KAN) to monitor tool wear status, decomposing complex high-dimensional functions into simple one-dimensional function combinations to handle complex data fitting problems, and realizing accurate mapping between cutting vibration signals and tool wear, specifically: Where: Φ q,p is a spline curve, Φ q,p and Φ q is a univariate continuous function, X is the feature set from the vibration signal, and f is the true mapping between the feature set and the tool wear value.

3. The manufacturing system of the wireless intelligent tool handle based on FPC flexible packaging according to claim 1 is characterized in that: The wireless intelligent tool handle comprises: a tool handle base and a turning part connected in sequence and a composite material reinforcement layer arranged outside the turning part, wherein: an annular groove and two module grooves are arranged on the composite material reinforcement layer.

4. The manufacturing system of the wireless intelligent tool handle based on FPC flexible packaging according to claim 1 is characterized in that: The composite material reinforcement layer is realized by the following method: i) Establish different toolholder models for simulation, determine the maximum thickness of the turning part and the direction of laying while ensuring the overall strength and stiffness of the toolholder, and determine the number of layers of composite material laid according to the wall thickness of the turning part of the toolholder; ii) Apply the resin evenly on the turning part of the tool handle, wrap the composite material layer by layer on the cylindrical surface, use a roller to remove bubbles to ensure that the composite material fits tightly, and use a vacuum pump to extract air to create a negative pressure environment to remove internal bubbles and improve adhesion. After the composite material is wrapped for the second time, vacuum is drawn again, and finally high-temperature curing is performed. The FPC flexible module slot and the soft cable layout slot are processed through micro-machining technology; iii) The static structural strength, dynamic structural strength and natural frequency of the prepared composite material reinforced tool handle are verified, and the FPC flexible module is assembled and a signal acquisition test is performed.

5. The manufacturing system of wireless intelligent tool handle based on FPC flexible package according to claim 1 is characterized in that: The two module slots are respectively provided with an FPC-packaged signal acquisition module and an FPC-packaged charging module, and the charging module is connected to a lithium battery to stabilize the input voltage between the signal acquisition module and the wireless transmission module.

6. The manufacturing system of the wireless intelligent tool handle based on FPC flexible packaging according to claim 5 is characterized in that: The FPC-packaged signal acquisition module comprises: an acceleration sensor, a microprocessor and a Wi-Fi transmission module, wherein: the acceleration sensor acquires the vibration signal of the tool handle during the cutting process and transmits it to the microprocessor through serial communication; the microprocessor transmits the processed signal to the Wi-Fi module as the server based on the serial communication; the Wi-Fi module communicates with the host computer as the client based on the TCP / IP protocol.

7. The manufacturing system of the wireless intelligent tool handle based on FPC flexible packaging according to claim 5 is characterized in that: The charging module includes: a power management chip, a voltage stabilizing chip and an annular lithium battery, wherein: the power management chip realizes the conversion, distribution and monitoring of electric energy to ensure the power supply requirements of the acquisition module; the voltage stabilizing chip provides a stable 3.3V voltage for the acquisition module to limit voltage fluctuations; the annular lithium battery is arranged in the annular groove and is the power supply of the acquisition module.

8. The manufacturing system of wireless intelligent tool handle based on FPC flexible package according to claim 1 is characterized in that: A protective shell is arranged outside the composite material reinforcement layer.

Images