TEMPERATURE INTERPOLATION DEVICE

Abstract

Temperature interpolation device (100), comprising: a calculation parameter setting unit (104) which records a previously selected calculation parameter for a temperature interpolation formula for the interpolation of lost temperature data; a temperature data read and write unit (102) which, during the operation of a target device whose temperature data is being read, continues to record both time information and temperature data on a non-volatile storage medium (61) and checks the time information recorded on the non-volatile storage medium (61) at predetermined time intervals to determine whether or not data has been lost in a predetermined time period; and a pre-temperature estimation unit (103) which, when determining that the data has been lost, uses the temperature data read and write unit (102) to estimate and interpolate the lost data by means of the temperature interpolation formula, using the data recorded on the non-volatile storage medium (61) and the calculation parameters recorded in the calculation parameter setting unit (104).

Description

Scope of the invention

[0001] The present invention relates to temperature interpolation devices. Related technology

[0002] A conventional method for compensating for thermal displacement is known (see, for example, publication JP 2017-144527A). Publication JP 2017-144527A discloses that "in a method for compensating for thermal displacement in a machine tool according to the present invention, a compensation value determined by performing an operation based on a trend value of measurement data for a machine temperature and a delay value of the heat transfer is used to compensate for a feed command of an NC device, and that this compensates for the position of a tool that is in the middle of machining a workpiece."

[0003] A method for compensating for the thermal displacement of a tool is also known (see, for example, publication JP 2004-148443A). Publication JP 2004-148443A discloses that "in a method for compensating for the thermal displacement of a tool according to the invention according to claim 1, a first machine tool temperature is measured near the bearing of a spindle and a second machine tool temperature is measured in a part where the thermal stability of the machine tool is high. Using a predetermined linear delay formula based on a difference between the first and second machine tool temperatures, the magnitude of the thermal displacement of the tool is estimated, and an estimated compensation value is used to compensate for the position of the tool."

[0004] However, with the conventional technologies described above, it is desirable that in the event of a loss of temperature data caused by a power outage, network disruption or the like, it is possible to refer back to the lost temperature data from the past.

[0005] On the other hand, a method for interpolating lost data is known (see, for example, publication JP 2018-120 785 A). Publication JP 2018-120 785 A discloses: “As a measure to prevent a reduction in the estimation accuracy of a deteriorated state in the event of a loss of the recording d of temperature profile data D1, it may be considered to use the temperature Tb of an accumulator 110 in a normal time period before and after a loss period to interpolate the temperature Tb during the loss period. Although different methods are provided as methods for interpolating data, it showsFig. 4 an example of linear interpolation". SUMMARY OF THE INVENTION

[0006] There is a demand for a temperature interpolation device that can appropriately interpolate temperature data during the operation of a machine tool in a period of time in which the temperature data is lost.

[0007] The problem is solved by a temperature interpolation device with the features of the independent claim. Further embodiments are the subject matter of the dependent claims.

[0008] (1) One aspect of the present disclosure relates to a temperature interpolation device comprising: a calculation parameter setting unit which records a previously selected calculation parameter for a temperature interpolation formula for the purpose of interpolating lost temperature data; a temperature data read and write unit which, during the operation of a target device whose temperature data is being read, continues to record both time information and temperature data on a non-volatile storage medium and which checks the time information recorded on the non-volatile storage medium at predetermined time intervals in order to determine whether or not data has been lost in a given time period;and a pre-temperature estimation unit which, when determining that the data has been lost, uses the temperature data read and write unit to estimate and interpolate the lost data using the temperature interpolation formula, based on the data recorded on the non-volatile storage medium and the calculation parameter recorded in the calculation parameter setting unit.

[0009] According to one aspect, it is possible to provide a temperature interpolation device that can appropriately interpolate temperature data during the operation of a machine tool in a period of time in which the temperature data is lost. BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is a diagram showing the configuration of a first embodiment; Fig. Figure 2 is a block diagram showing a temperature interpolation device, a heat displacement compensation device and a numerical control according to the first embodiment; Fig. Figure 3 is a diagram illustrating a loss of interpolated temperature data according to the first embodiment; Fig. Figure 4 is a diagram showing the recording of retrieved temperature data in a non-volatile memory according to the first embodiment; Fig. 5 is a diagram that represents a determination of whether or not a loss of temperature data has occurred according to the first embodiment; Fig. Figure 6 is a diagram illustrating the estimation and interpolation of the lost data according to the first embodiment; Fig. Figure 7 is a diagram representing a heat displacement compensation based on the temperature data recorded on the non-volatile memory, performed by a compensation displacement quantity calculation unit according to the first embodiment; Fig. Figure 8 is a diagram showing a difference in the time constant depending on the positions where temperature sensors are attached, according to a second embodiment; Fig. Figure 9 is a graph representing the difference between a temperature increase and a temperature decrease caused by a difference in the time constant in the second embodiment; and Fig. Figure 10 is a diagram illustrating the calculation of a parameter for a temperature rise curve according to the second embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0010] The following is with reference to the Fig. 1 to 7 describe a temperature interpolation device 100 according to a first embodiment of the present disclosure. Fig. Figure 1 is a diagram showing the overall configuration including a machine tool 1. Fig. Figure 2 is a block diagram showing the temperature interpolation device 100, a heat displacement compensation device 60 and a numerical control 70. Fig. Figure 3 is a diagram representing a loss of temperature data that is interpolated. Fig. Figure 4 is a diagram showing the recording of retrieved temperature data in a non-volatile memory 61. Fig. Figure 5 is a diagram that represents a determination of whether or not a loss of temperature data has occurred. Fig. Figure 6 is a diagram illustrating the estimation and interpolation of lost data. Fig. Figure 7 is a diagram that represents a heat displacement compensation based on the temperature data recorded in the non-volatile memory 61 using a compensation displacement quantity calculation unit.

[0011] First, an overview is described that includes the temperature interpolation device 100. The temperature interpolation device 100 is included in a configuration that features a heat displacement compensation function for estimating a thermal displacement using a heat displacement compensation model and temperature data. This function then adds a compensation quantity corresponding to the thermal displacement to an axis motion quantity. In this configuration, the heat displacement compensation model is created, for example, using machine learning, and the compensation quantity is calculated using the heat displacement compensation model and operating state data (such as temperature data).If a difference between an estimated value and an actually measured value does not correspond to a threshold or is smaller than it, the heat displacement compensation model is recalculated.

[0012] More precisely, as in Fig. Figure 1 shows temperature sensors 2 and displacement sensors 3 provided in the machine tool 1, making it possible to detect temperatures and displacements. The machine tool 1 includes various machine tools, such as a machining center (with an x-axis, a y-axis, and a z-axis), a lathe (which has only an x-axis and a z-axis), and the like.

[0013] The recorded temperatures and displacements are used as training data 4, and a compensation model calculation 5 is performed. The compensation model calculation 5 refers to the computational process for determining a heat displacement compensation model 5-3 by machine learning. More precisely, the compensation model calculation 5 is a computational process in which a learning software 5-1 performs machine learning 5-2 to create the heat displacement compensation model 5-3 based on the training data 4. The learning software 5-1 can be executed on a given computer. The training data 4 can be data that was previously retrieved before the work.

[0014] The specific heat displacement compensation model 5-3 is then used, and a compensation output 6 is executed during the actual work process. The compensation output 6 can be executed, for example, by a CNC device 6-2 (CNC, computerized numerical control). The CNC device 6-2 calculates an instruction (axis movement quantity) and outputs it to the machine tool 1. At this point, it uses the heat displacement compensation model 5-3 to calculate a compensation quantity and adds it to the axis movement quantity. The axis movement quantity compensated by the CNC device 6-2, as described above, is output to the machine tool 1. In this way, the machine tool 1 performs an operation according to the axis movement quantity that has undergone compensation 6-3.On the other hand, temperature sensors 2 are provided in the machine tool 1, and a temperature retrieval unit 6-1 reads the temperatures from the temperature sensors 2 and supplies them to the CNC device 6-2. Based on the supplied temperatures, the CNC device 6-2 uses the heat displacement compensation model 5-3 to perform the compensation parameter calculation output 6. Subsequently, the same processing is repeated. In this way, it is possible to perform work where interference from heat is unlikely.

[0015] More precisely, the heat displacement compensation function is described as follows: Fig. Figure 2 shows that the heat displacement compensation device 60 and the numerical control 70 implement this system. The heat displacement compensation device 60 comprises the non-volatile memory 61, which serves as the non-volatile storage medium, a compensation displacement quantity calculation unit 62, and a compensation execution unit 63. The numerical control 70 comprises an axis control unit 71. Each time the temperature data is read and retrieved by the temperature sensors 2, the temperature data is recorded in the non-volatile memory 61 in such a way that the temperature data is associated with time information. When the heat displacement compensation function is executed, the heat displacement compensation model for estimating the thermal displacement and the operating state data related to the thermal displacement are recorded.

[0016] The compensation displacement calculation unit 62 calculates the compensation value based on the operating state data (such as temperature data) and the compensation model and sends the compensation value to the compensation execution unit 63. The compensation execution unit 63 sends the compensation value corresponding to the temperatures to the axis control unit 71.

[0017] The axis control unit 71 is a unit that issues a command to the machine tool 1 and feeds the command, essentially corresponding to a machining program, into the machine tool's system. Axis control information within the machine tool's system is used for this purpose, and the execution is carried out accordingly. The compensation value corresponding to the temperatures is received by the compensation execution unit 63, added to the command, and then the command is sent to the machine tool 1.

[0018] If a loss of temperature data is caused by the temperature sensors 2 of the machine tool 1, the temperature interpolation device 100 interpolates the lost part of the data and outputs the interpolated temperature data together with the data read by the temperature sensors 2 to the heat displacement compensation device 60 at the same time.

[0019] More precisely, the temperature interpolation device comprises 100, as shown in Fig. Figure 2 shows a time information generation unit 101, a temperature data read and write unit 102, a pre-temperature estimation unit 103, and a calculation parameter setting unit 104. It is known that at predetermined parts where the temperature sensors 2 of the machine tool 1 are provided in the present embodiment, differences between the temperatures read by the temperature sensors 2 follow a cubic formula. This is used as a basic assumption, and the temperature data in a section where the temperature data has been lost are interpolated as follows.

[0020] For example, if, as in Fig. Figure 3 shows that if temperature data is lost in a section (a "section where the data was lost" in the middle) from a point in time immediately after the start of a temperature drop until a point in time immediately after the start of a subsequent temperature rise due to a power outage, network disruption, or similar event, a cubic formula is first determined using Lagrange's interpolation formula, which includes the temperature data for such a section as a given parameter. Then, using the determined cubic formula, the value of the temperature data for the section where the temperature data was lost is interpolated.

[0021] More precisely, the Lagrange interpolation formula with respect to the data (x1, y1), (x2, y2), (x3, y3), ..., (xn, yn) of n pairs of times x and temperatures y is given by the following formulas. y=∑k=1nPk(x)Pk(xk)yk Pk(x)=∏i=1n(x−xi)(x−xk)

[0022] Here, a third-order polynomial can be determined using data from four points (four pairs). Although the time points of the data elements are preferably somewhat separated, the magnitude of the temperature change is small if the time points are excessively close together, whereas the reliability of the temperature data decreases if the time points are excessively far apart. Therefore, the calculation parameter setting unit 104 records the following prerequisites for interpolation as calculation parameters used for interpolating the temperature data. [Prerequisites for interpolation] Interpolation formula: Lagrange interpolation formula Order: third order Number of data elements used for interpolation: 4 Distance between data elements used for interpolation: 30 minutes

[0023] Each time the temperature data is read and retrieved from the temperature sensors 2, the time information creation unit 101 generates the time information. The temperature data read and write unit 102 records the retrieved temperature data and the time information generated by the time information creation unit 101 in the non-volatile memory 61 of the heat displacement compensation device 60 in such a way that the temperature data and the time information are correlated. More precisely, the temperatures are, as in Fig. 4 shown, for example, at intervals of one minute, the temperature sensors 2 are read and retrieved, and the times and temperatures are compiled in pairs by the temperature data read and write unit 102 in such a way that they have a one-to-one relationship and are recorded in the non-volatile memory 61.

[0024] Simultaneously, the temperature data read and write unit 102 checks the time information in the non-volatile memory 61 at regular intervals to determine whether or not lost data has been generated. More precisely, the temperature data read and write unit 102 reads, as described in Fig. Figure 5 shows two data elements (hereinafter referred to as "latest data" and "penultimate data") that are closest to the current time, retrieved from non-volatile memory 61, and comparing the timestamps. If the interval between the timestamps of the two read data elements corresponds to a predetermined time or is greater than a predetermined temperature data retrieval interval, it is determined that (in the Fig. 3 shown section, where the data was lost) lost data was generated, and the pre-temperature estimation unit 103 is, as in Fig. 5 shown, a temperature estimation creation command was supplied.

[0025] More precisely, the temperatures in the non-volatile storage medium 61 are determined according to the Fig. 4 and Fig. The temperature sensors 2 read and retrieve the temperatures at one-minute intervals, and the times at which the temperatures are read and the read temperatures are paired by the temperature data read and write unit 102 in such a way that they have a one-to-one relationship and recorded in the non-volatile memory 61. Here, no time information or temperature data is available for the 5 minutes between a time "2018 / 10 / 23 19:02" and a time "2018 / 10 / 23 19:07". If, during the 5 minutes in which the interval is so large that it exceeds the specified interval of 1 minute and corresponds to 2 minutes or more, a loss of temperature data occurs, as described above, the temperature data read and write unit 102 determines, as in Fig. Figure 5 shows that data loss has occurred, which results in the temperature data read and write unit 102 sending the pre-temperature estimation unit 103 the estimation temperature creation command. The following formula, for example, is used as the determination formula: Tt−Tt−1≥SamplingRate+thresh (Tt: time of last data, SamplingRate: data retrieval interval, thresh: arbitrary threshold).

[0026] If the temperature data read and write unit 102 determines that data loss has occurred, as in Fig. As shown in Figure 6, the pre-temperature estimating unit 103 uses the times and temperature data remaining in the non-volatile memory 61, and the computation parameters recorded in the non-volatile memory 61, to estimate the temperature data for the lost time using the Lagrange interpolation formula. The pre-temperature estimating unit 103 then instructs the temperature data read and write unit 102 to record the estimated temperature data in the non-volatile memory 61. In this way, the temperature data read and write unit 102 records the temperature data for the lost time in the non-volatile memory 61 for interpolation.

[0027] In the configuration described above, the system detects that the temperature data collected by the temperature sensors 2 was lost during the specified time period. Therefore, the temperature data for the specified time period in which the data was lost is interpolated. Subsequently, the compensation displacement calculation unit 62 estimates the thermal displacement based on the temperature data collected by the temperature sensors 2, the temperature data interpolated by the temperature interpolation device 100, and the heat displacement compensation model, as described in Fig. 7 shown, and thereby the compensation execution unit 63 performs the heat displacement compensation function of adding the compensation quantity corresponding to the thermal displacement to the axis movement quantity.

[0028] The present embodiment described above achieves the following results. In this embodiment, the temperature data read and write unit 102 continues recording both time information and temperature data in the non-volatile memory 61 during the operation of the machine tool 1, which serves as the target device whose temperature data is being read. To determine whether data has been lost during a given time period, the unit checks the time information recorded in the non-volatile memory 61 at predetermined time intervals. In this way, it is possible to interpolate the temperature data during the operation of the machine tool 1 for the time period in which the temperature data was lost.

[0029] In the present embodiment, the pre-temperature estimation unit 103 uses the Lagrange interpolation formula as a temperature interpolation formula to estimate the lost data based on the temperature data, using the calculation parameters. In this way, it is possible to interpolate the temperature data with high accuracy when a temperature rise and fall follow a cubic formula, taking into account the thermal properties of parts of the machine tool 1 on which temperature sensors 2 are provided, thus enabling a highly accurate interpolation of a thermal displacement.

[0030] A second embodiment of the present disclosure is then described. The second embodiment differs from the first in that the lost data is interpolated using an exponential function when the time constant of a temperature change is high, depending on the position of the machine tool 1 on which the temperature sensors 2 are provided. Since the remaining configurations of the second embodiment are identical to those of the first embodiment, a description of the same configurations as those according to the first embodiment is omitted. Fig. Figure 8 is a diagram showing a difference in the time constant depending on the positions where the temperature sensors 2 are attached. Fig. Figure 9 is a graph that represents the difference between a temperature increase and a temperature decrease caused by the difference in the time constant. Fig. Figure 10 is a diagram that illustrates the calculation of a parameter for a temperature rise curve.

[0031] As in Fig. As shown in Figure 8, machine tool 1 has a part 11 with a high time constant and a part 12 with a low time constant. As shown in Fig. As shown in Figure 9, the shape of a curve indicating a temperature change varies considerably depending on the magnitude of the time constant. In particular, the displacement values ​​determined by the displacement sensors 3 change when, for example, the user of machine tool 1 stops and then restarts its operation. Therefore, taking the time constant into account is necessary for a highly accurate estimation of the lost temperature data.

[0032] More precisely, temperature rise data are retrieved from each of several temperature sensors 2 of the machine tool 1, and a decay constant λ is determined. As in Fig. As shown in Figure 10, it is assumed that a saturation temperature value at a temperature increase is 100% and that the time interval until the temperature reaches a temperature of 63.2% is τ, and therefore the decay constant λ, which is a positive number, is determined using the following relational formula: λ=1 / t. The determined value λ is recorded in advance in the calculation parameter setting unit 104 in such a way that it is assigned to the individual temperature sensors 2.

[0033] It is known that in the specified section where the temperature sensor 2 of the machine tool 1 is provided in the present embodiment and whose time constant is high, the temperature transition read by the temperature sensor 2 is generally an exponential temperature transition. When the temperature drops, this process is mathematically expressed, for example, by the following differential equation: dN / dt=−λN(t), where N(t) is a reduction quantity at time t. The differential equation is solved, resulting in the following exponential function: N(t)=N0e-λt where N0 is an initial value. The differential equation is transformed into the following equation: N(t)=N0e-λt+N∞ where it is assumed that N ∞ The temperature limit is...

[0034] The initial value N0 and the limit N ∞are defined as follows. N0 = Temperature of the last data point − Temperature of the penultimate data point N∞=Temperature of the penultimate data point − (Temperature of the new data point after the last data point)

[0035] The present embodiment described above achieves the following result. In this embodiment, the pre-temperature estimation unit 103 uses the exponential function as a temperature interpolation formula to estimate the lost data based on the previously retrieved temperature rise data. This calculation parameter is based on the time interval until a temperature with a predetermined ratio relative to the saturation temperature is reached from an initial rise temperature of 0°C. In this way, the lost temperature data can be estimated not only when the temperature decreases (falls) but also when it increases (rises), resulting in a smooth curve.

[0036] The present embodiments have been described above. Although the embodiments described above are preferred embodiments, the present invention is not limited to only the embodiments described above, and embodiments with various modifications can be implemented in practice. For example, the variants described below can be implemented in practice.

[0037] In particular, the configurations of the calculation parameter setting unit, the temperature data read and write unit, the pre-temperature estimation unit, and the like are not limited to the configurations of the calculation parameter setting unit 104, the temperature data read and write unit 102, the pre-temperature estimation unit 103, and the like according to the embodiments described above. The configurations of the heat displacement compensation device for performing the heat displacement compensation function, the numerical control, and the like are not limited to the configurations of the heat displacement compensation device 60, which includes the non-volatile memory 61, the compensation displacement quantity calculation unit 62, and the compensation execution unit 63, the numerical control 70, which includes the axis control unit 71, and the like.Although the Lagrange interpolation formula is used as the temperature interpolation formula, there is no restriction to this configuration. Although the temperature data follow the cubic or exponential function, there is no restriction to this configuration. Although the non-volatile memory 61 is used as the non-volatile storage medium in the embodiments described above, there is no restriction to this configuration. Although the initial rise temperature value is 0°C in the embodiments described above, there is no restriction to this configuration. EXPLANATION OF THE REFERENCE SYMBOLS 61 Non-volatile memory (non-volatile storage medium) 100 Temperature interpolation device 102 Temperature data read and write unit 103 Pre-temperature estimation unit 104 Calculation parameter setting unit

Claims

[1] Temperature interpolation device (100) comprising: a calculation parameter setting unit (104) which records a previously selected calculation parameter for a temperature interpolation formula for the interpolation of lost temperature data; a temperature data read and write unit (102) which, during the operation of a target device whose temperature data is being read, continues to record both time information and temperature data on a non-volatile storage medium (61) and checks the time information recorded on the non-volatile storage medium (61) at predetermined time intervals to determine whether or not data has been lost in a predetermined time period; and a pre-temperature estimation unit (103) which, when determining that the data has been lost, uses the temperature data read and write unit (102) to estimate and interpolate the lost data by means of the temperature interpolation formula, using the data recorded on the non-volatile storage medium (61) and the calculation parameters recorded in the calculation parameter setting unit (104). [2] Temperature interpolation device (100) according to claim 1, wherein the temperature data read and write unit (102) uses the Lagrange interpolation formula as the temperature interpolation formula to estimate the lost data based on the temperature data using the calculation parameter. [3] Temperature interpolation device (100) according to claim 1, wherein the temperature data read and write unit (102) uses an exponential function as a temperature interpolation formula to estimate the lost data from the temperature data when the temperature rises based on previously retrieved temperature rise data, using the calculation parameter based on a time interval until a temperature with a predetermined ratio with respect to a saturation temperature value is reached from an initial rise temperature value.

Images