TWI632968B - Prediction method of electrical discharge machining accuracy - Google Patents

Abstract

A method for predicting the accuracy of electrical discharge machining includes: extracting a plurality of key machining characteristic values of an electrical discharge machining machine; providing the key machining characteristic values to an automated server to establish an electrical discharge machining accuracy prediction model; and sensing the number of electrical discharge machining machines in real time Set the discharge voltage signal, discharge current signal, and electrode consumption value predicted online as input values, and send it to the discharge machining accuracy prediction model (which contains the electrode consumption prediction model), and output the electrode consumption prediction value in real time and modify the discharge processing accuracy prediction model. ; And real-time sensing of multiple discharge voltage signals, discharge current signals, and electrode consumption values predicted online on the EDM machine as input values and transmitting to the revised EDM accuracy prediction model to output EDM accuracy prediction values in real time As output value.

Description

Prediction method of electrical discharge machining accuracy

The invention relates to a method for predicting the accuracy of electrical discharge machining, and more particularly to a method for predicting the accuracy of electrical discharge machining, which can predict the accuracy of electrical discharge machining.

When the EDM is performing EDM, the workpiece quality is not as good as expected due to poor slag discharge, abnormal short circuit or electrode consumption. Furthermore, the general electrical discharge machine cannot effectively predict the machining accuracy of the workpiece during the machining process, and can only perform offline machining accuracy measurement of the processed workpiece through a measurement device.

In order to solve the problem that the machining process cannot immediately detect the lack of machining accuracy, a prediction model is currently used to monitor the actual EDM sensing information (such as discharge voltage / current) during the EDM process of EDM to predict the workpiece. Processing accuracy. Historical sample data is often used to construct the first set of prediction models. Then, the prediction model is applied to the subsequent practical application environment.

For example, EDM measures the electrode consumption to accurately calculate the actual machining depth, and automatically performs iterative machining and inspection to ensure the precision of EDM machining of fine holes. However, the discharge electrode consumption measurement is performed offline, which greatly affects the productivity of processing equipment.

In view of this, there is a need to provide a method for predicting the accuracy of electrical discharge machining to solve the aforementioned problems.

The main object of the present invention is to provide an electrical discharge machining accuracy. The prediction method is based on establishing or modifying a prediction model of electrical discharge machining accuracy, which can predict the electrical discharge machining accuracy.

In order to achieve the above object, the present invention provides a method for predicting the accuracy of electrical discharge machining, which includes the following steps: extracting a plurality of key processing feature values of an electrical discharge machine using a modeling substrate extraction method, wherein the key processing feature values include an electrode Consumption value; provide an automatic server for these key processing characteristic values to build a discharge machining accuracy prediction model; modify the discharge processing machine with a set of discharge voltage signal, discharge current signal and offline measured electrode consumption value A plurality of key processing characteristic values, wherein the key processing characteristic values include the electrode consumption value; providing the corrected key processing characteristic values to the automated server to modify the EDM accuracy prediction model; and real-time sensing the Multiple sets of discharge voltage signals, discharge current signals, and electrode consumption values that provide offline measurement are used as input values for the EDM and sent to the revised EDM accuracy prediction model, with at least one EDM accuracy prediction value being output in real time as output value.

The present invention further provides a method for predicting the accuracy of electrical discharge machining, which includes the following steps: extracting a plurality of key processing characteristic values of an electrical discharge machine by a modeled substrate extraction method, wherein the key processing characteristic values include an electrode consumption value; The key processing characteristic values provide an automated server to establish a discharge machining accuracy prediction model; a plurality of sets of discharge voltage signals, discharge current signals, and online predicted electrode consumption values are used as input values and transmitted to the discharge machining accuracy prediction model (Which contains the electrode consumption prediction model), in order to output the electrode consumption prediction value in real time and modify a plurality of key processing characteristic values of the electrical discharge machine; provide the corrected key processing characteristic values to the automated server to correct the discharge Prediction model of machining accuracy; and real-time sensing of multiple sets of discharge voltage signal, discharge current signal and electrode consumption value provided online prediction as input values, and send to the revised discharge machining accuracy prediction model to output at least one discharge machining accuracy in real time The predicted value is used as the output value.

When using offline measured electrode consumption or online predicted electricity When the consumption value is extreme, the present invention mainly can establish or modify the prediction model to complete the prediction method of the accuracy of electrical discharge machining. Through the data pre-processing technology, find out the corresponding predicted or established model and its input and output values. Correspondence to achieve EDM accuracy prediction. Because the accuracy of the EDM process products is directly affected by the size and appearance of the instant discharge electrode, if it is measured offline, it will also affect the productivity of the EDM equipment. The method for predicting the accuracy of electrical discharge machining of the present invention can effectively solve the problem of insufficient accuracy prediction ability of the existing electrical discharge machining, so as to effectively improve the accuracy prediction ability.

In order to make the above and other objects, features, and advantages of the present invention more obvious, the following description will be described in detail with reference to the accompanying drawings.

200‧‧‧Prediction system for EDM accuracy

210‧‧‧EDM

211‧‧‧ Spindle

211a‧‧‧electrode

212‧‧‧controller

220‧‧‧Data Processing Module

221‧‧‧Control feedback unit

222a‧‧‧Image Sensor

222‧‧‧Sensing unit

223‧‧‧Processing unit

300‧‧‧Measuring machine

400‧‧‧ automation server

A1, A2‧‧‧ electrode consumption area

d1, d2‧‧‧ axial consumption value

P1‧‧‧Workpiece

S10 ~ S20‧‧‧‧steps

S110 ~ S150‧‧‧step

S210 ~ S220‧‧‧step

S310 ~ S330‧‧‧step

S410 ~ S430‧‧‧step

w1, w2‧‧‧ radial consumption value

FIG. 1 is a schematic structural diagram of a prediction system for electrical discharge machining accuracy according to an embodiment of the present invention; FIG. 2a is a block schematic diagram of a method for establishing a prediction model according to an embodiment of the present invention; and FIG. 2b is a diagram illustrating the establishment of an embodiment of the present invention Step flow chart of the method of predictive model; FIG. 2c is a flow chart of the method of modeling the substrate extraction method according to an embodiment of the present invention; FIG. 3 is a schematic diagram of the processing electrode characteristics of the electrical discharge machine of the present invention; A block diagram of the input and output of the EDM precision prediction model of one embodiment; FIG. 4b is a schematic flowchart of a method for predicting the accuracy of EDM machining according to an embodiment of the present invention.

FIG. 5 is a schematic block diagram of a method for correcting a first-type EDM machining accuracy prediction model according to another embodiment of the present invention; FIG. 6 is a flowchart showing steps of a first-type EDM machining accuracy prediction method according to another embodiment of the present invention; FIG. 7 is a schematic block diagram of a method for correcting a second type of electrical discharge machining accuracy prediction model according to another embodiment of the present invention; and FIG. 8 is a flowchart showing steps of a method of predicting a second type of electrical discharge machining accuracy according to another embodiment of the present invention schematic diagram.

Please refer to FIG. 1, which is a schematic structural diagram of a prediction system for electrical discharge machining accuracy according to an embodiment of the present invention. The prediction system 200 for the accuracy of electrical discharge machining in this embodiment includes an electrical discharge machining machine 210 and a data processing module 220 that are connected to the electrical discharge machining machine 210 by messages. The electric discharge machine 210 has a main shaft 211, and the main shaft 211 is provided with an electrode 211a, and a workpiece P1 can be processed through the electrode 211a. The data processing module 220 includes a control feedback unit 221, a sensing unit 222 and a processing unit 223. The control feedback unit 221 is mainly used to sense the distance signal as a judgment to determine whether the sensing unit 222 is to capture process data when the EDM 210 is processed on the line (the process data mainly includes the electrode consumption value and the discharge voltage signal And discharge current signals). The processing unit 223 can establish a plurality of processing features based on the process data captured by the sensing unit 222, and can also measure the measurement values of different measurement items processed by a measurement machine 300 (these measurements). Value refers to the measured value of the quality of the workpiece processed by the EDM machine), and then key processing feature values related to the processing accuracy are collected from the processing features to provide to an automated server 400 and establish an EDM An accuracy prediction model is used to predict the machining accuracy of the electric discharge machine 210. For example, the automated server 400 is a neural network-like (Neural Network, NN) method and regression analysis (Regression Analysis) method (such as partial least squares method PLS) to establish the accuracy prediction model of electrical discharge machining. For those who want to know, the prediction model used in this embodiment may be a processing accuracy prediction model established by an automated server as disclosed in Patent No. TW 1349867, but this is not intended to limit the present invention.

FIG. 2a shows a block diagram of a method for establishing a prediction model according to an embodiment of the present invention. First provide the discharge voltage signal, discharge current signal and electrode consumption value, then use the modeling substrate extraction method, and finally establish a prediction model. FIG. 2b shows a flowchart of steps in a method for establishing a prediction model according to an embodiment of the present invention. The method for establishing a prediction model includes the following steps: in step S10, a plurality of key processing feature values of an electrical discharge machine are extracted by a modeling substrate extraction method, wherein the key processing feature values include an electrode consumption value; and, In step S20, an automatic server is provided for the key processing characteristic values to establish a prediction model for the accuracy of EDM.

FIG. 2c shows a flowchart of steps in a method for extracting a modeled substrate according to an embodiment of the present invention. The modeling substrate extraction method of this embodiment mainly collects process data (such as electrode consumption value, discharge voltage signal, and discharge current signal) during the discharge process of the electrical discharge machine, and then establishes multiple processing characteristics based on these process data and applies Data Pre-Processing Technology extracts key processing feature values from these processing features that can be used to estimate processing accuracy. These aggregated key processing characteristic values can be mainly provided to the automated server to establish the EDM precision prediction model for predicting the machining accuracy of EDM.

Please refer to FIG. 2c and FIG. 1 again. The modeling substrate extraction method of this embodiment includes the following steps. First, step S110 is performed. According to a predetermined processing instruction and workpiece characteristics, the electrical discharge machining machine 210 respectively processes and processes a plurality of workpieces P1 to obtain a plurality of sets of process data. In one example, each set of process data mainly includes an electrode consumption value signal, a discharge voltage signal, and a discharge current signal. In an example, the process data is the processing of the EDM during the EDM process. Conditions (such as open circuit voltage and pulse on / off time, etc.) and machine status (such as actual discharge voltage waveform and discharge current waveform, etc.). In this embodiment, a high-voltage probe and a current meter are used to sense the discharge voltage and discharge current of the electric discharge machine 210. Furthermore, an image sensor 222a can be used to measure the electrode consumption value (shown in FIG. 1). Please refer to FIG. 3, the electrode consumption value may be the consumption degree of the head edge of an electrode 211a, for example, the axial consumption value d1, d2, the radial consumption value w1, w2 of the electrode 211a and the electrode consumption area A1 , A2.

After obtaining the process data of the EDM 210, step S120 is performed to establish a plurality of processing features by using the process data. In this embodiment, a data pre-processing technique can be applied to establish processing characteristics from the electrode consumption value, the discharge voltage signal, and the discharge current signal, and these processing characteristics can be used to estimate the processing accuracy. In this embodiment, the processing characteristics include electrode consumption value, discharge frequency, open circuit ratio, short circuit ratio, average short circuit time, short circuit time standard deviation, average short circuit current, short circuit current standard deviation, average delay time, delay time standard deviation, average Discharge peak current, peak current standard deviation, average discharge time, discharge time standard deviation, average discharge energy, and discharge energy standard deviation.

In processing characteristics, the average delay time and short circuit ratio are established from the discharge voltage signal. Among them, the average delay time is defined as the time difference from the time point when a sufficient open circuit voltage has been established until the voltage pulse passes through the gap between the electrode and the workpiece and the discharge current starts. Short circuit ratio is defined as the number of short circuit pulses (SCP) divided by the number of discharge pulses, where the short-circuit pulse is within a discharge pulse period when the open-circuit voltage value continues to be less than the specified voltage threshold. The discharge pulse period is recorded as a short-circuit pulse.

In the processing characteristics, the discharge frequency, the average discharge peak current, and the average discharge time are established from the discharge current signal. Among them, the average discharge frequency (Average spark frequency) is defined as within one pulse time, If the peak value of the current wave exceeds the minimum threshold peak value, it is defined as the occurrence of current sparks, and the discharge frequency is defined as the total number of sparks during the sampling period. Average peak discharge current is defined as the average of all discharge peak currents during the sampling period. Among them, the peak current is a large current value that reaches the workpiece through the electrode during the pulse period. The average discharge current pulse duration is defined as the average of all current pulse durations during the sampling period. The duration of the current pulse is the time difference between the start point and the end point of the discharge current waveform.

In the processing characteristics, the average short-circuit time, open circuit ratio, average discharge energy, and average short-circuit current are jointly established based on the discharge current signal and the discharge voltage signal. Among them, the average short-circuit time is related to the short-circuit duration, and the short-circuit duration is defined as when there are multiple continuous short-circuits in a discharge pulse period (more than two consecutive pulses are required), the short-circuit duration is multiple continuous During the short circuit period, the time difference between the first short circuit peak and the last short circuit peak. The open circuit ratio is defined as the number of open circuits divided by the total number of discharge pulses during the sampling period. Among them, in a certain pulse time, when the voltage peak ends and does not follow the current peak, it is called an open circuit. If an open circuit occurs, it means that a voltage peak (Ignition Voltage) fails to lead the subsequent current peak (Discharge Current), and this voltage peak is an invalid pulse. The average discharge energy is mainly used to maintain the stability of the EDM process to ensure the processing quality, and the discharge energy (E) formula of the i-th discharge is as follows (1): Where t _ei is the discharge duration, U _i is the discharge voltage, and I _pi is the discharge peak current. This formula assumes that the discharge voltage remains unchanged during the discharge process.

Those who want to know, according to the foregoing disclosure, the short-circuit time standard deviation, short-circuit current standard deviation, delay time standard deviation, peak current standard deviation, The calculation method of the standard deviation of the discharge time standard deviation and the standard deviation of the discharge energy is well known to those having ordinary knowledge in the technical field to which the present invention pertains, and therefore will not be repeated here.

Please refer to FIG. 2c again. After establishing a plurality of processing features, step S130 may be performed to obtain a plurality of sets of measured values of the measuring machine for measuring the workpieces, where each set of measured values is the electrical discharge machine respectively. Process the measured value of the quality of the workpiece. In this embodiment, a mold steel with a length, a width, and a height of 30 mm, 30 mm, and 10 mm is used as a workpiece, and an electrode with a diameter of 3 mm is used in combination. Wherein, in this embodiment, blind hole processing is taken as an example. A blind hole is formed in the workpiece by drilling the workpiece through an electrode, and each blind hole has an opening above the upper surface of the workpiece. The measurement items in this embodiment are the geometric and dimensional accuracy of each blind hole, which are the roughness of the bottom surface of the blind hole, the roundness of the upper opening, the size of the upper opening, the roundness of the bottom circle of the blind hole, and the blind hole. The size of the bottom circle, and each measurement item has a corresponding measurement value. The measuring machine 300 performs the measurement of the measurement item on the workpiece P1 to obtain a corresponding measurement value.

After obtaining the measured values, step S140 is then performed to perform a correlation analysis step to obtain multiple correlation values between the processing features and the measured values. In an embodiment, MATLAB tools can be used and the following relationship (2) can be used to observe the relationship between processing features and measured values: Among them: i is used to indicate the i-th workpiece, X is the processing feature, Represents the average value of the processing feature group, Y represents the measured value of different measurement items, Represents the average of the measurement values of different measurement items.

After the correlation value between the processing feature and the measured value is found, then step S150 is performed, and a processing feature with a larger correlation value is selected from the correlation values as a plurality of key processing feature values. Among them, this implementation An example is to select a processing feature whose absolute correlation value is greater than 0.3.

The following describes the key processing feature values related to the measurement values of different measurement items that are summarized by this embodiment. When the measured value is the measured value of the bottom surface roughness of the workpiece, the key processing characteristic values include the electrode consumption value, discharge frequency, short-circuit ratio, average short-circuit time, average short-circuit current, average delay time, delay time standard deviation, and average discharge. Peak current, average discharge time, average discharge energy, and standard deviation of discharge energy.

When the measured value is the measured value of the upper opening diameter of the blind hole of the workpiece, the key processing characteristic values include the electrode consumption value, the discharge frequency, the short-circuit ratio, the average short-circuit time, the average short-circuit current, the average delay time, and the standard deviation of the delay time , Average discharge peak current, average discharge time, average discharge energy and standard deviation of discharge energy. When the measured value is the measured roundness of the upper opening of the blind hole of the workpiece, the key processing characteristic values include the electrode consumption value, open circuit ratio, short circuit ratio, average short circuit time, average short circuit current, average discharge peak current, Standard deviation of peak current, average discharge time, standard deviation of discharge time, average discharge energy, and standard deviation of discharge energy.

When the measured value is the measured value of the diameter of the bottom circle of the blind hole of the workpiece, the key processing characteristic values include the electrode consumption value, open circuit ratio, short circuit ratio, average short circuit time, average short circuit current, average delay time, delay standard deviation , Average discharge peak current, average discharge time, average discharge energy and standard deviation of discharge energy. When the measurement value is the measurement of the true roundness of the bottom circle of the blind hole of the workpiece, the key processing characteristic values include electrode consumption value, discharge frequency, open circuit ratio, short circuit ratio, average delay time, standard deviation of delay time, average discharge Peak current, standard deviation of peak current, standard deviation of discharge time, average discharge energy, and standard deviation of discharge energy.

In this way, by extracting the extracted key processing feature values into a database of an automated server, a model base is established and an electric discharge machining accuracy prediction model training is performed, thereby predicting the machining accuracy.

For example, first use an automated server to measure the thickness of the bottom surface of the blind hole. Roughness measurement and its key processing characteristics are used to model and test. Wherein, in this embodiment, a neural network-like (NN) method and a regression analysis method (such as a partial least square method PLS) are used to establish an accuracy prediction model to predict the roughness of the bottom surface of the hole. The prediction results are shown in Table 1 below. From Table 1, we can observe that the mean absolute errors (MAE) of NN and PLS are 0.633 and 0.556 μm, and the 95% Max Errors of NN and PLS are 0.597 and 0.524 μm, respectively. It is less than 1.501 μm, which is less than twice the standard deviation of the real measurement (Real Y), which means that with the extracted key processing feature values, the roughness of the bottom surface of the hole can be predicted using these two models.

As shown in Table 2 below, when using an automated server to model and test the true roundness of the blind hole bottom circle and its key processing feature values, the MAEs of NN and PLS are 0.005 and 0.005mm, respectively. The 95% Max Error with PLS is 0.01 and 0.011mm, respectively, which is less than 0.015, which is twice the standard deviation of Real Y, representing the extracted key processing feature values. The model can be used to predict the true roundness of the bottom circle of the blind hole.

FIG. 4a shows a block diagram of the input and output of an electrical discharge machining accuracy prediction model according to an embodiment of the present invention. First provide the discharge voltage signal and discharge current signal, then use the established prediction model, and finally perform the discharge machining accuracy prediction. FIG. 4b is a schematic flowchart of a method for predicting the accuracy of electrical discharge machining according to an embodiment of the present invention. The method for predicting the accuracy of electrical discharge machining includes the following steps. In step S210, a prediction model of the accuracy of electrical discharge machining is provided. In step S220, a plurality of sets of discharge voltage signals and discharge current signals of the electric discharge machine are sensed as input values and transmitted to the electric discharge machining accuracy prediction model, and at least one electric discharge machining accuracy prediction value is output in real time as For example, the roughness of the bottom surface of the blind hole of the workpiece, the true roundness of the upper opening, the size of the upper opening, the true roundness of the bottom circle of the blind hole, and the predicted value of the size of the bottom circle of the blind hole.

FIG. 5 shows a block diagram of a method for correcting the first type of EDM accuracy prediction model according to another embodiment of the present invention. First provide the discharge voltage signal, the discharge current signal and the electrode consumption value measured offline, then modify the prediction model, and finally perform the discharge machining accuracy prediction. FIG. 6 is a schematic flowchart of steps in a method for predicting the accuracy of a first type of electrical discharge machining according to another embodiment of the present invention. The method for predicting the accuracy of electrical discharge machining further includes the following steps. In step S310, a plurality of key machining characteristic values of the electrical discharge machining machine are corrected with a set of discharge voltage signals, discharge current signals, and electrode consumption values measured offline. Please refer to FIG. 3 again, the electrode consumption value measured offline can be the consumption degree of the head edge of an electrode 211a, such as the axial consumption value d1, d2, the radial consumption value w1, w2 of the electrode 211a, and the electrode consumption area A1, A2. In step S320, the modified key processing feature values are provided to the automated server, and the discharge machining accuracy prediction model is modified by a neural network-like method and a regression analysis method. In step S330, a plurality of sets of discharge voltage signals, discharge current signals, and electrode consumption values provided for offline measurement of the electric discharge machine are sensed in real time as input values, and are transmitted to the modified discharge machining accuracy prediction model for immediate output. At least one predicted value of EDM accuracy as an output value, such as a blind hole in a workpiece The predicted value of the roughness of the bottom surface of the hole, the true roundness of the upper opening, the size of the upper opening, the true roundness of the bottom circle of the blind hole, and the size of the bottom circle of the blind hole. In this embodiment, the present invention uses the electrode consumption values measured offline to modify the discharge machining accuracy prediction model, and then the discharge voltage signal, discharge current signal, and electrode consumption values measured offline are input to the corrected discharge machining accuracy prediction. model. The invention is assisted by the analysis of the electrode consumption image, which can integrate offline measurement of the electrode consumption image, and the correction prediction model can improve the original electric discharge machining accuracy prediction ability.

FIG. 7 shows a block diagram of a method for correcting a second type of EDM accuracy prediction model according to another embodiment of the present invention. First provide the discharge voltage signal, the discharge current signal and the electrode consumption value predicted online, then modify the prediction model, and finally perform the discharge machining accuracy prediction. FIG. 8 is a schematic flowchart showing the steps of a second type of EDM accuracy prediction method according to another embodiment of the present invention. The method for predicting electrical discharge machining accuracy further includes the following steps: In step S410, a plurality of sets of discharge voltage signals, electrical discharge current signals, and electrode consumption values predicted online are transmitted as input values to the electrical discharge machining accuracy prediction model (which includes electrodes). Consumption prediction model), in order to output the electrode consumption values predicted on the line in real time and modify multiple key processing feature values of the EDM (such as using a fit regularization classification method or using a Glmnet regression model). Please refer to FIG. 3 again, the predicted electrode consumption value on the line may be the predicted value of the head edge consumption of an electrode 211a, such as the axial consumption d1, d2, the radial consumption w1, w2 and Electrode consumption area A1, A2. In step S420, the corrected key processing feature values are provided to the automated server, and the discharge machining accuracy prediction model is modified by a neural network-like method and a regression analysis method. In step S430, a plurality of sets of discharge voltage signals, discharge current signals, and electrode consumption values provided online for prediction are input as input values, and transmitted to the modified discharge machining accuracy prediction model to output at least one discharge machining accuracy prediction in real time. The value is used as an output value, such as the predicted value of the roughness of the bottom surface of the blind hole of the workpiece, the true roundness of the upper opening, the size of the upper opening, the true roundness of the bottom circle of the blind hole, and the size of the bottom circle of the blind hole. In this embodiment, the present invention uses The electrode consumption value predicted online was used to modify the discharge machining accuracy prediction model, and then the discharge voltage signal, discharge current signal, and the electrode consumption value predicted online were input to the revised discharge machining accuracy prediction model. The online prediction of the electrode consumption value in the present invention can greatly reduce the offline measurement time of the electrode, so as to increase the productivity of the electrical discharge machine and improve the overall processing prediction accuracy.

When the electrode consumption value measured offline or the electrode consumption value predicted online is used, the present invention mainly can establish or modify a prediction model to complete the prediction method of electrical discharge machining accuracy. Through data pre-processing technology, find out the corresponding The revised prediction model and the correspondence between its input and output values are used to achieve the accuracy prediction of EDM. Because the accuracy of the EDM process products is directly affected by the size and appearance of the instant discharge electrode, if it is measured offline, it will also affect the productivity of the EDM equipment. The method for predicting the accuracy of electrical discharge machining of the present invention can effectively solve the problem of insufficient accuracy prediction ability of the existing electrical discharge machining, so as to effectively improve the accuracy prediction ability.

Please refer to FIG. 1 again. In yet another embodiment, the control feedback unit 221 may be a laser ranging device and is disposed on the main shaft 211. Thereby, the sensing unit 222 can select whether to acquire the process data of the electric discharge machine 210 (ie, the electrode consumption value, the discharge voltage signal and the discharge current signal) by sensing the position of the electrode 211a through the laser ranging device. For example, the laser ranging device is used to determine whether the moving direction of the main shaft 211 is the feeding direction. If the determination result is yes, it means that the main shaft 211 is being processed. Therefore, the sensing unit 222 can further obtain the manufacturing process during processing. data. Conversely, if the determination result is no, the sensing unit 222 does not need to retrieve process data. It can be known that the required data amount captured by the sensing unit 222 can be effectively filtered by controlling the feedback unit 221. In other embodiments, the control feedback unit 221 may be an optical ruler. Alternatively, the sensing unit 222 can also directly obtain the coordinate position of the main shaft 211 from the controller 212 of the electric discharge machining machine 210 to determine whether to retrieve relevant process data, and the same can be achieved by the screening sensing unit 222. Purpose of data volume.

In summary, it only describes the implementation or examples of the technical means adopted by the present invention to solve the problem, and is not intended to limit the scope of patent implementation of the present invention. That is, all changes and modifications that are consistent with the meaning of the scope of patent application of the present invention, or made according to the scope of patent of the present invention, are covered by the scope of patent of the present invention.

Claims (10)

A method for predicting the accuracy of electrical discharge machining includes the following steps: extracting a plurality of key processing feature values of an electrical discharge machine using a modeling substrate extraction method, wherein the key processing feature values include an electrode consumption value; The characteristic value provides an automated server to establish a prediction model of electrical discharge machining accuracy; and a plurality of sets of electrical discharge voltage signals and electrical discharge current signals of the electrical discharge machine are sensed in real time as input values and transmitted to the established electrical discharge machining accuracy prediction model , Taking at least one predicted value of EDM accuracy as an output value in real time. The method for predicting the accuracy of electrical discharge machining as described in item 1 of the scope of the patent application, further includes the following steps: Correcting multiple keys of the electrical discharge machine with a set of discharge voltage signal, discharge current signal, and electrode consumption value measured offline Processing characteristic values; providing the key processing characteristic values after correction to the automated server to modify the prediction model of the discharge machining accuracy; and real-time sensing of multiple sets of discharge voltage signals, discharge current signals of the discharge processing machine, and providing offline The measured electrode consumption value is used as an input value, and is transmitted to the revised prediction model of electrical discharge machining accuracy, and at least one predicted value of electrical discharge machining accuracy is immediately output as an output value. The method for predicting the accuracy of electrical discharge machining as described in item 1 of the scope of the patent application, further includes the following steps: taking multiple sets of discharge voltage signals, electrical discharge current signals, and electrode consumption values predicted online as input values, and transmitting them to the electrical discharge machining accuracy prediction Model, the electrical discharge machining accuracy prediction model includes an electrode consumption prediction model to output the electrode consumption values predicted on the line in real time and modify a plurality of key machining characteristic values of the electric discharge machining machine; and provide the corrected key machining characteristic values to the An automated server to modify the prediction model of discharge machining accuracy; and to sense multiple sets of discharge voltage signals, discharge current signals, and electrode consumption values that provide online predictions as input values, and send them to the revised discharge machining accuracy prediction model, Take at least one predicted value of EDM accuracy as an output value. The method for predicting the accuracy of electrical discharge machining according to item 1 of the scope of patent application, wherein the step of extracting a plurality of key machining characteristic values of an electrical discharge machining machine includes: using the electrical discharge machining machine to separately process multiple workpieces to obtain multiple sets of processes Data, where each set of process data includes an electrode consumption value, a discharge voltage signal and a discharge current signal; use the set of process data to establish a plurality of processing features; obtain a measuring machine to measure a plurality of the workpieces Measured values, each of which is a measured value of the workpiece processed by the electrical discharge machine according to the process data; a correlation analysis step is performed to obtain the processing characteristics and the measurements A plurality of correlation values among the values; and a processing feature with a large correlation value is selected from the correlation values as the key processing feature values. The method for predicting the accuracy of electrical discharge machining as described in item 4 of the scope of the patent application, wherein the step of selecting processing features with a large correlation value among the correlation values is to select processing features with absolute correlation values greater than 0.3. The method for predicting discharge machining accuracy as described in item 4 of the scope of patent application, wherein the machining characteristics include the electrode consumption value, a discharge frequency, an open circuit ratio, a short circuit ratio, an average short circuit time, and a short circuit time standard deviation. , An average short-circuit current, a short-circuit current standard deviation, an average delay time, a delay time standard deviation, an average discharge peak current, a peak current standard deviation, an average discharge time, a discharge time standard deviation, an average discharge energy And a standard deviation of discharge energy. The method for predicting the accuracy of electrical discharge machining as described in item 6 of the scope of the patent application, wherein the measurement value is a roughness measurement value of the workpieces, and the key processing characteristic values include the electrode consumption value, a discharge frequency, A short-circuit ratio, an average short-circuit time, an average short-circuit current, an average delay time, a delay time standard deviation, an average discharge peak current, an average discharge time, an average discharge energy, and a discharge energy standard deviation. The method for predicting the accuracy of electrical discharge machining as described in item 6 of the scope of patent application, wherein the electrical discharge machining machine performs a drilling step on each of the workpieces, and forms a blind hole on each of the workpieces, where each Each of the blind holes has an upper opening on one of the upper surfaces of each of the workpieces and a bottom circle of the blind hole; wherein the measurement value is a measurement value of the size of the upper opening, and the key processing characteristic values include the Electrode consumption value, a discharge frequency, a short-circuit ratio, an average short-circuit time, an average short-circuit current, an average delay time, a delay time standard deviation, an average discharge peak current, an average discharge time, an average discharge energy, and Standard deviation of discharge energy. The method for predicting the accuracy of electrical discharge machining as described in item 6 of the scope of patent application, wherein the electrical discharge machining machine performs a drilling step on each of the workpieces, and forms a blind hole on each of the workpieces, where each Each of the blind holes has an opening on one of the upper surfaces of each of the workpieces and a bottom circle of the blind hole; wherein the measured value is a measurement of the size of the bottom circle of the blind hole, and the key processing characteristic values Including the electrode consumption value, an open circuit ratio, a short circuit ratio, an average short circuit time, an average short circuit current, an average delay time, a delay time standard deviation, an average discharge peak current, an average discharge time, and an average discharge energy And a standard deviation of discharge energy. The method for predicting the accuracy of electrical discharge machining according to item 1 of the scope of the patent application, wherein the electrode consumption value is an axial consumption value, a radial consumption value and an electrode consumption area of a head edge of an electrode.