JP2019111648A5

JP2019111648A5 -

Info

Publication number: JP2019111648A5
Application number: JP2019082287A
Authority: JP
Filing date: 2019-04-23
Publication date: 2020-09-03
Anticipated expiration: 2037-03-21

Claims

A calculation formula for estimating the amount of thermal displacement of the machine element based on a measurement data group including temperature data of the machine element and its surroundings and / or operation state data of the machine element of a machine tool having a thermally expanding machine element. A machine learning device that is optimized by machine learning.
A measurement data acquisition unit that acquires the measurement data group,
A thermal displacement amount acquisition unit that acquires an actually measured value of the thermal displacement amount of the machine element,
A storage in which the measurement data group acquired by the measurement data acquisition unit is used as input data, and the measured value of the thermal displacement of the machine element acquired by the thermal displacement acquisition unit is associated with each other as a label and stored as teacher data. Department and
Thermal displacement prediction calculation that calculates the thermal displacement of the machine element based on the measurement data group by performing machine learning based on the measurement data group and the measured value of the thermal displacement of the machine element. It is equipped with a calculation formula learning unit that sets formulas.
The calculation formula learning unit is an estimated value of the thermal displacement amount of the mechanical element calculated by substituting the measurement data group within a predetermined period stored in the storage unit as teacher data into the thermal displacement amount prediction calculation formula. And, based on the difference from the measured value of the thermal displacement amount of the mechanical element within the predetermined period stored as a label in the storage unit, the thermal displacement amount prediction calculation formula is set .
The thermal displacement amount prediction calculation formula is a machine learning device that uses a plurality of time shift elements of temperature data included in the measurement data group .

The machine learning apparatus according to claim 1, wherein the thermal displacement amount prediction calculation formula uses information related to a tendency of temperature change derived from a plurality of time shift elements of the temperature data.

The measurement data acquisition unit further
A second measurement data group is acquired by adding measurement data to the measurement data group and / or excluding the measurement data from the measurement data group.
The second measurement data group is stored as input data in the storage unit.
The calculation formula learning unit further
The machine learning device according to claim 1 or 2 , wherein a second thermal displacement amount prediction calculation formula for calculating the thermal displacement amount of the machine element is set based on the second measurement data group.

Further, a contribution determination unit for determining the contribution of the measurement data included in the measurement data group to the prediction of the amount of thermal displacement is provided.
The contribution determination unit
The first is the error between the first thermal displacement predicted value calculated by the first thermal displacement prediction formula set based on the measurement data group including the measurement data for which the contribution is to be calculated and the measured thermal displacement value. The error and the second thermal displacement predicted value and the thermal displacement measured by the second thermal displacement prediction formula set based on the second measurement data group excluding the measurement data for which the contribution is calculated. The machine learning device according to claim 3 , wherein the contribution degree of the measurement data of the contribution degree calculation target is determined based on the difference between the second error and the value.

Among the currently acquired measurement data groups, the optimization measurement data selection unit that selects the optimization measurement data group consisting of the combination of measurement data with the highest accuracy using a preset number of measurement data is further added. Prepare,
The optimized measurement data selection unit
From the currently acquired measurement data group, the measurement data group obtained by removing the measurement data having the smallest contribution determined by the contribution determination unit is selected as the first measurement data group.
The measurement data group obtained by removing the measurement data having the smallest contribution determined by the contribution determination unit from the i (1 ≦ i) th measurement data group is selected as the (i + 1) th measurement data group. The machine learning device according to claim 4 , wherein an optimized measurement data group composed of a preset number of measurement data is selected by repeatedly performing the above.

The machine learning device according to any one of claims 1 to 5 , wherein the thermal displacement amount prediction calculation formula uses a first-order lag element of measurement data included in the measurement data group.

The machine learning device according to any one of claims 1 to 6, wherein the thermal displacement amount prediction calculation formula is set based on machine learning by a neural network.

The calculation formula learning unit sets a prediction calculation formula for the amount of thermal displacement based on machine learning using multiple regression analysis in consideration of the L2 regularization term, according to any one of claims 1 to 6. The machine learning device described.

The machine learning device according to any one of claims 1 to 6, wherein the calculation formula learning unit sets a prediction calculation formula for the amount of thermal displacement by using sparse regularization learning.

It is further equipped with a detection unit that detects measurement data included in the measurement data group that does not contribute to improving the accuracy of thermal displacement prediction.
The detection unit
The machine learning device according to claim 9, wherein the machine learning device detects based on the prediction calculation formula of the thermal displacement amount set by the sparse regularization learning.

The machine learning device according to any one of claims 1 to 10, which is included in the control device of the machine tool.

Corresponds to the thermal displacement amount of the machine element calculated from the measurement data group based on the thermal displacement amount prediction calculation formula set by the machine learning device according to any one of claims 1 to 11. A correction amount calculation unit that calculates the correction amount, and
A correction unit that corrects the machine position of the machine element based on the correction amount of the machine element calculated by the correction amount calculation unit.
Machine tool thermal displacement compensator equipped with.

The thermal displacement correction device according to claim 12, wherein the thermal displacement correction device is included in the control device of the machine tool.