CN106514434B - A kind of milling cutter wear monitoring method based on data - Google Patents

Abstract

The invention belongs to the technical field related to tool wear detection, and discloses a data-based milling tool wear monitoring method, which includes the following steps: (1) collecting three-phase output current signals of a spindle drive motor of a CNC machine tool when the milling tool is working; (2) cleaning the collected three-phase output current signals; (3) extracting features representing the wear of the milling tool from the three-phase output current signals after self-cleaning using the compressed sensing method and key point theory (4) According to the three-phase output current signal of the spindle drive motor collected in real time during normal processing of a certain milling tool, the characteristic signal index is calculated online, and then the wear of the milling tool is monitored online in real time. The data-based milling tool wear monitoring method provided by the invention reduces the cost and realizes real-time monitoring of the milling tool wear.

Description

A data-based monitoring method for milling tool wear

technical field

The invention belongs to the technical field related to tool wear detection, and more specifically relates to a data-based milling tool wear monitoring method.

Background technique

At present, with the development of computer and automation technology, the field of mechanical processing is developing towards the direction of intelligent manufacturing. The most basic requirement of the manufacturing and processing system is that the processing system can automatically and effectively monitor and adjust the faults that occur during the production and processing process. The wear of the basic element tool in the machining process will cause the vibration of the machine tool, the decrease of the surface quality of the workpiece and the accuracy of the machining dimension. Therefore, the research on the feature extraction of tool wear is of great significance for the monitoring of tool conditions.

Over the past few decades, machining condition monitoring has been extensively studied, especially for issues such as tool wear, workpiece deformation, and chatter. However, for thin-walled parts with large spans or high height-to-thickness ratios, there is still no mature monitoring method because of large curvature changes and easy deformation during processing, resulting in changes in cutting force and affecting machining accuracy. Processing state identification is a multi-factor and non-linear problem. The comprehensive consideration of multiple factors constitutes different processing states and normal or abnormal physical signal characteristics under different parameters.

The commonly used methods of monitoring tool wear can be divided into direct measurement method and indirect measurement method. The direct measurement method is to directly measure the average wear amount of the middle part of the wear zone on the knife face. Most of the tool wear detection methods are based on the relevant characteristics of the tool volume loss. The wear value of the tool is directly obtained through contact measurement or CCD imaging. This method is susceptible to Influenced by the processing environment, it is inconvenient to carry out online measurement in real-time processing. The indirect measurement method realizes indirect measurement by measuring physical quantities related to tool wear, such as cutting force, acoustic emission signal, etc., and establishing the corresponding relationship between tool wear and these quantities. In actual monitoring, due to the interference of vibration and measurement noise, it is easy to make mistakes in judging the wear of the tool by indirect measurement, resulting in misjudgment, and because the boundary between normal wear and abnormal wear of the tool has certain uncertainty, it is determined in advance. Thresholds are more difficult. For example, the Chinese patent application number 201310442967.0 discloses a tool wear detection method, which collects acoustic emission signals of various wear states, current signals in the machine tool spindle, cutting speed, cutting depth and feed rate as condition attributes, and establishes Decision table, the BP neural network is trained and learned through the genetic algorithm, and then the trained neural network is used to predict the degree of tool wear. However, there are still some shortcomings, such as the need to obtain some inconvenient acoustic emission signals, etc., the sensor layout is troublesome, the data calculation is complicated, and the cost is high, which is not conducive to popularization and use.

Contents of the invention

In view of the above defects or improvement needs of the prior art, the present invention provides a data-based milling tool wear monitoring method, which applies the compressive sensing method and key point theory to the tool wear monitoring in CNC machine tool processing, and extracts the tool The key characteristics of wear to realize real-time online prediction of tool wear; the compressive sensing method and key point theoretical algorithm adopted in the milling tool wear monitoring method are simple and easy to implement and are suitable for online calculation, which can realize real-time prediction of tool life in industrial sites, Real-time prediction of tool wear, timely replacement of tools with serious damage, saving costs; and the milling tool wear monitoring method only collects the three-phase output current signal of the spindle drive motor of the CNC machine tool during the CNC machining process, and arranges the Hall The current sensor can conveniently collect the current signal of the spindle drive motor. The arrangement of the sensor does not affect the normal processing process of the machine tool, does not change the physical structure of the machine tool itself, and is easy to implement.

To achieve the above object, the present invention provides a data-based milling tool wear monitoring method, which includes the following steps:

(1) Collect the three-phase output current signal of the spindle drive motor of the CNC machine tool when the milling tool is working;

(2) cleaning the collected three-phase output current signal;

(3) extract the characteristic coefficient characterizing the wear of the milling tool from the three-phase output current signal after self-cleaning by adopting the compressive sensing method and the key point theory;

(4) Calculate the characteristic signal index online according to the three-phase output current signal of the spindle drive motor collected in real time during normal processing of a certain milling tool, and then perform real-time online monitoring of the wear of the milling tool.

Further, the three-phase output current signals are collected by Hall current sensors.

Further, the purpose of cleaning the three-phase output current signals is to remove pathological and redundant data in the three-phase output current signals, thereby providing reliable data for the extraction of milling tool wear characteristics.

Further, the characteristic signal index reflects the relationship between the three-phase output current signal of the spindle drive motor and the corresponding wear degree of the milling tool.

Further, the update formula of the characteristic signal index is:

Γ _k ＝D(S _k -S _baseline )

In the formula, D is the Euclidean distance between two vectors, S _k is the characteristic coefficient extracted by the compressive sensing method when the current signal data is collected for the kth time, and S _baseline represents the characteristic coefficient of the basic current signal data (when there is no fault).

Generally speaking, compared with the prior art, the above technical solutions conceived by the present invention provide a data-based milling tool wear monitoring method, which applies the compressive sensing method and key point theory to the tool in CNC machine tool processing Wear monitoring, from which the key features of tool wear are extracted to realize real-time online prediction of tool wear; the compressive sensing method and key point theoretical algorithm adopted in the milling tool wear monitoring method are simple and easy to implement and are suitable for online calculation, which can realize industrial Real-time prediction of tool life on site, real-time prediction of tool wear, timely replacement of tools for severely damaged tools, and cost savings; and the milling tool wear monitoring method only collects three-phase data of the spindle drive motor of the CNC machine tool during the CNC machining process. The current signal of the spindle drive motor can be collected conveniently by arranging the Hall current sensor to output the current signal. The arrangement of the sensor does not affect the normal processing process of the machine tool and does not change the physical structure of the machine tool itself, which is easy to realize.

Description of drawings

Fig. 1 is a flowchart of a data-based milling tool wear monitoring method provided by a preferred embodiment of the present invention.

FIG. 2 is a waveform diagram of the three-phase output current of the spindle drive motor involved in the data-based milling tool wear monitoring method in FIG. 1 .

FIG. 3 is a current waveform diagram of the corresponding CNC machine tool spindle under different wear conditions involved in the data-based milling tool wear monitoring method in FIG. 1 .

Fig. 4 is a waveform diagram of the variation of the spindle current when the CNC machine tool automatically switches the tool involved in the data-based milling tool wear monitoring method in Fig. 1 .

Fig. 5 is a waveform diagram of 400 curves obtained during the machining of the same tool involved in the data-based milling tool wear monitoring method in Fig. 1 .

Detailed ways

In order to make the object, technical solution and advantages of the present invention clearer, the present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described here are only used to explain the present invention, not to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not constitute a conflict with each other.

Please refer to Figure 1, the data-based milling tool wear monitoring method provided by the preferred embodiment of the present invention, the milling tool wear monitoring method can greatly improve the reliability and production efficiency of processing equipment, and fully explore the value of intelligent manufacturing big data The chain provides support for future smart manufacturing. The milling tool wear monitoring method collects the spindle current signal in milling tool processing, then cleans the collected current signal, analyzes the causality and correlation of the cleaned current signal, finds out some hidden physical laws, extracts It can characterize the characteristic coefficient of tool wear, analyze the current signal collected in real time, calculate the characteristic signal index of tool wear, realize the real-time monitoring of tool wear, replace the tool that is about to be damaged in advance, and reduce the production cost.

In the present embodiment, the described data-based milling tool wear monitoring method mainly includes the following steps:

Step 1, collecting the three-phase output current signal of the spindle drive motor of the corresponding CNC machine tool when the milling tool is working. In this embodiment, the monitoring method of milling tool wear based on data is described by taking the monitoring of milling tool wear when CNC machine tools process engine blades as an example. In this embodiment, the Hall current sensor is used to collect the three-phase output current signal of the spindle drive motor of the CNC machine tool when the blade is processed under different wear degrees of the milling tool; it can be understood that in other embodiments, other types of current signals can also be used. Current sensors, such as Rogowski current sensors; a large number of repeated experiments, collecting the curves of the three-phase output current signal of the spindle drive motor changing with time for a large number of tools under different wear conditions and the three-phase output of the spindle drive motor when the tool is switched The change curve of current is shown in Figure 2 and Figure 4. Figure 3 is the waveform diagram of the spindle current under different wear conditions of the tool.

Step 2, cleaning the collected three-phase output current signals to ensure the reliability of the three-phase output current signals. Specifically, the collected three-phase output current signals are removed through a data cleaning model to remove redundant, pathological, and noisy data, so as to provide a reliable data source for subsequent extraction of milling tool wear characteristics. In order to facilitate the analysis and processing, 400 curves of the spindle current signal corresponding to the same tool are superimposed together to obtain the waveform shown in Figure 5.

Step 3, using the compressed sensing method and the key point theory to extract the characteristic coefficient representing the wear of the milling tool from the three-phase output current signal after cleaning.

Step 4: Calculate the characteristic signal index online according to the three-phase output current signal of the spindle drive motor collected in real time during normal processing of a milling tool, and then monitor the wear of the milling tool online.

Specifically, according to the three-phase output current signal of the spindle drive motor collected in real time when a certain milling tool is processing engine blades (as shown in the middle section of Figure 4), the characteristic signal index Γ is calculated online, and then the tool wear is monitored online , Replace the severely worn tools in advance to save processing costs. Wherein the characteristic signal index Γ accurately reflects the correlation between the three-phase output current signal of the spindle drive motor and the corresponding milling tool wear degree, and the update formula of the characteristic signal index Γ is:

Γ _k ＝D(S _k -S _baseline )

In the formula, D is the Euclidean distance between two vectors, S _k is the characteristic coefficient extracted by the compressive sensing method when the current signal data is collected for the kth time, and S _baseline represents the characteristic coefficient of the basic current signal data (when there is no fault).

The data-based tool wear detection method of the present invention is to apply the idea of machine learning in the mechanical field, and to extract the milling tool wear from the three-phase output current signal after self-cleaning using the compressed sensing method and the key point theory. characteristic coefficient.

The data-based milling tool wear monitoring method provided by the present invention applies the compressive sensing method and key point theory to the tool wear monitoring in CNC machine tool processing, and extracts the key features of tool wear from it to realize real-time online prediction of tool wear ; The compressive sensing method and the key point theoretical algorithm adopted by the milling tool wear monitoring method are simple and easy to implement and are suitable for online calculation, which can realize real-time prediction of tool life on the industrial site, real-time prediction of tool wear, and timely detection of severely damaged tools. Replace the cutting tool to save cost; and the milling tool wear monitoring method only collects the three-phase output current signal of the spindle drive motor of the CNC machine tool in the CNC machining process, and the current signal of the spindle drive motor can be collected conveniently by arranging the Hall current sensor , the layout of the sensor does not affect the normal processing process of the machine tool, does not change the physical structure of the machine tool itself, and is easy to implement.

It is easy for those skilled in the art to understand that the above descriptions are only preferred embodiments of the present invention, and are not intended to limit the present invention. Any modifications, equivalent replacements and improvements made within the spirit and principles of the present invention, All should be included within the protection scope of the present invention.

Claims (3)

1. a kind of milling cutter wear monitoring method based on data, it comprises the following steps：

(1) the three-phase output current signal of the spindle drive motor of corresponding numerically-controlled machine tool when gathering milling cutter work；

(2) the three-phase output current signal collected is cleaned；

(3) use in the three-phase output current signal after compression sensing method and key point theory self-cleaning and extract characterization The characteristic coefficient of the milling cutter abrasion；

(4) based on the following formula in line computation and renewal characteristic signal exponential gamma_k, this feature signal exponential gamma_kIt is described for reflecting Dependency relation between the three-phase output current signal of spindle drive motor and the corresponding milling cutter degree of wear, and then to milling Cutting knife tool abrasion carries out real time on-line monitoring：

In above formula, D is two vectorial Euclidean distances, S_kIt is that kth time uses step (3) when gathering the three-phase output current signal The characteristic coefficient that the mode extracts, S_baselineCurrent signal data when representing base current signal data namely not being out of order Characteristic coefficient.

2. the milling cutter wear monitoring method based on data as claimed in claim 1, it is characterised in that：The three-phase output Current signal is collected using Hall current sensor.

3. the milling cutter wear monitoring method based on data as claimed in claim 1, it is characterised in that：It is defeated to the three-phase It is to remove in the three-phase output current signal morbid state, the data of redundancy to go out current signal to carry out cleaning, and then is milling The extraction of tool wear feature provides reliable data.