Electrical discharge machining apparatus and electrical discharge machining method

Abstract

An electrical discharge machining device (1A) is provided with: a power supply device (20) that supplies power to a machining gap between a tool electrode (4) and a workpiece (3); a current detection unit (9) that detects an inter-electrode current value, which is the current value of the current flowing from the power supply device to the machining inter-electrode; a voltage command detection unit (12) that detects an applied voltage value, which is a command value for the voltage value of the voltage applied between the machining electrodes; and a calculation device (10A) that acquires the inter-electrode current value and the applied voltage value, and estimates the electrical characteristics from the inter-electrode current value and the applied voltage value using a learning model for estimating the electrical characteristics inherent to the device from the inter-electrode current value and the applied voltage value. A correction value for correcting the inter-electrode current value and the applied voltage value is determined such that the estimated electrical characteristic is a reference electrical characteristic.

Description

Technical Field

[0001] This invention relates to an electrical discharge machining apparatus and a method for machining workpieces using electrical discharge energy. Background Technology

[0002] For an electrical discharge machining (EDM) apparatus that generates a discharge between the machining electrode formed by the tool electrode and the workpiece, and uses the discharge energy to machine the workpiece, it is desirable to achieve stable machining corresponding to the EDM apparatus.

[0003] The processing condition adjustment device described in Patent Document 1 learns the adjustment actions of processing conditions (voltage waveform, current waveform, etc.) corresponding to the processing state (processing voltage, processing current, processing speed, etc.), and predicts the adjustment actions of processing conditions based on the processing state when a predetermined processing has been performed.

[0004] Patent Document 1: International Publication No. 2022 / 210472 Summary of the Invention

[0005] However, in the technology of the aforementioned Patent Document 1, the common adjustment actions of multiple electrical discharge machining devices are learned, which has the following problem: the machining error caused by the inherent electrical characteristics of the electrical discharge machining device cannot be corrected, resulting in a decrease in machining quality.

[0006] The present invention was made in view of the above circumstances, and its object is to provide an electrical discharge machining apparatus that can suppress the degradation of machining quality.

[0007] To solve the above-mentioned problems and achieve the objective, the electrical discharge machining apparatus of the present invention includes: a power supply device that supplies power to the machining electrode space between the tool electrode and the workpiece; a current detection unit that detects the current value of the current flowing from the power supply device to the machining electrode space, i.e., the electrode space current value; a voltage command detection unit that detects the command value of the voltage applied to the machining electrode space, i.e., the applied voltage value; and a calculation unit that acquires the electrode space current value and the applied voltage value, uses a learning model for inferring the inherent electrical characteristics of the apparatus based on the electrode space current value and the applied voltage value, infers the electrical characteristics based on the electrode space current value and the applied voltage value, and determines a correction value for correcting the electrode space current value and the applied voltage value so that the inferred electrical characteristics become a reference electrical characteristic, i.e., a reference characteristic.

[0008] The effects of the invention

[0009] The electrical discharge machining apparatus of the present invention achieves the effect of suppressing the degradation of machining quality. Attached Figure Description

[0010] Figure 1 This is a diagram showing the structure of the electrical discharge machining apparatus according to Embodiment 1.

[0011] Figure 2 This is a diagram showing the structure of the learning device included in the electrical discharge machining apparatus according to Embodiment 1.

[0012] Figure 3 This is a diagram illustrating an example of the structure of the neural network used in the learning device of Embodiment 1.

[0013] Figure 4 This is a flowchart illustrating the processing flow of the electrical discharge machining apparatus according to Embodiment 1 when correcting electrical characteristics.

[0014] Figure 5 This is a diagram showing an example of the waveform of the inter-electrode current value obtained by the electrical discharge machining apparatus according to Embodiment 1.

[0015] Figure 6 This diagram illustrates the process by which the electrical discharge machining apparatus of Embodiment 1 infers impedance and parasitic capacitance values.

[0016] Figure 7 This is a diagram illustrating the range that can be corrected by the electrical discharge machining apparatus according to Embodiment 1.

[0017] Figure 8 This is a diagram showing the structure of the electrical discharge machining apparatus according to Embodiment 2.

[0018] Figure 9 This is a diagram showing the processing circuit of the processing section in the case where the processing electrodes of the electrical discharge machining apparatus according to Embodiment 2 are in a separated state.

[0019] Figure 10 It means and Figure 9 The diagram shows the equivalent circuit corresponding to the processing circuit.

[0020] Figure 11 This is a diagram used to illustrate the impedance and phase delay calculated by the electrical discharge machining apparatus according to Embodiment 2 when the machining electrodes are in a separated state.

[0021] Figure 12 This is a graph showing the frequency characteristics calculated by the electrical discharge machining apparatus according to Embodiment 2 when the machining electrodes are in a separated state.

[0022] Figure 13 This is a diagram showing the processing circuit of the processing section in the case where the processing electrodes of the electrical discharge machining apparatus according to Embodiment 2 are in a short-circuit state.

[0023] Figure 14 It means and Figure 13 The diagram shows the equivalent circuit corresponding to the processing circuit.

[0024] Figure 15 This is a diagram used to illustrate the impedance and phase delay calculated by the electrical discharge machining apparatus according to Embodiment 2 when the machining electrodes are in a short-circuit state.

[0025] Figure 16 This is a graph showing the frequency characteristics calculated by the electrical discharge machining apparatus according to Embodiment 2 when the machining electrodes are in a short-circuit state.

[0026] Figure 17 This is a flowchart illustrating the processing flow of the electrical characteristics correction performed by the electrical discharge machining apparatus according to Embodiment 2.

[0027] Figure 18 This is a diagram illustrating an example of the display screen showing the image when the electrical characteristics of the electrical discharge machining apparatus according to Embodiment 3 are deviated.

[0028] Figure 19 This is a diagram illustrating an example of a display screen shown by the electrical discharge machining apparatus according to Embodiment 3 after correcting for deviations in electrical characteristics.

[0029] Figure 20 This is a diagram illustrating an example of the hardware structure of the computing device involved in Implementation Method 1. Detailed Implementation

[0030] Hereinafter, the electrical discharge machining apparatus and electrical discharge machining method according to embodiments of the present invention will be described in detail with reference to the accompanying drawings.

[0031] Implementation Method 1

[0032] Figure 1 This is a diagram showing the structure of the electrical discharge machining apparatus according to Embodiment 1. The electrical discharge machining apparatus 1A applies a voltage pulse to the machining electrode space formed by the tool electrode (machining electrode) 4 and the workpiece 3, bringing the tool electrode 4 and the workpiece 3 close together, thereby generating a discharge between the machining electrodes, and machining the workpiece 3 by the heat of the generated electric arc.

[0033] The electrical discharge machining apparatus 1A includes a machining section 2A, a power supply unit 20, and a computing unit 10A. The power supply unit 20 is a machining power source connected in series with the machining electrode between the tool electrode 4 and the workpiece 3, supplying power to the machining electrode. The power supply unit 20 supplies power to the machining electrode by outputting an AC or DC voltage to the machining electrode.

[0034] The machining unit 2A includes a tool electrode 4 and a current detection unit 9. The machining unit 2A uses electricity supplied from the power supply unit 20 to perform electrical discharge machining on the workpiece 3. The current detection unit 9 is connected in series between the power supply unit 20 and the machining electrode, and measures the current value (hereinafter sometimes referred to as the inter-electrode current value) flowing from the power supply unit 20 through the inter-electrode. The current detection unit 9 sends the measured inter-electrode current value to the arithmetic unit 10A. The arithmetic unit 10A is a computer that performs calculations on the information used during electrical discharge machining.

[0035] The power supply unit 20 is connected to the machining electrodes via a cable (not shown). This cable has electrical characteristics (electrical load) such as parasitic inductance PX1, parasitic capacitance (described later), and impedance in its electrical circuit. The parasitic inductance PX1, parasitic capacitance, and impedance significantly affect the current flowing between the machining electrodes. The impedance is a value calculated based on the parasitic inductance PX1 and parasitic capacitance. In Embodiment 1, the electrical discharge machining apparatus 1A suppresses the degradation of machining quality by correcting the impedance and parasitic capacitance values.

[0036] Even if the length, thickness, and number of cables connecting the power supply device 20 to the processing electrodes are fixed, the electrical characteristics will change due to mechanical characteristics such as the spatial path, bending state, and contact with the connection. Therefore, processing characteristics such as the surface roughness of the processed surface of the workpiece 3 and the processed dimensions of the workpiece 3 will vary depending on the cable condition. For example, even if the EDM apparatus 1A is the same model, if the mechanical characteristics of the cable change due to individual differences, the current flowing between the electrodes will change, and sometimes the processing result will not fall within the expected dimensional tolerances and surface roughness tolerances. Therefore, in Embodiment 1, in order to correct the electrical characteristics (impedance and parasitic capacitance values ​​in Embodiment 1) themselves, the EDM apparatus 1A corrects the applied voltage value and the inter-electrode current value, which will be described later.

[0037] The arithmetic unit 10A is connected to the processing unit 2A, the power supply unit 20, and the display 25. The connection lines between the arithmetic unit 10A and the processing unit 2A are omitted from the diagram. The arithmetic unit 10A includes a control unit 11, a voltage command detection unit 12, an arithmetic unit 13, a learning device 14, and a correction value determination unit 16. The control unit 11 controls the processing unit 2A, the power supply unit 20, and the display 25.

[0038] The voltage command detection unit 12 detects the command value (hereinafter sometimes referred to as the applied voltage value) of the voltage applied between the processing electrodes. The voltage command detection unit 12 detects the applied voltage value based on the voltage command output by the control unit 11 to the power supply device 20. The voltage command detection unit 12 sends the detected applied voltage value to the learning device 14.

[0039] The arithmetic unit 13 performs calculations on various information. The display 25 displays information from the electrical discharge machining apparatus 1A according to instructions received from the apparatus. The display 25 shows the status of the electrical discharge machining apparatus 1A, machining conditions, etc.

[0040] The learning device 14 performs machine learning on the relationship between the inter-electrode current value and the measured impedance in the discharge circuit during learning, and infers the impedance corresponding to the inter-electrode current value in the discharge circuit during inference.

[0041] Furthermore, the learning device 14 calculates the amount of electricity flowing between the processing electrodes by integrating the inter-electrode current value over time during learning. During learning, the learning device 14 performs machine learning on the relationship between the amount of electricity, the applied voltage value, and the parasitic capacitance multiplied by a constant (hereinafter referred to as constant α). Hereinafter, the parasitic capacitance multiplied by the constant α is sometimes referred to as the parasitic capacitance value αC. During inference, the learning device 14 infers the parasitic capacitance value αC corresponding to the amount of electricity and the applied voltage value.

[0042] The parasitic capacitance value αC is an inherent value of the electrical discharge machining apparatus 1A. In Embodiment 1, the electrical discharge machining apparatus 1A learns the parasitic capacitance value αC, thereby inferring its own inherent parasitic capacitance value αC. That is, the electrical discharge machining apparatus 1A learns and infers its own inherent parasitic capacitance value αC for each electrical discharge machining apparatus 1A.

[0043] Generally, if the electrical characteristics of a discharge machining apparatus, such as impedance, change, the current flowing between the machining electrodes during discharge will change. Furthermore, the machining accuracy, such as the surface roughness of the machined surface of the workpiece 3 and the dimensions of the workpiece 3 after machining, is affected by the current flowing between the machining electrodes during discharge. Therefore, due to the change in current, the machining accuracy also changes.

[0044] In order to correct the electrical characteristics themselves, the electrical discharge machining apparatus 1A of Embodiment 1 corrects the applied voltage value and the inter-electrode current value, so that stable processing quality can be achieved regardless of the size of the workpiece 3, the type of wire of the wire electrode, the state of the cable, etc.

[0045] Furthermore, the EDM apparatus 1A can correct the applied voltage and inter-electrode current values ​​without adjusting the processing conditions provided by the EDM apparatus manufacturer, thus making processing quality management easier.

[0046] Furthermore, in order to correct the inherent electrical characteristics of the electrical discharge machining apparatus 1A, the applied voltage value and the inter-electrode current value are corrected, thus suppressing individual differences in the machining accuracy of the electrical discharge machining apparatus 1A. In addition, in order to correct the electrical characteristics, the electrical discharge machining apparatus 1A can also correct the discharge frequency, the voltage interval, the peak value of the discharge current, the pulse width of the discharge current, the axis feed speed, etc.

[0047] Here, the detailed structure of the learning device 14 will be described. Figure 2 This diagram illustrates the structure of the learning device included in the electrical discharge machining apparatus according to Embodiment 1. The learning device 14 includes a state observation unit 141, a data acquisition unit 142, a learning unit 143, and an inference unit 144.

[0048] The state observation unit 141 obtains the applied voltage value from the voltage command detection unit 12. Additionally, the state observation unit 141 obtains the inter-electrode current value from the current detection unit 9. During learning and inference, the state observation unit 141 sends both the applied voltage value and the inter-electrode current value to the learning unit 143.

[0049] During learning, the data acquisition unit 142 acquires the impedance measured for the discharge circuit. The impedance can be measured by any method. Additionally, during learning, the data acquisition unit 142 acquires the measured parasitic capacitance value αC between the processing electrodes. The parasitic capacitance value αC can be measured by any method. The data acquisition unit 142 transmits the impedance and the parasitic capacitance value αC to the learning unit 143.

[0050] The learning unit 143 learns the impedance (predicted value) corresponding to the inter-electrode current value based on a dataset (learning data) created from a combination of the inter-electrode current value received from the state observation unit 141 and the impedance received from the data acquisition unit 142. That is, the learning unit 143 uses the learning data containing the inter-electrode current value and the impedance to generate a learning model (hereinafter sometimes referred to as the impedance learning model) for inferring the impedance based on the inter-electrode current value.

[0051] Here, the dataset used to generate the impedance learning model is data that correlates the inter-electrode current values ​​as state variables with the impedance as decision data. An impedance learning model is generated for each electrical discharge machining (EDM) apparatus 1A. That is, EDM apparatus 1A does not use data from other apparatuses, but instead uses data measured specifically for its own apparatus to generate the impedance learning model.

[0052] Furthermore, the learning unit 143 calculates the amount of electricity flowing through the processed electrodes by performing time integration on the inter-electrode current value received from the state observation unit 141. The learning unit 143 learns the parasitic capacitance value αC corresponding to the applied voltage value and the calculated amount of electricity based on a dataset created from a combination of the parasitic capacitance value αC between the processed electrodes, the applied voltage value received from the state observation unit 141, and the calculated amount of electricity. That is, the learning unit 143 uses learning data containing the parasitic capacitance value αC, the applied voltage value, and the amount of electricity to generate a learning model (hereinafter sometimes referred to as the parasitic capacitance learning model) for inferring the parasitic capacitance value αC based on the applied voltage value and the amount of electricity.

[0053] Here, the dataset used to generate the parasitic capacitance learning model is data that correlates the applied voltage value and electrical charge as state variables with the parasitic capacitance value αC as decision data. A parasitic capacitance learning model is generated for each electrical discharge machining (EDM) apparatus 1A. That is, the EDM apparatus 1A does not use data from other apparatuses, but instead uses data measured specifically for its own apparatus to generate the parasitic capacitance learning model.

[0054] When the applied voltage value is set to the applied voltage value V and the charge is set to the charge Q, the learning unit 143 learns the parasitic capacitance value αC corresponding to the applied voltage value and the charge by learning the relationship of the following equation (1).

[0055] Q=αC×V (1)

[0056] The impedance learning model and parasitic capacitance learning model generated by the learning unit 143 are stored in a storage device (not shown) provided in the electrical discharge machining apparatus 1A. The inference unit 144 uses the impedance learning model to infer the impedance based on the applied voltage value. Specifically, the inference unit 144 infers the impedance corresponding to the applied voltage value by inputting the applied voltage value into the impedance learning model.

[0057] Furthermore, the inference unit 144 uses a parasitic capacitance learning model to infer the parasitic capacitance value αC based on the applied voltage value and the electrical charge. Specifically, the inference unit 144 infers the parasitic capacitance value αC corresponding to the applied voltage value and the electrical charge by inputting the applied voltage value and the electrical charge into the parasitic capacitance learning model. The inference unit 144 sends the inference result, i.e., the impedance and the parasitic capacitance value αC, to the correction value determination unit 16.

[0058] The correction value determination unit 16 corrects at least one of the voltage value of the voltage applied by the power supply device 20 to the processing electrode and the current value of the current flowing through the processing electrode based on the impedance and parasitic capacitance value αC received from the inference unit 144.

[0059] Furthermore, the learning device 14 may be a separate device from the electrical discharge machining apparatus 1A, connected to the electrical discharge machining apparatus 1A via a network. Alternatively, the learning device 14 may reside on a cloud server.

[0060] The learning device 14, for example, learns the impedance of the discharge circuit or the parasitic capacitance value αC between the processing electrodes through so-called teacher-guided learning, according to a neural network model. Here, teacher-guided learning refers to a method that learns the features of the dataset by providing a large set of data containing a certain input and a result (label) to the machine learning device, and infers the result based on the input.

[0061] The neural network consists of the following layers: an input layer composed of multiple neurons, an intermediate layer (hidden layer) composed of multiple neurons, and an output layer composed of multiple neurons. The intermediate layer can be one layer or two or more layers. Furthermore, a portion of the processing performed by the learning device 14 can be performed by the arithmetic unit 13.

[0062] Figure 3 This is a diagram illustrating an example of the structure of the neural network used in the learning device of Embodiment 1. For example, in Figure 3 In the case of the 3-layer neural network shown, if multiple inputs are fed into the input layer (X1 to X3), the values ​​are multiplied by weights W1 (w11 to w16) and then fed into the intermediate layer (Y1 to Y2). The result is then further multiplied by weights W2 (w21 to w26) and output from the output layer (Z1 to Z3). This output varies depending on the values ​​of weights W1 and W2.

[0063] Here, the learning device 14 used Figure 3 The processing performed by the neural network in the case of an impedance learning model will be explained. In the case of an impedance learning model, the neural network learns the impedance corresponding to the inter-electrode current value through so-called teacher-guided learning, based on a dataset created by a combination of the inter-electrode current value obtained by the state observation unit 141 and the impedance obtained by the data acquisition unit 142.

[0064] That is, in the case of the impedance learning model, the neural network learns by adjusting weights W1 and W2 so that the output from the output layer, when the input is the inter-electrode current value (the first input), is close to the second input, i.e., the impedance (the correct answer). In this case, the learning device 14 learns the correspondence between the inter-electrode current value and the impedance, thereby generating an impedance learning model that can output the appropriate impedance when the inter-electrode current value is input. As described above, the learning device 14 learns the learning model, which can output the impedance as the correct answer when the inter-electrode current value is used as input.

[0065] Next, the learning device 14 used Figure 3 The process performed by the neural network in the case of a parasitic capacitance learning model will be explained. In this case, the neural network learns the parasitic capacitance value αC corresponding to the applied voltage value and the electrical quantity based on a dataset created by a combination of the applied voltage value and the electrical quantity obtained by the state observation unit 141 and the parasitic capacitance value αC obtained by the data acquisition unit 142, through so-called teacher-guided learning.

[0066] That is, in the case of the parasitic capacitance learning model, the neural network learns by adjusting weights W1 and W2 so that the output from the output layer, given the applied voltage and charge as the first input, is close to the second input, i.e., the parasitic capacitance value αC (the correct answer). In this case, the learning device 14 learns the correspondence between the applied voltage and charge and the parasitic capacitance value αC, thereby generating a learning model that can output the appropriate parasitic capacitance value αC given the applied voltage and charge as input. As described above, the learning device 14 learns the learning model, which can output the parasitic capacitance value αC as the correct answer when the applied voltage and charge are given as input.

[0067] Furthermore, the learning device 14 can also learn the impedance or parasitic capacitance value αC through so-called unsupervised learning. Unsupervised learning refers to a method that learns the distribution of input data by simply providing a large amount of input data to the machine learning device, and learns about compression, classification, shaping, etc., based on the input data even without providing corresponding teacher output data. In unsupervised learning, the learning device 14 can cluster the features of the dataset into similar groups. By using the clustering results to set a benchmark, the learning device 14 can allocate outputs that make the benchmark the best, thereby achieving output prediction. Additionally, the learning device 14 can also learn the impedance or parasitic capacitance value αC through reinforcement learning. Representative methods of reinforcement learning include Q-learning and TD-learning.

[0068] Furthermore, the learning algorithm used by the learning device 14 can also be deep learning, which involves learning by extracting the feature quantities themselves. Alternatively, the learning device 14 can also perform machine learning using other known methods, such as genetic programming, functional logic programming, and support vector machines.

[0069] The electrical discharge machining (EDM) apparatus 1A generates an impedance learning model and a parasitic capacitance learning model by performing machining under various machining conditions. The EDM apparatus 1A performs machining with various inter-electrode current values ​​and measures the impedance during each machining process. The learning device 14 of the EDM apparatus 1A generates an impedance learning model by learning the impedance corresponding to the inter-electrode current value based on a dataset created according to the combination of inter-electrode current value and impedance.

[0070] Furthermore, the electrical discharge machining (EDM) apparatus 1A performs machining with various applied voltage values ​​and inter-electrode current values, and measures the parasitic capacitance value αC during each machining process. The learning device 14 of the EDM apparatus 1A calculates the electrical charge flowing through the machining inter-electrode by integrating the inter-electrode current values ​​during each machining process over time. The learning device 14 generates a parasitic capacitance learning model by learning the parasitic capacitance value αC corresponding to the applied voltage value and the electrical charge based on a dataset created from a combination of applied voltage value, parasitic capacitance value αC, and electrical charge.

[0071] When the electrical discharge machining (EDM) apparatus 1A actually processes the workpiece 3, it uses an impedance learning model to infer the impedance and a parasitic capacitance learning model to infer the parasitic capacitance value αC. The EDM apparatus 1A calculates correction values ​​corresponding to the inferred impedance and the inferred parasitic capacitance value αC. The correction value corresponding to the inferred impedance corresponds to the difference between the reference impedance and the inferred impedance. Similarly, the correction value corresponding to the inferred parasitic capacitance value αC corresponds to the difference between the reference parasitic capacitance value αC and the inferred parasitic capacitance value αC. As described above, the EDM apparatus 1A calculates correction values ​​corresponding to the deviation between the inferred electrical characteristics and the reference electrical characteristics (reference characteristics).

[0072] The correction value calculated by the electrical discharge machining apparatus 1A includes a correction value (voltage correction value) for correcting the voltage value of the voltage applied by the power supply device 20 to the machining electrode and a correction value (current correction value) for correcting the current value of the current flowing through the machining electrode.

[0073] The voltage correction value is a correction value relative to the voltage reference value, and the current correction value is a correction value relative to the current reference value. The electrical discharge machining apparatus 1A corrects the voltage reference value by using the voltage correction value and corrects the current reference value by using the current correction value, thereby correcting the inherent impedance and parasitic capacitance value αC of the electrical discharge machining apparatus 1A.

[0074] Figure 4 This is a flowchart illustrating the processing flow of the electrical characteristics correction performed by the electrical discharge machining apparatus according to Embodiment 1. The electrical discharge machining apparatus 1A stores an impedance learning model and a parasitic capacitance learning model obtained through learning.

[0075] The electrical discharge machining apparatus 1A sets the machining parameters according to the instructions from the user (step S10). The machining parameters are physical quantities that are used to perform electrical discharge machining on the workpiece 3. An example of the machining parameters is the thickness of the workpiece 3, the material of the workpiece 3, the position of the upper and lower nozzles, and the diameter of the tool electrode 4, i.e., the wire electrode diameter.

[0076] The electrical discharge machining apparatus 1A begins machining the workpiece 3 using machining parameters and obtains a machining waveform (step S20). The machining waveform here includes the waveform of the inter-electrode current value measured by the current detection unit 9 and the waveform of the applied voltage value detected by the voltage command detection unit 12.

[0077] The learning device 14 of the electrical discharge machining apparatus 1A infers the impedance based on the inter-electrode current value by inputting the inter-electrode current value into the impedance learning model (step S30). In other words, the learning device 14 infers the impedance component among the components that affect the machining quality based on the inter-electrode current value.

[0078] Furthermore, the learning device 14 calculates the amount of electricity flowing between the processing electrodes by integrating the inter-electrode current value over time. The learning device 14 infers the parasitic capacitance value αC based on the applied voltage and amount of electricity input to the parasitic capacitance learning model (step S40). In other words, the learning device 14 infers the parasitic capacitance component that affects processing quality based on the applied voltage and amount of electricity.

[0079] Here, we will explain the process of inferring impedance based on the inter-electrode current value and the process of inferring parasitic capacitance αC based on the applied voltage value and the electrical charge. Figure 5 This is a diagram showing an example of the waveform of the inter-electrode current value obtained by the electrical discharge machining apparatus according to Embodiment 1. Figure 6 This diagram illustrates the process by which the electrical discharge machining apparatus of Embodiment 1 infers impedance and parasitic capacitance values. Figure 5 as well as Figure 6 The horizontal axis of the graph shown represents time, and the vertical axis represents the inter-electrode current value.

[0080] The current detection unit 9 measures the inter-electrode current value at a specific period, and therefore the obtained inter-electrode current value is as follows: Figure 5 The values ​​shown are fragmented. The inference unit 144 calculates the waveform 40 of the continuously changing inter-electrode current value based on the fragmented inter-electrode current value through interpolation processing and the like.

[0081] The inference unit 144 infers the impedance based on the waveform 40 of the inter-electrode current value and the impedance learning model.

[0082] Furthermore, the inference unit 144 calculates the electrical charge by integrating the inter-electrode current value over time. When the applied voltage is constant, the electrical charge is proportional to the parasitic capacitance value αC. That is, when the applied voltage is constant, the time integral of the inter-electrode current value is proportional to the parasitic capacitance value αC. Based on a learning model calculated from the waveform 40 of the inter-electrode current value, the applied voltage value, and the parasitic capacitance, the inference unit 144 infers the parasitic capacitance value αC based on the electrical charge and the applied voltage value.

[0083] Furthermore, the inference unit 144 can also infer the parasitic capacitance value αC using the impedance immediately after discharge. In this case, the inference unit 144 differentiates the inter-electrode current value immediately after discharge with respect to time and calculates the slope 41 of the inter-electrode current value immediately after discharge. That is, the inference unit 144 calculates the time increase of the inter-electrode current value immediately after discharge. The time increase of the inter-electrode current value immediately after discharge is proportional to the impedance immediately after discharge. The inference unit 144 uses the calculated impedance immediately after discharge to infer the parasitic capacitance value αC. As a result, the inference unit 144 can improve the accuracy of inferring the parasitic capacitance value αC.

[0084] The electrical discharge machining apparatus 1A calculates the correction values ​​(voltage correction value and current correction value) used to correct the impedance and parasitic capacitance value αC (step S50). The electrical discharge machining apparatus 1A determines the calculated correction value as the current correction value of the electrical discharge machining apparatus 1A (step S60).

[0085] The electrical discharge machining (EDM) apparatus 1A executes steps S10 to S60 at the start of machining. Alternatively, the EDM apparatus 1A may execute step S10 at the start of machining and simultaneously execute steps S20 to S60. In this case, if the machining waveform changes, the EDM apparatus 1A executes steps S30 to S60.

[0086] Here, the range that the comparative electrical discharge machining apparatus (hereinafter referred to as the comparative electrical discharge machining apparatus) can correct and the range that the electrical discharge machining apparatus 1A of Embodiment 1 can correct will be explained.

[0087] The contrast EDM apparatus undergoes machining tests upon leaving the factory, and various adjustments are made based on the results of these tests. This process adjusts for machining errors inherent in the mechanical assembly of the contrast EDM apparatus; however, the machining tests are time-consuming and costly.

[0088] Furthermore, a method is provided for adjusting the processing conditions applied to a comparative electrical discharge machining (EDM) apparatus using a processing condition adjustment device. In this method, the processing condition adjustment device learns the adjustment actions of processing conditions corresponding to processing states such as processing voltage, thereby predicting the adjustment actions of processing conditions based on the processing state during a predetermined processing operation.

[0089] This machining condition adjustment device learns from the common adjustment actions among multiple EDM devices, therefore it cannot eliminate the machining errors generated by each individual EDM device. Consequently, the machining condition adjustment device cannot correct machining errors caused by the inherent electrical characteristics (impedance, parasitic capacitance αC, etc.) of the comparative EDM devices, resulting in reduced machining quality.

[0090] For example, in a contrast electrical discharge machining (SDM) apparatus, the inherent electrical characteristics of the SDM apparatus change due to the mechanical installation difference during installation, but the SDM apparatus cannot correct the machining error corresponding to the mechanical installation difference.

[0091] Furthermore, in a comparative electrical discharge machining (EDM) apparatus, if the state of the apparatus changes due to the passage of time, the inherent electrical characteristics of the apparatus will change, but the apparatus cannot correct the machining error corresponding to the change in its state.

[0092] Furthermore, in a comparative electrical discharge machining (EDM) apparatus, if the state of the machining fluid changes, the inherent electrical characteristics of the comparative EDM apparatus will change, but the comparative EDM apparatus cannot correct the machining error corresponding to the change in the state of the machining fluid.

[0093] Furthermore, in contrast electrical discharge machining (EDM) apparatuses, the inherent electrical characteristics of the apparatus change due to component replacement, but the contrast EDM apparatus cannot correct the machining errors caused by component replacement.

[0094] Furthermore, the processing condition adjustment device does not eliminate errors generated by each individual process; it merely establishes a correspondence between processing conditions and their adjustment actions. Therefore, with respect to the processing condition adjustment device, the adjustments to the processing conditions change each time the type of workpiece being processed or the type of wire electrode is changed, which is time-consuming.

[0095] For example, whenever the type of workpiece being processed, the type of electrode, or the state of the cable connecting the power supply to the processing electrode in a contrast electrical discharge machining (SEM) apparatus changes, the SEM apparatus must adjust its processing conditions accordingly. Therefore, for a contrast electrical discharge machining (SEM) apparatus, each set of processing conditions provided by the manufacturer needs to be individually calibrated, making processing condition management difficult.

[0096] On the other hand, the electrical discharge machining apparatus 1A of Embodiment 1 generates an impedance learning model during learning for inferring impedance based on the applied voltage value, and uses the impedance learning model to infer impedance based on the applied voltage value during inference. Additionally, the electrical discharge machining apparatus 1A generates a parasitic capacitance learning model during learning for inferring parasitic capacitance value αC based on the applied voltage value and the electrical charge, and infers parasitic capacitance value αC based on the applied voltage value and the electrical charge during inference.

[0097] Furthermore, the electrical discharge machining apparatus 1A calculates voltage correction values ​​and current correction values ​​for correcting impedance and parasitic capacitance value αC, and uses the voltage correction values ​​and current correction values ​​to correct the voltage and current during machining.

[0098] As described above, even if the inherent electrical characteristics of the electrical discharge machining apparatus 1A change, it can infer the changed electrical characteristics and calculate the voltage correction value and current correction value based on the electrical characteristics, thus making it easy to correct machining errors.

[0099] Figure 7 This diagram illustrates the range that can be corrected by the electrical discharge machining apparatus according to Embodiment 1. The electrical discharge machining apparatus 1A can correct for individual differences in machining errors caused by mechanical installation discrepancies, machining errors caused by changes in mechanical state over time, machining errors caused by the state of the machining fluid, and machining errors caused by differences in replaced parts.

[0100] For example, even if the inherent electrical characteristics of the electrical discharge machining apparatus 1A change due to mechanical installation errors, the electrical discharge machining apparatus 1A can correct the machining errors corresponding to the mechanical installation errors by inferring the electrical characteristics and using voltage and current correction values ​​corresponding to the electrical characteristics.

[0101] Furthermore, even when mechanical condition differences arise due to changes over time, the electrical discharge machining apparatus 1A can correct machining errors corresponding to mechanical condition differences by inferring electrical characteristics and using voltage and current correction values ​​corresponding to those electrical characteristics.

[0102] Furthermore, even when there is a state difference in the processing fluid, since the electrical discharge machining apparatus 1A infers the electrical characteristics and uses the voltage correction value and current correction value corresponding to the electrical characteristics, it is also possible to correct the processing error corresponding to the state difference in the processing fluid.

[0103] Furthermore, even if the replaced parts have differences (such as differences in wire diameter of the wire electrode), the electrical discharge machining apparatus 1A can correct the machining errors caused by the differences in the replaced parts by inferring the electrical characteristics and using voltage correction values ​​and current correction values ​​corresponding to the electrical characteristics.

[0104] Furthermore, even if the type of workpiece 3 processed by the electrical discharge machining apparatus 1A changes or the type of wire electrode changes, since the electrical discharge machining apparatus 1A infers the electrical characteristics and uses voltage correction values ​​and current correction values ​​corresponding to the electrical characteristics, there is no need to change the adjustment of the processing conditions.

[0105] As described above, in Embodiment 1, the electrical discharge machining apparatus 1A infers the electrical characteristics and corrects the electrical characteristics using voltage correction values ​​and current correction values ​​corresponding to the electrical characteristics. Therefore, it is possible to correct the machining errors caused by the electrical characteristics without relying on individual differences caused by the installation difference of the machine (machine), changes over time, or changes in the setting environment.

[0106] Furthermore, since the electrical discharge machining apparatus 1A corrects for deviations in each individual component of the electrical discharge machining apparatus 1A, it can suppress variations in machining characteristics and eliminates the need for individual adjustments to machining conditions.

[0107] As described above, the electrical discharge machining apparatus 1A of Embodiment 1 uses a learning model to infer the inherent electrical characteristics of the apparatus based on the inter-electrode current value and the applied voltage value. Furthermore, the electrical discharge machining apparatus 1A corrects the inter-electrode current value and the applied voltage value so that the inferred electrical characteristics become a reference electrical characteristic, i.e., a reference characteristic. Therefore, the electrical discharge machining apparatus 1A can correct the inherent electrical characteristics of the apparatus, thus maintaining constant machining accuracy and suppressing a decrease in machining quality.

[0108] Implementation Method 2

[0109] Next, use Figures 8 to 17 Embodiment 2 will be described. The electrical discharge machining apparatus 1A of Embodiment 1 can infer the electrical characteristics present in the machining circuit based on the inter-electrode current value and the applied voltage value detected during machining. When using the electrical discharge machining apparatus 1A, the user needs to perform preparatory machining to obtain the inter-electrode current value and the applied voltage value before actual machining. In Embodiment 2, preparatory machining is not required before machining, and the electrical characteristics of the machining circuit are inferred. The machining circuit includes electrical circuits such as cables, machining electrodes, tool electrodes 4, and workpiece 3, etc.

[0110] In Embodiment 2, the electrical characteristics include at least one of impedance, parasitic capacitance αC, parasitic inductance, processing fluid resistance, and wiring resistance. In Embodiment 2, the parasitic capacitance αC, parasitic inductance, and other electrical characteristics are inferred by comparing the frequency characteristics of the impedance and the frequency characteristics of the phase delay with the frequency characteristics calculated based on the circuit equations of the equivalent circuit of the processing circuit. The phase delay is the phase difference between the inter-electrode current value and the applied voltage value.

[0111] Figure 8 This is a diagram showing the structure of the electrical discharge machining apparatus according to Embodiment 2. For Figure 8 Among the structural elements Figure 1 The structural elements of the electrical discharge machining apparatus 1A of Embodiment 1 that perform the same function are labeled with the same reference numerals, and repeated descriptions are omitted.

[0112] The electrical discharge machining apparatus 1B can deduce the electrical characteristics of the machining circuit without pre-processing. Compared with the electrical discharge machining apparatus 1A, the electrical discharge machining apparatus 1B has a machining section 2B instead of a machining section 2A, and an arithmetic unit 10B instead of an arithmetic unit 10A. That is, the electrical discharge machining apparatus 1B has a machining section 2B, a power supply device 20, and an arithmetic unit 10B.

[0113] In addition to the structural elements present in the machining unit 2A, the machining unit 2B also has a voltage detection unit 8. Compared to the arithmetic unit 10A, the arithmetic unit 10B lacks a voltage command detection unit 12. Furthermore, compared to the arithmetic unit 10A, the arithmetic unit 10B replaces the learning unit 14 with an inference unit 15.

[0114] The voltage detection unit 8 is electrically connected in parallel with the processing electrodes and detects the voltage value of the voltage applied from the power supply device 20 to the processing electrodes, which is then used as the applied voltage value. The voltage detection unit 8 sends the detected applied voltage value to the arithmetic unit 10B.

[0115] The current detection unit 9 is connected in series between the machining electrode space and the power supply device 20, and detects the current value supplied from the power supply device 20 to the machining electrode space as the electrode space current value. The current detection unit 9 sends the detected electrode space current value to the arithmetic unit 10B.

[0116] The control unit 11 of the arithmetic unit 10B controls the power supply unit 20 and the display 25. The control unit 11 controls, for example, the parameters of the power supply unit 20 (applied voltage value, inter-electrode current value, discharge frequency, voltage interval time, peak value of discharge current, pulse width of discharge current, shaft feed speed, etc.).

[0117] The arithmetic unit 13 of the arithmetic device 10B calculates the frequency characteristics of the impedance and the frequency characteristics of the phase delay (phase delay) based on the output value from the voltage detection unit 8, i.e., the applied voltage value, and the output value from the current detection unit 9, i.e., the inter-electrode current value.

[0118] The inference device 15 of the arithmetic unit 10B infers the parasitic capacitance value αC and parasitic inductance present in the processing circuit based on the frequency characteristics of the impedance output from the arithmetic unit 13 and the frequency characteristics of the phase delay. Furthermore, regardless of the circuit structure, the power supply device 20 of Embodiment 2 is preferably a device capable of applying both AC and DC voltages to the processing electrodes.

[0119] Here, we will explain the method by which the electrical discharge machining apparatus 1B infers the parasitic capacitance αC and parasitic inductance on the machining circuit (between the machining electrodes). In the separated state where the tool electrode 4 is not in contact with the workpiece 3, the parasitic capacitance αC dominates; in the short-circuit state where the tool electrode 4 is in contact with the workpiece 3, the parasitic inductance dominates. Therefore, the electrical discharge machining apparatus 1B infers the parasitic capacitance αC between the machining electrodes based on the frequency characteristics of the machining circuit in the separated state, and infers the parasitic inductance on the machining circuit based on the frequency characteristics of the machining circuit in the short-circuit state.

[0120] First, the method for inferring the parasitic capacitance value αC from the electrical discharge machining apparatus 1B will be explained. Figure 9 This is a diagram showing the processing circuit of the processing section in the case where the processing electrodes of the electrical discharge machining apparatus according to Embodiment 2 are in a separated state. Figure 10 It means and Figure 9 The diagram shows the equivalent circuit corresponding to the processing circuit. For example... Figure 9 As shown, in the machining circuit 51 of machining section 2B, when the tool electrode 4 is not in contact with the workpiece 3, i.e., when the machining electrodes are separated, the machining circuit 51 is operated by... Figure 10 The equivalent circuit 55 shown is represented. The processing circuit 51 in the separated state is the first processing circuit.

[0121] In the equivalent circuit 55, a parasitic inductance PX1 is generated between the power supply device 20 and the tool electrode 4, a parasitic capacitance value C1 is generated between the tool electrode 4 and the workpiece 3 as a parasitic capacitance value αC, and a resistance of the machining fluid, namely the machining fluid resistance R1, is generated between the machining electrodes.

[0122] In this state, the electrical discharge machining apparatus 1B applies an AC voltage from the power supply device 20 and detects both the inter-electrode current value and the applied voltage value. The arithmetic unit 10B calculates the impedance and phase delay based on the detected inter-electrode current value and the applied voltage value.

[0123] Figure 11This is a diagram used to illustrate the impedance and phase delay calculated by the electrical discharge machining apparatus according to Embodiment 2 when the machining electrodes are in a separated state. Figure 11 The upper graph shown represents time on the horizontal axis and the applied voltage value on the vertical axis. Additionally, Figure 11 The horizontal axis of the lower-level graph shown represents time, and the vertical axis represents the inter-electrode current value.

[0124] The arithmetic unit 10B of the electrical discharge machining apparatus 1B calculates the impedance IZ based on the detected inter-electrode current value Ip and the applied voltage value Vp. The inter-electrode current value Ip is the maximum value (peak value) of the absolute value of the inter-electrode current value Ip, corresponding to the amplitude of the waveform of the inter-electrode current value. The applied voltage value Vp is the maximum value of the absolute value of the applied voltage value Vp, corresponding to the amplitude of the waveform of the applied voltage value. There is a relationship between the inter-electrode current value Ip, the applied voltage value Vp and the impedance IZ as shown in Equation (2), therefore the arithmetic unit 10B uses Equation (2) to calculate the impedance IZ.

[0125] Z = Vp / Ip (2)

[0126] Furthermore, the arithmetic unit 10B calculates the phase delay (angle) θ based on the period T1 of the applied voltage value Vp and the phase delay time t1. The phase delay time t1 is the time of phase delay of the inter-electrode current value Ip relative to the applied voltage value Vp. The relationship between the period T1 and the phase delay time t1 is given by the following equation (3), therefore the arithmetic unit 10B uses equation (3) to calculate the phase delay θ.

[0127] θ = 2π × (t / T) (3)

[0128] The arithmetic unit 10B here calculates the phase delay θ by calculating 2π×(t1 / T1). By performing the above-mentioned processes of calculating the impedance IZ and calculating the phase delay θ while changing the AC frequency of the processing power supply, the arithmetic unit 10B is able to obtain the frequency characteristics of the processing circuit 51 when the processing electrodes are in a separated state.

[0129] Figure 12 This is a graph showing the frequency characteristics calculated by the electrical discharge machining apparatus according to Embodiment 2 when the machining electrodes are in a separated state. Figure 12 The horizontal axis represents frequency, the left vertical axis represents impedance, and the right vertical axis represents phase delay. Figure 12 In this context, the frequency characteristics of the impedance in the separated state are represented as impedance IZ1, and the frequency characteristics of the phase delay in the separated state are represented as phase delay θ1.

[0130] The inference device 15 of the electrical discharge machining apparatus 1B infers the parasitic capacitance value C1 between the machining electrodes by comparing the obtained frequency characteristics with the frequency characteristics calculated based on the circuit equation of the equivalent circuit 55. Specifically, the inference device 15 infers the parasitic capacitance value C1 between the machining electrodes by comparing the frequency characteristics of the impedance and phase delay calculated for the separated state with the frequency characteristics of the impedance and phase delay calculated based on the circuit equation of the equivalent circuit 55. The circuit equation of the equivalent circuit 55 in the separated state is the first circuit equation. Furthermore, the frequency characteristics of the machining circuit 51 in the separated state are the first frequency characteristics, and the frequency characteristics of the circuit equation of the equivalent circuit 55 in the separated state are the second frequency characteristics.

[0131] Furthermore, the impedance obtained during the inference of the parasitic capacitance value C1 is consistent with the resistance R1 of the processing fluid present between the processing electrodes. Therefore, the inference device 15 can also calculate the resistance R1 of the processing fluid as a value corresponding to the state of the processing fluid.

[0132] Next, the method for inferring parasitic inductance using the electrical discharge machining apparatus 1B will be explained. Figure 13 This is a diagram showing the processing circuit of the processing section in the case where the processing electrodes of the electrical discharge machining apparatus according to Embodiment 2 are in a short-circuit state. Figure 14 It means and Figure 13 The diagram shows the equivalent circuit corresponding to the processing circuit. For example... Figure 13 As shown, in the machining circuit 52 of machining section 2B, when the tool electrode 4 is in contact with the workpiece 3, i.e., in a short-circuit state between the machining electrodes, the machining circuit 52 is operated by... Figure 14 The equivalent circuit 56 shown is represented. The processing circuit 52 in the short-circuit state is the second processing circuit.

[0133] In the equivalent circuit 56, a wiring resistance R2 is generated on the equivalent circuit 56, a parasitic capacitance value C2 is generated between the tool electrode 4 and the workpiece 3 as a parasitic capacitance value αC, and a resistance of the machining fluid, namely the machining fluid resistance R1, is generated between the tool electrode 4 and the workpiece 3. Figure 14 (Not shown in the diagram) and parasitic inductance PX2. The wiring resistance R2 on the equivalent circuit 56 includes the resistance of the wiring present on the equivalent circuit 56, the contact resistance between the wiring and the tool electrode 4, etc. In addition, the machining fluid resistance R1 can be ignored under the short-circuit condition between the machining electrodes.

[0134] In this state, the electrical discharge machining apparatus 1B applies an AC voltage from the power supply device 20 and detects the inter-electrode current value and the applied voltage value. The arithmetic unit 10B calculates the impedance and phase delay based on the detected inter-electrode current value and applied voltage value. The arithmetic unit 10B performs the same processing as in the separated state between the machining electrodes to calculate the impedance and phase delay in the short-circuit state between the machining electrodes.

[0135] Figure 15 This is a diagram used to illustrate the impedance and phase delay calculated by the electrical discharge machining apparatus according to Embodiment 2 when the machining electrodes are in a short-circuit state. Figure 15 The upper graph shown represents time on the horizontal axis and the applied voltage value on the vertical axis. Additionally, Figure 15 The horizontal axis of the lower-level graph shown represents time, and the vertical axis represents the inter-electrode current value.

[0136] In the same case where the arithmetic unit 10B of the electrical discharge machining apparatus 1B is separated from the machining electrode, the impedance IZ is calculated based on equation (2), the detected inter-electrode current value Ip, and the applied voltage value Vp.

[0137] Furthermore, in the case where the arithmetic unit 10B is separated from the processing electrode, the phase delay θ is calculated based on equation (3), the period T2 of the applied voltage value Vp, and the phase delay time t2. The phase delay time t2 is the time of phase delay of the inter-electrode current value Ip relative to the applied voltage value Vp.

[0138] The arithmetic unit 10B here calculates the phase delay θ by calculating 2π×(t2 / T2). By performing the above-mentioned processes of calculating the impedance IZ and calculating the phase delay θ while changing the AC frequency of the processing power supply, the arithmetic unit 10B can obtain the frequency characteristics of the processing circuit 52 when the processing electrodes are in a short-circuit state.

[0139] Figure 16 This is a graph showing the frequency characteristics calculated by the electrical discharge machining apparatus according to Embodiment 2 when the machining electrodes are in a short-circuit state. Figure 16 The horizontal axis represents frequency, the left vertical axis represents impedance, and the right vertical axis represents phase delay. Figure 16 In this context, the frequency characteristic of the impedance under short-circuit conditions is expressed as impedance IZ2, and the frequency characteristic of the phase delay under short-circuit conditions is expressed as phase delay θ2.

[0140] Similar to the process for inferring the parasitic capacitance value C1, the inference device 15 of the electrical discharge machining apparatus 1B infers the parasitic inductance PX2 on the machining circuit by comparing the obtained frequency characteristics with the frequency characteristics calculated based on the circuit equations of the equivalent circuit 56. Specifically, the inference device 15 infers the parasitic inductance PX2 on the machining circuit by comparing the frequency characteristics of the impedance and phase delay calculated for the short-circuit state with the frequency characteristics of the impedance and phase delay calculated based on the circuit equations of the equivalent circuit 56. The circuit equation of the equivalent circuit 56 in the short-circuit state is the second circuit equation. Furthermore, the frequency characteristics of the machining circuit 52 in the short-circuit state are the third frequency characteristics, and the frequency characteristics of the circuit equations of the equivalent circuit 56 in the short-circuit state are the fourth frequency characteristics.

[0141] Furthermore, the impedance obtained during the deduction of parasitic inductance PX2 is consistent with the wiring resistance R2 on the machining circuit. Therefore, the deduction device 15 can also detect the wiring resistance R2 as a value corresponding to the installation state of the machining circuit. The installation state of the machining circuit includes the state of the wiring present on the machining circuit, the state of the tool electrode 4, and the contact state between the wiring and the tool electrode 4. That is, the wiring resistance R2 includes the resistance of the wiring itself, the resistance of the tool electrode 4, and the resistance of the contact portion between the wiring and the tool electrode 4.

[0142] Electrical discharge machining apparatus 1B passes through Figures 9 to 16 The processing described herein allows for the prior estimation of parasitic capacitance C1, parasitic inductance PX2, and impedance on the processing circuit, even without actual discharge processing.

[0143] Figure 17 This is a flowchart illustrating the processing flow of the electrical characteristics correction performed by the electrical discharge machining apparatus according to Embodiment 2. The electrical discharge machining apparatus 1B sets the processing electrodes to a separated state (step S110).

[0144] The control unit 11 sets the frequency of the processing power supply to the power supply device 20 (step S120). The power supply device 20 applies an AC voltage of the set frequency to the processing electrodes (step S130).

[0145] The voltage detection unit 8 and the current detection unit 9 detect the voltage value and the current value, respectively (step S140). The calculation unit 13 calculates the impedance and phase delay based on the detected values ​​(voltage value and current value) (step S150).

[0146] The control unit 11 determines whether the calculation of impedance and phase delay at all set frequencies has been completed (step S160). If the calculation of impedance and phase delay at all set frequencies has been completed (step S160, No), the control unit 11 sets a new frequency for the processing power supply to the power supply device 20 (step S170). Then, the electrical discharge machining apparatus 1B executes the processes of steps S130 to S160. The electrical discharge machining apparatus 1B repeats the processes of steps S130 to S160 until the calculation of impedance and phase delay at all set frequencies is completed.

[0147] If the calculation of impedance and phase delay at all set frequencies has been completed (step S160, Yes), the control unit 11 determines whether the measurement when the processing poles are in a short-circuit state has been completed (step S180).

[0148] If the measurement of the machining electrode gap being in a short-circuit state has not yet been completed (step S180, No), the electrical discharge machining apparatus 1B sets the machining electrode gap to a short-circuit state (step S190). Then, the electrical discharge machining apparatus 1B executes the processes of steps S120 to S180. The electrical discharge machining apparatus 1B repeats the processes of steps S120 to S180 until the measurement of the machining electrode gap being in a short-circuit state is completed.

[0149] If the measurement is completed when the electrode spacing is short-circuited (step S180, Yes), the inference device 15 infers the parasitic capacitance value αC and the parasitic inductance (step S200). Specifically, the inference device 15 infers the parasitic capacitance value αC between the electrode spacing by comparing the frequency characteristics of the impedance and the frequency characteristics of the phase delay calculated in the separated state between the electrode spacing with the frequency characteristics of the impedance and the frequency characteristics of the phase delay calculated based on the circuit equations of the separated state.

[0150] In addition, the inference device 15 infers the parasitic inductance by comparing the frequency characteristics of the impedance and the frequency characteristics of the phase delay calculated in the short-circuit state between the processing poles with the frequency characteristics of the impedance and the frequency characteristics of the phase delay calculated according to the circuit equations in the short-circuit state.

[0151] Furthermore, the inference device 15 can also infer the machining fluid resistance R1 based on the impedance calculated when the machining electrodes are separated. Additionally, the inference device 15 can also infer the wiring resistance R2 based on the impedance calculated when the machining electrodes are short-circuited.

[0152] After the inference device 15 infers the electrical characteristics, the correction value determination unit 16 corrects at least one of the voltage value of the voltage applied by the power supply device 20 to the machining electrode and the current value of the current flowing through the machining electrode based on at least one of the impedance, parasitic capacitance value C1, parasitic inductance PX2, machining fluid resistance R1 and wiring resistance R2.

[0153] As described above, the electrical discharge machining apparatus 1B of Embodiment 2 infers electrical characteristics such as parasitic capacitance C1 and parasitic inductance PX2 before machining by comparing the frequency characteristics of the impedance and phase delay of the machining circuits 51 and 52 with the frequency characteristics calculated based on the circuit equations of the equivalent circuits 55 and 56 of the machining circuits 51 and 52. Therefore, the electrical discharge machining apparatus 1B can infer the electrical characteristics of the machining circuits 51 and 52 before machining without pre-machining. Furthermore, the electrical discharge machining apparatus 1B can infer the electrical characteristics of the machining circuits 51 and 52 without prior learning.

[0154] Implementation Method 3

[0155] Next, use Figure 18 as well as Figure 19 Embodiment 3 will be described. In Embodiment 3, the electrical discharge machining apparatus 1A and 1B cause the display 25 to show a screen that prompts for the correction of electrical characteristics such as the parasitic capacitance value αC and a screen indicating that the correction is complete.

[0156] In Embodiment 3, the electrical discharge machining apparatuses 1A and 1B perform the same process. Therefore, the following describes the situation where the electrical discharge machining apparatus 1B causes the display 25 to show a screen prompting for the correction of electrical characteristics and a screen indicating that the correction is complete.

[0157] The electrical characteristics of the electrical discharge machining apparatus 1B, namely the parasitic capacitance value αC and impedance, are inferred through the processing described in Embodiment 2. Figure 18 This is a diagram illustrating an example of the display screen showing the image when the electrical characteristics of the electrical discharge machining apparatus according to Embodiment 3 are deviated.

[0158] When the electrical characteristics of the electrical processing apparatus 1B deviate from the reference characteristics, the arithmetic unit 10B causes the display 25 to show a screen prompting for correction of the electrical characteristics. Specifically, the electrical processing apparatus 1B simultaneously displays both a statement 251 explaining the correction and a button 252 for performing the correction in a window WX1 on the screen. The window WX1 can be a newly displayed window or a window that is continuously displayed on the screen.

[0159] The explanatory statement regarding the correction 251 is, for example, "Correct the parasitic capacitance value with AA." or "Correct the impedance with BB." These are texts indicating the correction content of electrical characteristics.

[0160] Button 252 is used to correct deviations in electrical characteristics from reference characteristics. If button 252 is clicked, touched, or input via keyboard, the electrical discharge machining apparatus 1B corrects the electrical characteristics by adjusting the voltage and current values.

[0161] The electrical discharge machining apparatus 1B displays both a calibration instruction 251 and a calibration button 252 on the display 25, allowing the user to easily instruct on the calibration of electrical characteristics by clicking, touching, or typing on the button 252. This reduces deviations in the machining characteristics of each individual component of the electrical discharge machining apparatus 1B without requiring complex operations.

[0162] Figure 19 This diagram illustrates an example of a screen displayed on the display after the electrical characteristics of the electrical discharge machining apparatus according to Embodiment 3 have been corrected. If the electrical characteristics correction is complete, the arithmetic unit 10B of the electrical discharge machining apparatus 1B causes the display 25 to display a message 253 indicating that the correction is complete in window WX2 on the screen. Window WX2 is either a newly displayed window on the screen or a window that is continuously displayed on the screen.

[0163] Message 253 indicating that calibration is complete includes text indicating that calibration is complete and text indicating the degree of electrical deviation (text indicating the amount of deviation of the electrical characteristic). That is, message 253 includes text indicating that calibration of the deviation of the electrical characteristic from the reference characteristic has been completed and text indicating the amount of deviation between the electrical characteristic and the reference characteristic. Message 253 indicating calibration completion includes text such as "Calibration complete," "Parasitic capacitance deviation = AA," or "Impedance deviation = BB."

[0164] As described above, in Embodiment 3, the electrical discharge machining apparatus 1B displays a window WX1 on the display 25 that prompts for the correction of electrical characteristics, so that the user can easily grasp the content of the electrical characteristic correction and can easily instruct the correction of electrical characteristics.

[0165] Furthermore, the electrical discharge machining apparatus 1B displays a window WX2 on the display 25 indicating that the electrical characteristic calibration has been completed. Therefore, the user can easily ascertain that the electrical characteristic calibration is complete and can easily understand the calibration details. In other words, the user can easily understand that the electrical discharge machining apparatus 1B has been individually adjusted.

[0166] Next, the hardware structure of arithmetic devices 10A and 10B will be described. Furthermore, since arithmetic devices 10A and 10B have the same hardware structure, the hardware structure of arithmetic device 10A will be described here.

[0167] Figure 20 This diagram illustrates an example of the hardware structure of the computing device according to Embodiment 1. The computing device 10A can be implemented using an input device 300, a processor 100, a memory 200, and an output device 400. Examples of the processor 100 are CPUs (Central Processing Units, also known as central processing units, processing devices, microprocessors, microcomputers, DSPs (Digital Signal Processors)) or system LSIs (Large Scale Integration). Examples of the memory 200 are RAM (Random Access Memory) or ROM (Read Only Memory).

[0168] The arithmetic unit 10A is implemented by the processor 100 reading and executing an arithmetic program stored in the memory 200 that is executable by a computer and is used to perform the actions of the arithmetic unit 10A. The program used to perform the actions of the arithmetic unit 10A, i.e., the arithmetic program, can also be described as a program that enables the computer to execute the process or method of the arithmetic unit 10A.

[0169] The arithmetic program executed by the processor 100 has a modular structure including the arithmetic unit 10A. These structural elements are loaded onto the main storage device and generated on the main storage device. Specifically, the arithmetic program executed by the processor 100 has a modular structure including the control unit 11, the voltage instruction detection unit 12, the arithmetic unit 13, and the learning device 14. These structural elements are loaded onto the main storage device and generated on the main storage device.

[0170] The input device 300 receives user-defined information and sends it to the processor 100 or the memory 200. The memory 200 stores calculation programs, etc. Additionally, the memory 200 serves as a temporary storage area, i.e., a shared area, when the processor 100 performs various processes. The output device 400 outputs the calculation results from the processor 100 to the processing unit 2A, the power supply unit 20, the display 25, etc.

[0171] The computing program can also be stored as an installable or executable file on a computer-readable storage medium and provided as a computer program product. Alternatively, the computing program can be provided to the computing device 10A via a network such as the Internet. Furthermore, the functions of the computing device 10A can be implemented partly through dedicated hardware such as dedicated circuits and partly through software or firmware. For example, the learning device 14 in the computing device 10A can also be implemented via… Figure 20 It is implemented using the hardware structure described in the document.

[0172] The structure shown in the above embodiments is an example, and it can be combined with other known technologies, and the embodiments can be combined with each other. It is also possible to omit or change a part of the structure without departing from the spirit.

[0173] Explanation of the label

[0174] 1A, 1B Electrical Discharge Machining Apparatus; 2A, 2B Machining Unit; 3 Workpiece; 4 Tool Electrode; 8 Voltage Detection Unit; 9 Current Detection Unit; 10A, 10B Calculation Unit; 11 Control Unit; 12 Voltage Command Detection Unit; 13 Calculation Unit; 14 Learning Unit; 15 Inference Unit; 16 Correction Value Determination Unit; 20 Power Supply Unit; 25 Display; 40 Waveform; 41 Slope; 51, 52 Machining Circuit; 55, 56 Equivalent Circuit; 100 Processor; 141 Status Observation Unit; 142 Data Acquisition Unit; 143 Learning Unit; 144 Inference Unit; 200 Memory; 251 Description Statement; 252 Button; 253 Message; 300 Input Device; 400 Output Device; C1, C2, αC Parasitic Capacitance Values; Ip Inter-electrode Current Values; IZ, IZ1, IZ2 Impedance, PX1, PX2 parasitic inductance, Q charge, R1 processing fluid resistance, R2 wiring resistance, T1, T2 period, V, Vp applied voltage value, w11 to w16, w21 to w26, W1, W2 weights, WX1, WX2 window.

Claims

1. An electrical discharge machining apparatus, characterized in that, have: A power supply device that supplies power to the machining electrode space between the tool electrode and the workpiece. The current detection unit detects the current value, i.e., the inter-electrode current value, of the current flowing from the power supply device to the processing inter-electrode. The voltage command detection unit detects the command value, i.e., the applied voltage value, of the voltage applied between the processing electrodes; as well as The computing device acquires the inter-electrode current value and the applied voltage value, uses a learning model to infer the inherent electrical characteristics of the device based on the inter-electrode current value and the applied voltage value, infers the electrical characteristics based on the inter-electrode current value and the applied voltage value, and determines a correction value to correct the inter-electrode current value and the applied voltage value so that the inferred electrical characteristics become a reference electrical characteristic, i.e., a reference characteristic.

2. The electrical discharge machining apparatus according to claim 1, characterized in that, The electrical characteristics include the impedance of the processing circuit and the parasitic capacitance between the processing electrodes, the processing circuit including the power supply and the processing electrodes. The learning model includes an impedance learning model for inferring the impedance and a parasitic capacitance learning model for inferring the parasitic capacitance value. The computing device uses the impedance learning model to infer the impedance based on the inter-electrode current value, calculates the amount of electricity flowing through the processing inter-electrode based on the inter-electrode current value, and uses the parasitic capacitance learning model to infer the parasitic capacitance value based on the amount of electricity and the applied voltage value, so as to determine the correction value in such a way that the inferred impedance and the parasitic capacitance value become the reference impedance and the reference parasitic capacitance value.

3. The electrical discharge machining apparatus according to claim 1 or 2, characterized in that, The computing device acquires learning data including the inter-electrode current value, the applied voltage value, and the electrical characteristics, and uses the learning data to generate the learning model.

4. An electrical discharge machining apparatus, characterized in that, have: A power supply device that supplies power to the machining electrode space between the tool electrode and the workpiece. The current detection unit detects the current value, i.e., the inter-electrode current value, of the current flowing from the power supply device to the processing inter-electrode. The voltage detection unit detects the voltage value, i.e., the applied voltage value, of the voltage applied between the processing electrodes; as well as The computing device, based on the inter-electrode current value and applied voltage value when the processing inter-electrode is in a separated state, a first circuit equation of a first processing circuit including the power supply device and the processing inter-electrode in the separated state, the inter-electrode current value and applied voltage value when the processing inter-electrode is in a short-circuit state, and a second circuit equation of a second processing circuit including the power supply device and the processing inter-electrode in the short-circuit state, infers the inherent electrical characteristics of the device and determines a correction value for correcting the inter-electrode current value and the applied voltage value so that the inferred electrical characteristics become a reference electrical characteristic, i.e., a reference characteristic.

5. The electrical discharge machining apparatus according to claim 4, characterized in that, The inherent electrical characteristics of this device are the parasitic capacitance between the processing electrodes and the parasitic inductance in the processing circuit. The computing device calculates the frequency characteristics (i.e., the first frequency characteristic) of the first processing circuit, compares the frequency characteristics (i.e., the second frequency characteristic) of the first circuit equation with the first frequency characteristic, and thereby infers the parasitic capacitance value between the processing electrodes. The computing device calculates the frequency characteristics (i.e., the third frequency characteristic) of the second processing circuit, and compares the frequency characteristics (i.e., the fourth frequency characteristic) of the second circuit equation with the third frequency characteristic, thereby inferring the parasitic inductance on the processing circuit. The computing device determines the correction value in such a way that the parasitic capacitance and the parasitic inductance become the reference characteristics.

6. The electrical discharge machining apparatus according to claim 5, characterized in that, The first frequency characteristic refers to the frequency characteristic of the impedance of the first processing circuit and the frequency characteristic of the phase delay between the inter-electrode current value and the applied voltage value of the first processing circuit. The second frequency characteristic refers to the frequency characteristic of the impedance in the first circuit equation and the frequency characteristic of the phase delay in the first circuit equation. The third frequency characteristic is the frequency characteristic of the impedance of the second processing circuit and the frequency characteristic of the phase delay of the second processing circuit. The fourth frequency characteristic is the frequency characteristic of the impedance in the second circuit equation and the frequency characteristic of the phase delay in the second circuit equation.

7. The electrical discharge machining apparatus according to claim 6, characterized in that, The computing device calculates the first frequency characteristic based on the inter-electrode current value and the applied voltage value under the separated state, and calculates the third frequency characteristic based on the inter-electrode current value and the applied voltage value under the short-circuit state.

8. The electrical discharge machining apparatus according to any one of claims 1 to 7, characterized in that, The computing device causes a display connected to the computing device to display a button for correcting the deviation of the electrical characteristic relative to the reference characteristic and text indicating the correction content of the electrical characteristic.

9. The electrical discharge machining apparatus according to claim 8, characterized in that, The computing device causes the display to show text indicating that the correction of the deviation of the electrical characteristic relative to the reference characteristic has been completed, and text indicating how much deviation there is between the electrical characteristic and the reference characteristic.

10. A method for electrical discharge machining, characterized in that, Includes the following steps: In the power supply step, the power supply device of the electrical discharge machining apparatus supplies power to the machining electrode between the tool electrode and the workpiece. In the current detection step, the electrical discharge machining device detects the current value, i.e., the inter-electrode current value, of the current flowing from the power supply device to the machining electrodes. In the voltage command detection step, the electrical discharge machining device detects the command value, i.e., the applied voltage value, of the voltage applied between the machining electrodes. as well as The calculation steps involve the electrical discharge machining apparatus acquiring the inter-electrode current value and the applied voltage value, using a learning model to infer the inherent electrical characteristics of the apparatus based on the inter-electrode current value and the applied voltage value, inferring the electrical characteristics based on the inter-electrode current value and the applied voltage value, and determining a correction value to correct the inter-electrode current value and the applied voltage value so that the inferred electrical characteristics become the reference electrical characteristics, i.e., the reference characteristics.

11. A method for electrical discharge machining, characterized in that, Includes the following steps: In the power supply step, the power supply device of the electrical discharge machining apparatus supplies power to the machining electrode between the tool electrode and the workpiece. In the current detection step, the electrical discharge machining device detects the current value, i.e., the inter-electrode current value, of the current flowing from the power supply device to the machining electrodes. In the voltage detection step, the electrical discharge machining device detects the voltage value of the voltage applied between the machining electrodes, i.e., the applied voltage value. as well as The calculation steps involve the electrical discharge machining apparatus inferring its inherent electrical characteristics based on the inter-electrode current value and applied voltage value when the machining electrodes are in a separated state, the first circuit equation of the first machining circuit including the power supply device and the machining electrodes in the separated state, the inter-electrode current value and applied voltage value when the machining electrodes are in a short-circuit state, and the second circuit equation of the second machining circuit including the power supply device and the machining electrodes in the short-circuit state. The apparatus then determines a correction value to adjust the inter-electrode current value and the applied voltage value so that the inferred electrical characteristics become the reference electrical characteristics, i.e., the reference characteristics.

Images