A thermal displacement compensation device and a computer-readable recording medium on which the program is stored.

Abstract

A thermal displacement correction device according to the present invention comprises: a temperature data acquisition unit that acquires temperature data pertaining to the temperature of each part of a machine; a displacement data acquisition unit that acquires displacement data pertaining to the amount of thermal displacement of the machine; a temperature data storage unit that stores temperature data used in machine learning; a temperature data comparison unit that compares the temperature data measured by the temperature data acquisition unit and the temperature data stored by the temperature data storage unit; and a determination unit that, on the basis of the result of comparison by the temperature data comparison unit, determines whether it is necessary to measure the amount of thermal displacement of the machine using the displacement data acquisition unit.

Description

Technical Field

[0001] The present invention relates to a thermal displacement correction device and a computer-readable recording medium recording a program.

Background Art

[0002] Thermal deformation of manufacturing machines such as machine tools is one of the factors that hinder machining accuracy. When performing machining on a manufacturing machine, in order to ensure machine accuracy, it is necessary to correct thermal deformation due to temperature changes considering internal heat sources such as motors of the manufacturing machine and external heat sources such as sunlight, lighting, and air conditioning.

[0003] Techniques for predicting the amount of thermal deformation due to temperature changes based on such heat sources and reducing the influence of thermal deformation by correcting the feed amount in each coordinate axis direction have been widely known. For example, as a technique for correcting thermal displacement due to heat generation of a manufacturing machine, a machine learning model for estimating the amount of thermal displacement of each part of the machine tool based on the temperature state of the machine tool detected by a temperature sensor is created, and the thermal displacement is corrected using the machine learning model (for example, Patent Document 1, etc.).

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] Models for estimating thermal displacement are generated, for example, by manufacturers of manufacturing machinery. These manufacturers might install their machinery in a variable-temperature chamber to reproduce various temperature environments and collect training data for model generation. They then generate the model using machine learning based on the collected data. However, if the temperature environment at the manufacturing site differs from what was anticipated, the model generated by the manufacturer may not be able to estimate thermal displacement with sufficient accuracy. In such cases, it becomes necessary to collect training data at appropriate times within the manufacturing site and perform additional training using this data.

[0006] The training data used for additional training requires temperature data, including the temperature of each part of the manufacturing machine and the ambient temperature of the installation environment, as well as displacement data, including the displacement of the manufacturing machine. Figure 5 is a schematic diagram showing an example of a manufacturing machine used when creating training data. The temperature of each part of the manufacturing machine can be obtained by measuring it with temperature sensors attached to each part of the machine. The displacement of the manufacturing machine can be obtained by measuring the position of the measurement reference point with a touch probe attached to the main spindle. Additional training is then performed using the temperature data and displacement data obtained in this way.

[0007] Temperature data can be measured even during workpiece machining using temperature sensors attached to the manufacturing machine. However, displacement data cannot be measured without interrupting workpiece machining and attaching a touch probe in place of the tool. In other words, additional learning during actual operation requires stopping the machining process. This presents the problem of increased cycle time. Therefore, there is a need for a more efficient way to update the model.

[0008] Patent Document 1 describes a method for determining whether a model needs to be updated based on adjustment values ​​calculated from correction values ​​obtained by actually running the machining program. Additional measurements are performed only if it is determined that a model update is necessary. This reduces the number of additional measurements required, thereby streamlining the model update process. However, since the adjustment values ​​in this technology vary depending on the machining environment and machining program, actual measurements must be taken to determine the adjustment values ​​when changing the machining environment or machining program. Therefore, even when applied to manufacturing machines where the machining environment and machining program are frequently changed, the effect of improving the efficiency of model updates is not fully realized. [Means for solving the problem]

[0009] The thermal displacement compensation device described in this disclosure stores pre-learned temperature data in a database. During actual operation of the manufacturing machine, the temperature measured by the temperature sensor is compared with the temperature data stored in the database. If similar temperature data exists, displacement measurement is not performed. Only when no similar temperature data exists is displacement measurement performed, and additional learning is carried out to solve the above problem.

[0010] Furthermore, one aspect of the present disclosure is a thermal displacement correction device having a function for estimating and correcting the thermal displacement of a machine by modeling the relationship between temperature and thermal displacement using machine learning, comprising: a temperature data acquisition unit that acquires temperature data relating to the temperature of each part of the machine; a displacement data acquisition unit that acquires displacement data relating to the amount of thermal displacement of the machine; a temperature data storage unit that stores the temperature data used in the machine learning; a temperature data comparison unit that compares the temperature data measured by the temperature data acquisition unit with the temperature data stored in the temperature data storage unit; and the comparison result of the temperature data comparison unit. If they are not similar, the displacement data acquisition unit measures the thermal displacement of the machine and updates the model. This is a thermal displacement compensation device equipped with a judgment unit that makes decisions.

[0011] Other aspects of the present disclosure include a computer-readable recording medium that records a program for operating a computer as a thermal displacement correction device having a function related to thermal displacement correction that estimates and corrects the thermal displacement of a machine by modeling the relationship between temperature and thermal displacement by machine learning, the recording medium comprising: a temperature data acquisition unit for acquiring temperature data relating to the temperature of each part of the machine; a displacement data acquisition unit for acquiring displacement data relating to the amount of thermal displacement of the machine; a temperature data storage unit for storing the temperature data used in the machine learning; a temperature data comparison unit for comparing the temperature data measured by the temperature data acquisition unit with the temperature data stored in the temperature data storage unit; and the comparison result of the temperature data comparison unit. If they are not similar, the displacement data acquisition unit measures the thermal displacement of the machine and updates the model. It is a computer-readable recording medium that stores a program that operates the computer, acting as a decision-making unit. [Effects of the Invention]

[0012] One aspect of this disclosure makes it possible to update a model for estimating thermal displacement more efficiently. [Brief explanation of the drawing]

[0013] [Figure 1] This is a hardware configuration diagram of a thermal displacement compensation device according to one embodiment. [Figure 2] Block diagram showing the function of a thermal displacement compensation device according to one embodiment. [Figure 3] This is a schematic diagram of a manufacturing machine equipped with a temperature sensor. [Figure 4] This diagram illustrates a temperature data set stored in the temperature data storage unit. [Figure 5] This is a schematic diagram showing an example of a manufacturing machine used when creating training data. [Modes for carrying out the invention]

[0014] Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a schematic hardware configuration diagram showing a main part of a thermal displacement correction device according to an embodiment of the present invention. The thermal displacement correction device 1 of the present invention can be mounted on a control device that controls a manufacturing machine 2 such as a machine tool. Further, the thermal displacement correction device 1 can be mounted on a personal computer provided together with a control device that controls the manufacturing machine 2, a personal computer connected to the control device via a wired / wireless network (not shown), a cell computer, a fog computer, a cloud server, or the like. In the present embodiment, an example in which the thermal displacement correction device 1 is mounted on a control device that controls the manufacturing machine 2 is shown.

[0015] The CPU 11 included in the thermal displacement correction device 1 according to the present embodiment is a processor that controls the thermal displacement correction device 1 as a whole. The CPU 11 reads out a system program stored in the ROM 12 via the bus 22 and controls the entire thermal displacement correction device 1 according to the system program. Temporary calculation data, display data, and various data input from the outside are temporarily stored in the RAM 13.

[0016] The non-volatile memory 14 is composed of, for example, a memory backed up by a battery not shown, an SSD (Solid State Drive), or the like, and the stored state is maintained even when the power of the thermal displacement correction device 1 is turned off. In the non-volatile memory 14, a machine control program and data read from an external device 72 via the interface 15, a machine control program and data input via the input device 71, a machine control program and data acquired from the manufacturing machine 2, and the like are stored. The data stored in the non-volatile memory 14 may be expanded in the RAM 13 at the time of execution / use. Further, various system programs such as a known analysis program are written in advance in the ROM 12.

[0017] The interface 15 is an interface for connecting the CPU 11 of the thermal displacement correction device 1 and an external device 72 such as a USB device. From the external device 72 side, for example, a control program stored in advance, data related to the operation of each manufacturing machine 2, etc. can be read. Also, the control program, setting data, etc. edited within the thermal displacement correction device 1 can be stored in external storage means via the external device 72. The PLC (Programmable Logic Controller) 16 performs input / output of signals and controls via the I / O unit 17 between the built-in sequence program in the thermal displacement correction device 1 and devices such as the manufacturing machine 2, peripheral devices of the manufacturing machine 2 (not shown), and sensors attached to the manufacturing machine 2.

[0018] The thermal displacement correction device 1 is connected to sensors such as a temperature sensor 3 attached to the manufacturing machine 2 and a touch probe 4 used for measuring thermal displacement. The temperature sensor 3 is used for measuring the temperature of each part of the manufacturing machine 2 and the ambient temperature. Also, the touch probe 4 is used for measuring the displacement amount of each part of the manufacturing machine 2.

[0019] To the display device 70, data such as each data read onto the memory, data obtained as a result of executing a program, etc. are output and displayed via the interface 18. Also, an input device 71 composed of a keyboard, a pointing device, etc. passes commands, data, etc. based on operations by an operator to the CPU 11 via the interface 19.

[0020] The axis control circuit 30, which controls each axis of the manufacturing machine 2, receives the axis movement command amount from the CPU 11 and outputs the axis command to the servo amplifier 40. The servo amplifier 40 receives this command and drives the servo motors 50 that move the axes of the manufacturing machine 2. The axis servo motors 50 have built-in position and speed sensors, and the position and speed feedback signals from these position and speed sensors are fed back to the axis control circuit 30 to perform position and speed feedback control. In the hardware configuration diagram of Figure 1, only one axis control circuit 30, servo amplifier 40, and servo motor 50 are shown, but in reality, as many as the number of axes of the manufacturing machine 2 to be controlled (for example, 3 for a manufacturing machine 2 with 3 linear axes, and 5 for a 5-axis machining center) are provided.

[0021] The spindle control circuit 60 receives a spindle rotation command to the spindle of the manufacturing machine 2 and outputs a spindle speed signal to the spindle amplifier 61. The spindle amplifier 61 receives this spindle speed signal and rotates the spindle motor 62 of the spindle at the commanded rotational speed to drive the tool. A position coder 63 is coupled to the spindle motor 62, and the position coder 63 outputs a feedback pulse synchronized with the rotation of the spindle, which is read by the CPU 11.

[0022] Interface 21 is an interface for connecting the thermal displacement compensation device 1 and the machine learning machine 100. The machine learning machine 100 includes a processor 101 that controls the entire machine learning machine 100, a ROM 102 that stores system programs, a RAM 103 for temporary storage in each process related to machine learning, and a non-volatile memory 104 used for storing learning models, etc. The machine learning machine 100 can observe various information obtainable by the thermal displacement compensation device 1 via interface 21 (for example, temperature data of each part of the manufacturing machine 2, ambient temperature data, displacement data of each part of the manufacturing machine 2, etc.). The thermal displacement compensation device 1 also receives information output from the machine learning machine 100 and performs actions such as correcting the amount of thermal displacement compensation, controlling the manufacturing machine 2, displaying on the display device 70, and transmitting information to other devices via a network (not shown).

[0023] Figure 2 is a schematic block diagram showing the functions of a thermal displacement compensation device 1 according to one embodiment of the present invention. Each function of the thermal displacement compensation device 1 according to this embodiment is realized by the CPU 11 of the thermal displacement compensation device 1 and the processor 101 of the machine learning machine 100, respectively, executing system programs and controlling the operation of each part of the thermal displacement compensation device 1 and the machine learning machine 100.

[0024] The thermal displacement correction device 1 of this embodiment includes a temperature data acquisition unit 110, a displacement data acquisition unit 120, a temperature data comparison unit 130, a judgment unit 140, an output unit 150, a learning unit 160, and an estimation unit 170. Furthermore, a temperature data storage unit 210, which is an area for storing temperature data used to generate a model by machine learning, is pre-prepared on the RAM 13 to non-volatile memory 14 of the thermal displacement correction device 1. In addition, a model storage unit 220 is pre-prepared on the RAM 103 to non-volatile memory 104 of the machine learning machine 100, which stores a model generated by performing machine learning using the temperature data stored in the temperature data storage unit 210 and the displacement data corresponding to the temperature data.

[0025] The temperature data acquisition unit 110 acquires a set DS of temperature data relating to the temperature of each part of the manufacturing machine 2. This temperature data set DS may include the temperature of the environment in which the manufacturing machine 2 is installed. The temperature data acquisition unit 110 may acquire temperature data, for example, from temperature sensors 3 attached to the manufacturing machine 2. This temperature data may be acquired when the manufacturing machine 2 is stopped, but it is preferable to acquire it when it is running. This is to acquire the differences in temperature of each part of the manufacturing machine 2 due to the contents of the processing program. As illustrated in Figure 3, it is preferable that multiple temperature sensors 3 are attached to each part of the manufacturing machine 2. The temperature data acquired by the temperature data acquisition unit 110 may be the instantaneous value of the temperature measured by each temperature sensor 3 at a predetermined point in time, but it is preferable that it be time-series data measured by the temperature sensors 3 over a predetermined period. This is because the amount of thermal displacement of each part of the manufacturing machine 2 is caused by the change in temperature of each part over a predetermined period in the past. Thus, it is preferable that the temperature data acquisition unit 110 acquires multiple temperature data measured over a predetermined period of time, representing the changes in ambient temperature and the temperature of each part of the manufacturing machine 2.

[0026] The displacement data acquisition unit 120 acquires displacement data related to the thermal displacement of each part of the manufacturing machine 2. The displacement data acquisition unit 120 may automatically measure the thermal displacement of each part of the manufacturing machine 2 in response to a command to acquire displacement data and acquire it as displacement data. Alternatively, the operator may manually measure the thermal displacement of each part of the manufacturing machine 2 by operating the manufacturing machine 2 and acquire it as displacement data. When acquiring displacement data related to the thermal displacement of the manufacturing machine 2, the displacement data acquisition unit 120 displays a message, for example, on a display device 70, instructing the operator to stop the processing operation of the manufacturing machine 2 and measure the thermal displacement. Upon receiving this message, the operator stops the processing operation of the workpiece by the manufacturing machine 2 and performs the thermal displacement measurement. Once the thermal displacement measurement is completed, the displacement data acquisition unit 120 acquires displacement data related to the thermal displacement of the manufacturing machine 2 based on the work results. If the manufacturing machine 2 is a machine tool, the thermal displacement of the manufacturing machine 2 can be measured by attaching a touch probe to the spindle and detecting a measurement reference with the touch probe, as illustrated in Figure 5. Such work cannot be performed while the workpiece is being processed. Therefore, the measurement of thermal displacement is performed by temporarily stopping the workpiece processing operation of manufacturing machine 2.

[0027] The temperature data comparison unit 130 compares the temperature data set DS acquired by the temperature data acquisition unit 110 with the temperature data set stored in the temperature data storage unit 210 at a predetermined timing. The predetermined timing may be at predetermined intervals or at any time instructed by the operator. The temperature data comparison unit 130 outputs the comparison result between the temperature data sets to the judgment unit 140.

[0028] The temperature data storage unit 210 stores multiple temperature datasets that were used in the machine learning process when generating the model to be stored in the model storage unit 220. Figure 4 shows an example of a temperature dataset stored in the temperature data storage unit 210. In the example in Figure 4, the temperature data Et of the installation environment of the manufacturing machine 2 is stored. j (t) and the temperature data T of each part of the machine measured by each temperature sensor 1 to i. 1j (t)~Temperature data Tij Datasets DS1 to DSj, including (t), are stored in the temperature data storage unit 210. In this case, the temperature data comparison unit 130 calculates the similarity between, for example, the temperature data set DS acquired by the temperature data acquisition unit 110 and each of the temperature data sets stored in the temperature data storage unit 210. For the similarity between individual time series data, well-known methods such as CCF (Cross Correlation Function) or DTW (Dynamic Time Warping) can be used. The similarity between the datasets can then be calculated by calculating statistical values ​​such as the mean for the similarity between each of the time series data that make up the dataset. Then, the highest similarity among the respective similarities Sim1 to Simj can be used as the comparison result.

[0029] The determination unit 140 determines, based on the comparison results between temperature datasets input from the temperature data comparison unit 130, whether it is necessary to measure the thermal displacement of the manufacturing machine 2 by the displacement data acquisition unit 120, and whether it is necessary to update the model stored in the model storage unit 220. For example, if the temperature dataset DS acquired by the temperature data acquisition unit 110 is not similar to the group of temperature datasets stored in the temperature data storage unit 210, the determination unit 140 determines that it is necessary to measure the thermal displacement of the manufacturing machine 2 and update the model. At this time, the determination unit 140 determines, for example, a predetermined threshold value (Sim) set in advance. th Then, it compares this with the similarity input from the temperature data comparison unit 130, and calculates the threshold. thThe system should determine that additional training is required for the model stored in the model storage unit 220 if the following conditions are met. If the decision unit 140 determines that it is necessary to measure the thermal displacement of the manufacturing machine 2, it instructs the displacement data acquisition unit 120 to measure the thermal displacement of the manufacturing machine 2 according to the operator's instructions. If the decision unit 140 determines that it is necessary to update the model, it instructs the learning unit 160 of the machine learning machine learning modeler 100 to perform machine learning using the set of temperature data acquired by the temperature data acquisition unit 110 and the displacement data acquired by the displacement data acquisition unit 120, according to the operator's instructions, and update the model. The decision unit 140 outputs the decision result to the output unit 150.

[0030] The output unit 150 outputs the result of the judgment made by the judgment unit 140 to the display device 70. At this time, the output unit 150 may also display the similarity between each dataset compared by the temperature data comparison unit 130 on the display device 70, in addition to the judgment result of the judgment unit 140. Furthermore, depending on the relationship between the similarity and the threshold, the operator may decide whether to measure the thermal displacement of the manufacturing machine 2 or update the model based on these displays. The output unit 150 may also output the result of the judgment made by the judgment unit 140 to the RAM 13 or non-volatile memory 14 or to an external device 72. The output unit 150 may also transmit the result of the judgment made by the judgment unit 140 to another computer such as a fog computer or host computer via a network (not shown).

[0031] Based on commands from the decision unit 140, the learning unit 160 performs machine learning processing using the temperature dataset acquired by the temperature data acquisition unit 110 and the displacement data acquired by the displacement data acquisition unit 120, and updates the model stored in the model storage unit 220. The temperature dataset used by the learning unit 160 for training is additionally stored in the temperature data storage unit 210.

[0032] During the processing operation of the manufacturing machine 2, the estimation unit 170 uses the temperature data set acquired by the temperature data acquisition unit 110 to perform thermal displacement estimation processing based on the model stored in the model storage unit 220. The thermal displacement estimated by the estimation unit 170 is used in the thermal displacement correction processing during the processing operation of the manufacturing machine 2.

[0033] In the thermal displacement correction device 1 according to this embodiment, which has the above configuration, the ambient temperature and the temperature of each part of the manufacturing machine 2 are measured during actual operation. These are then compared with a temperature dataset that has already been used to train the model. If a similar temperature dataset exists, it is assumed that the current temperature conditions are already reflected in the model. Even if the amount of thermal displacement of the manufacturing machine 2 is measured and trained, an improvement in the model's performance cannot be expected. Furthermore, there is a risk of overfitting. Therefore, in such cases, it is not necessary to stop the processing operation of the manufacturing machine 2 to measure the amount of thermal displacement. On the other hand, if there is no similar temperature dataset, it is highly likely that the current temperature conditions are not already reflected in the model. By acquiring current data and updating the model using the acquired data, an improvement in the model's performance can be expected. In this way, the thermal displacement correction device 1 according to this embodiment makes it possible to update the model for estimating the amount of thermal displacement more efficiently.

[0034] Ideally, a model for estimating the thermal displacement of manufacturing machinery should be generated before actual operation by, for example, installing the manufacturing machinery in a temperature-variable chamber, measuring the thermal displacement under multiple temperature conditions, and then performing machine learning using the obtained data. However, if the manufacturing machinery is large, it may not be possible to place it in a temperature-variable chamber, making it difficult to create a model using data obtained under multiple temperature conditions before actual operation. In such cases, data for machine learning must be collected at appropriate intervals while the manufacturing machinery is in actual operation on the factory floor, and the model updated each time. Even in such cases, since processing operations must be stopped to collect data, it is necessary to collect data to efficiently update the model without unnecessarily reducing the cycle time. In such cases as well, the thermal displacement compensation device according to this embodiment functions effectively.

[0035] As one modification of the thermal displacement correction device 1 according to this embodiment, the temperature data comparison unit 130 may use machine learning techniques when comparing the temperature dataset DS acquired by the temperature data acquisition unit 110 with the temperature dataset stored in the temperature data storage unit 210. More specifically, for example, a known cluster analysis technique may be used to create clusters of existing datasets from the temperature datasets stored in the temperature data storage unit 210, and the proximity of the distance between the temperature data acquisition unit 110 and the clusters may be considered as similarity for comparison. Then, if the distance from the cluster is greater than or equal to a predetermined threshold, the judgment unit 140 should determine that it is necessary to measure the thermal displacement of the manufacturing machine 2 and update the model. Known techniques such as regression analysis and k-means can be used for cluster analysis.

[0036] Although embodiments of the present invention have been described above, the present invention is not limited to the examples of embodiments described above, and can be implemented in various forms by making appropriate modifications. [Explanation of Symbols]

[0037] 1. Thermal displacement compensation device 2 Manufacturing machinery 3. Temperature sensor 4 Touch probes 11 CPU 12 ROM 13 RAM 14 Non-volatile memory 15,18,19,21 Interface 16 PLC 17 I / O Units 22 buses 30-axis control circuit 40 Servo Amplifier 50 servo motors 60 Spindle Control Circuit 61 Spindle Amplifier 62 Spindle motors 63 Position Coder 70 Display device 71 Input device 72 External equipment 100 machine learning machines 101 Processors 102 ROM 103 RAM 104 Non-volatile memory 110 Temperature data acquisition unit 120 Displacement data acquisition unit 130 Temperature data comparison section 140 Judgment Department 150 Output section 160 Learning Department 170 Estimation Department 210 Temperature data storage unit 220 Model Memory Unit

Claims

1. A thermal displacement correction device having a function related to thermal displacement correction that estimates and corrects the thermal displacement of a machine by modeling the relationship between temperature and thermal displacement using machine learning, A temperature data acquisition unit that acquires temperature data relating to the temperature of each part of the machine, A displacement data acquisition unit acquires displacement data related to the thermal displacement of the machine, A temperature data storage unit that stores the temperature data used in the aforementioned machine learning, A temperature data comparison unit compares the temperature data measured by the temperature data acquisition unit with the temperature data stored in the temperature data storage unit. A determination unit determines, if the comparison results from the temperature data comparison unit are not similar, to measure the thermal displacement of the machine using the displacement data acquisition unit and update the model. A thermal displacement compensation device equipped with the following features.

2. The aforementioned temperature data includes not only the temperature of each part of the machine, but also the temperature of the environment in which the machine is installed. The thermal displacement compensation device according to claim 1.

3. The determination unit makes a determination based on the similarity between the temperature data measured by the temperature data acquisition unit and the temperature data stored in the temperature data storage unit. The thermal displacement compensation device according to claim 1.

4. The aforementioned similarity is calculated using a machine learning method with the temperature data stored in the temperature data storage unit as training data. The thermal displacement compensation device according to claim 3.

5. A computer-readable recording medium containing a program that operates a computer as a thermal displacement correction device having a function related to thermal displacement correction that estimates and corrects the thermal displacement of a machine by modeling the relationship between temperature and thermal displacement using machine learning, A temperature data acquisition unit that acquires temperature data relating to the temperature of each part of the machine. A displacement data acquisition unit that acquires displacement data related to the thermal displacement of the machine. A temperature data storage unit that stores the temperature data used in the aforementioned machine learning, A temperature data comparison unit compares the temperature data measured by the temperature data acquisition unit with the temperature data stored in the temperature data storage unit. A determination unit determines, if the comparison results from the temperature data comparison unit are not similar, that the displacement data acquisition unit should measure the thermal displacement of the machine and update the model. A computer-readable recording medium that stores programs used to operate a computer.

Images