CN-117400062-B - Cutter wear monitoring method and device based on digital twin driving

CN117400062BCN 117400062 BCN117400062 BCN 117400062BCN-117400062-B

Abstract

The invention discloses a cutter wear monitoring method and a cutter wear detecting device based on digital twin driving, relates to a cutter wear monitoring method and a cutter wear detecting device, and aims to monitor and predict the cutter wear state in the milling process of a numerical control machine tool so as to realize three-dimensional visual monitoring of the cutter wear state. The device comprises a workpiece conveying mechanism, a depth camera bracket, a depth camera, a CCD camera, a tail end rotating mechanism, an X-axis moving mechanism, a Y-axis moving mechanism and a Z-axis moving mechanism. The method comprises the steps of constructing a digital space in a milling process of a numerical control machine tool, establishing connection between a physical space and the digital space, determining the position of a milling surface of a workpiece according to acquired data, controlling a triaxial movement mechanism by a digital twin system to drive a CCD camera to acquire images of the milling surface of the workpiece at a viewpoint, constructing a cutter monitoring model based on SVM, and acquiring image data in real time. The invention belongs to the technical field of intelligent manufacturing.

Inventors

LI MAOYUE
Sha Baiying
SU YUANQIANG
ZHANG CHENGLONG

Assignees

哈尔滨理工大学

Dates

Publication Date: 20260505
Application Date: 20231024

Claims (6)

1. A digital twin-driven tool wear monitoring method is characterized by being realized by a tool wear detection device, and comprises a workpiece conveying mechanism (1), a depth camera support (3), a depth camera (4), a CCD (charge coupled device) camera (5) and a tail end rotating mechanism (6), an X-axis moving mechanism (8), two Y-axis moving mechanisms (7) and two Z-axis moving mechanisms (9), wherein the depth camera support (3) is arranged on a support body of the conveying mechanism (1), the depth camera (4) is arranged on the depth camera support (3) above the workpiece conveying mechanism (1), the two Z-axis moving mechanisms (9) are symmetrically arranged on the support bodies on two sides of the workpiece conveying mechanism (1), each Z-axis moving mechanism (9) is provided with a Y-axis moving mechanism (7) in a sliding mode, the two Y-axis moving mechanisms (7) are vertically arranged relatively, the X-axis moving mechanism (8) is arranged on the two Y-axis moving mechanisms (7) in a sliding mode, the CCD (5) is arranged on the X-axis moving mechanism (8) through the tail end rotating mechanism (6) and the workpiece conveying mechanism (1), The method comprises the following steps: Firstly, constructing a virtual space in the milling process of a numerical control machine tool, and performing real-time milling simulation; establishing connection between a physical space and a virtual space, and acquiring multi-element data from the physical space, wherein the multi-element data comprises static information and dynamic information of the physical space; Step three, acquiring scene point clouds according to a depth camera, denoising and dividing the scene point clouds to obtain target point clouds, performing point clouding treatment on a workpiece model to obtain source point clouds, and registering the source point clouds and the target point clouds to obtain the pose of the workpiece in the scene; Extracting a workpiece coordinate system and the position of a milling surface from an NC program of a numerical control system, and combining the position of the milling surface with the pose of the workpiece to obtain the position of the milling surface on a detection device, and solving a viewpoint when acquiring images according to a model of the workpiece, the position of the milling surface and optical parameters of a camera; Controlling a triaxial movement mechanism by the digital twin system to drive a CCD camera to acquire milling surface images of the workpiece at the view point; constructing a cutter monitoring model based on an SVM, and building a data set to train a test model, so that the model has higher recognition precision and generalization capability; and step six, collecting image data in real time, inputting the image data into a model to obtain a cutter abrasion value, and sending out a warning to remind a cutter to change when the cutter abrasion monitoring value exceeds a cutter abrasion threshold.
2. The method for monitoring tool wear driven by digital twin according to claim 1, wherein the constructing of the virtual space in the first step comprises: three-dimensional modeling, namely establishing a three-dimensional model in a virtual space according to an actual numerical control machine tool and detection equipment in industrial modeling software, and optimizing the model in 3D modeling software; scene construction, namely importing the optimized model into three-dimensional visualization software to perform scene construction; The method comprises the steps of performing a behavior logic model construction, namely, after completing scene construction, giving specific behaviors and logic relations of virtual model motions of a virtual space, performing milling simulation on a numerical control machine tool driven by a plurality of data according to data acquired by an external sensor and a machine tool communication protocol in real time, variables bound by an association model and preset behavior actions, constructing father-son relations among model tree structures, and perfecting machine tool virtual space construction; and man-machine interaction, namely writing a system interface in three-dimensional visual software to realize three-dimensional visual monitoring of the abrasion state of the cutter.
3. The method for monitoring tool wear driven by digital twin according to claim 1, wherein the establishing a connection between the physical space and the virtual space in the second step comprises: Registering an OPC UA server in a numerical control system; Establishing TCP/IP network layer communication connection between the physical space and the virtual space through a Socket interface; After the physical space is communicated with the machine tool and the detection device in the virtual space, the OPC UA server is accessed to obtain the running state information of the machine tool and the milling surface image data of the workpiece, and the virtual space receives the multivariate data from the physical space and stores the multivariate data into the MySQL database.
4. The method of claim 1, wherein the static information comprises machine tool attribute data, workpiece attribute data and process attribute data, wherein the machine tool attribute data comprises a device model and a system model, the tool attribute data comprises a tool name, a tool material and a tool parameter, the workpiece attribute data comprises a workpiece name, a workpiece material and a workpiece size, the process attribute data comprises a clamping mode and a processing parameter, the processing parameter comprises a cutting speed, a feeding speed and a back cutting amount, and the dynamic information is machine tool state data.
5. A digital twin-driven tool wear monitoring method according to claim 1 is characterized in that an SVM machine learning model is constructed, a gray level co-occurrence matrix is used for extracting image contrast and secondary diagonal moment from an acquired image, the two extracted features are input into the SVM model to predict the wear state of a tool, a tool wear data set is constructed, a training set and a testing set of each wear state are randomly divided into the data set according to the proportion of 3:1, feature extraction and feature selection are respectively carried out on the training set and the testing set, a feature vector set is determined, normalization processing is carried out, parameter optimization is carried out on all points in the searching range by using a grid search method, so that most super-parameter combinations are obtained, cross verification is repeated for 3 times, and finally, the tool monitoring model is obtained.
6. The method for monitoring the wear of the digital twin-drive cutter according to claim 1, wherein the wear value of the cutter is obtained, and whether an early warning is sent out or not is judged according to a threshold value so as to remind the cutter replacement.

Description

Cutter wear monitoring method and device based on digital twin driving Technical Field The invention relates to a cutter wear monitoring method and a detection device, in particular to a cutter wear monitoring method and a detection device based on digital twin driving, and belongs to the technical field of intelligent manufacturing. Background With the advent of big data, internet of things, cloud computing, and other technologies, digital twinning has also been gradually proposed to facilitate the transformation of manufacturing. Digital twin is used as an advanced technology for trampling intelligent manufacturing and industry 4.0, and the attribute, the characteristic, the behavior, the performance and the like of a physical entity are described and modeled in a data driving mode, so that the digital twin is gradually applied to the fields of manufacturing workshops, ships, vehicles and the like. In recent years, digital twinning is gradually developed in the industrial field, and a new solution is provided for monitoring and optimizing mechanical products. In the field of machine manufacturing, a tool is used as a "tooth" of a machine tool, and the cutting state thereof has an important influence on the stable progress of the entire machining process. The cutter generates abrasion and damage phenomena under the action of thermal coupling in the cutting process. When the cutter is severely worn, the processing quality and efficiency of the workpiece are affected, and the machine tool can be stopped, so that the normal operation of the whole processing system is affected. During machine tool downtime, the tool wear causes a downtime of 20% of the total downtime. Therefore, whether the abrasion state of the cutter can be timely and accurately monitored is an important premise for guaranteeing the quality safety of the workpiece and the machine tool. In conventional machining processes, in order to achieve monitoring of the tool state, direct monitoring and indirect monitoring are generally classified. Direct monitoring is by direct visual measurement of the wear state of the cutting tool and indirect monitoring is by measuring the wear level of the tool using the process signal. Since the side area of the tool is difficult to measure when directly monitoring, the online capture process signal is used to monitor the wear state of the tool. The periodic contact of the tool with the workpiece during milling, due to the nature of this complex process mechanism, results in periodic process signals which lead to unpredictable and premature tool wear during milling, and the use of sensors which are in direct contact with the machine tool for measuring the relevant signals, can create an obstacle to machining. To overcome these problems, the present invention proposes a digital twin-drive non-haptic, low cost and high speed tool wear monitoring method. Disclosure of Invention The invention aims to provide a cutter wear monitoring method and a cutter wear detection device based on digital twin driving, which are used for monitoring and predicting the cutter wear state in the milling process of a numerical control machine tool, so as to realize three-dimensional visual monitoring of the cutter wear state. The technical scheme adopted for solving the technical problems is as follows: The tool wear detection device based on digital twin driving comprises a workpiece conveying mechanism, a depth camera support, a depth camera, a CCD (charge coupled device) camera, a tail end rotating mechanism, an X-axis moving mechanism, two Y-axis moving mechanisms and two Z-axis moving mechanisms, wherein the depth camera support is arranged on a conveyor framework body, the depth camera is arranged on the depth camera support above the workpiece conveying mechanism, the two Z-axis moving mechanisms are symmetrically arranged on the frame bodies on two sides of the workpiece conveying mechanism, each Z-axis moving mechanism is slidably provided with one Y-axis moving mechanism, the two Y-axis moving mechanisms are vertically arranged relatively, the X-axis moving mechanism is slidably arranged on the two Y-axis moving mechanisms, the CCD camera is arranged on the X-axis moving mechanism through the tail end rotating mechanism, and the workpiece is arranged on a conveyor belt of the workpiece conveying mechanism. A method of digital twin driven tool wear monitoring, the method comprising the steps of: Firstly, constructing a virtual space in the milling process of a numerical control machine tool, and performing real-time milling simulation; Establishing connection between a physical space and a virtual space, and acquiring multi-element data from the physical space; Determining the position of the milling surface of the workpiece according to the acquired data to obtain a viewpoint during image acquisition; Controlling a triaxial movement mechanism by the digital twin system to drive a CCD camera to acquire milling surface images o