Method for creating training data

Abstract

To provide a teacher data creation method which can automatically create correct and large amounts of teacher data for predicting the breakout of a continuous casting apparatus with Supervised Learning.SOLUTION: Time sequence molten metal face level data detected by molten metal level detection mean provided at a tundish 2 are differentiated with respect to time sequence temperature data detected by temperature detection means 4 installed in a mold 3 of a continuous casting apparatus to determine the data type of the time sequence temperature data, further, time sequence casting speed data detected by casting speed detection means are differentiated to determine the starting point and / or the terminal point of a constant region of the data type, and further, a region in which the casting speed data is reduced to a prescribed value is determined as a break out part by an existing break out detection device and the time sequence temperature data are labeled to create teacher data.SELECTED DRAWING: Figure 3

Claims

[Claim 1] A method for creating training data to be given to an AI in order to predict breakout in a continuous casting apparatus by the following processing performed on a computer, wherein, with respect to time-series temperature data detected by a temperature detection means installed in the mold of the continuous casting apparatus, among the patterns of change in time-series molten metal level data detected by a molten metal level detection means installed in the mold, the data portion of the time-series temperature data where the time-series molten metal level data is almost constant is determined to be the "steady portion", if the time-series molten metal level data shows a large value at the beginning of the time series, the data portion of the time-series temperature data is determined to be the "rising portion" preceding the "steady portion", and if the time-series molten metal level data shows a large value at the end of the time series, the data portion of the time-series temperature data is determined to be the "steady portion" A method for creating training data, wherein the following points are determined to be subsequent "falling edges," and the point at which the time-series casting speed data detected by the casting speed detection means changes abruptly is defined as the starting point which is the boundary between the "steady-state section" and the "rising edge section," and / or the ending point which is the boundary between the "steady-state section" and the "falling edge section," and a label is assigned to the time-series temperature data that can identify each data section which includes at least the data section of the "steady-state section" and at least one of the data sections of the "rising edge section" and the "falling edge section" may follow before and after it, and a label is assigned to the data section of the time-series temperature data after the point at which the casting speed data is reduced to a predetermined value by the existing breakout detection device, in place of the "falling edge section," to identify the "breakout section."

Images