CONVEYANCE ABNORMALITY DETECTING DEVICE AND CONVEYANCE ABNORMALITY DETECTING METHOD AS WELL AS METHOD FOR CONSTRUCTING MAXIMUM CURRENT PREDICTION MODEL FOR DETECTING CONVEYANCE ABNORMALITY

To enable a conveyance abnormality of a material of steel plates to be detected surely at low cost.SOLUTION: A conveyance abnormality detecting device detects a conveyance abnormality of a material of steel plates. The conveyance abnormality detecting device comprises: an obtaining part that obtains actual values of driving currents distributed respectively to a plurality of driving motors which drive a plurality of conveyance rollers that convey the material of steel plates; and a determining part that inputs a material specification parameter concerning the material of steel plates and operational parameters concerning a conveyance speed and acceleration of the material of steel plates as well as an injection quantity of cooling water, and outputs actual values of the driving currents distributed to the plurality of driving motors respectively, and determines whether or not a conveyance abnormality occurs in the material of steel plates, on the basis of a learnt maximum driving current prediction model to which machine learning is performed so that the maximum value of the driving currents is predicted in accordance with prescribed machine learning algorithm, the maximum value of the driving currents predicted by the learnt maximum driving current prediction model and the actual values of the driving currents obtained by the obtaining part.SELECTED DRAWING: Figure 3 【課題】 鋼板材料の搬送異常の検出を確実かつ低コストで実現する。【解決手段】 本発明は、鋼板材料の搬送異常を検出する搬送異常検出装置である。前記搬送異常検出装置は、前記鋼板材料を搬送する複数の搬送ローラを駆動する複数の駆動モータの各々に対する駆動電流の実績値を取得する取得部と、前記鋼板材料に関する材料諸元パラメータと前記鋼板材料の搬送速度及び加速度並びに冷却水の噴射量に関する操業パラメータとを入力とし、前記複数の駆動モータの各々に対する前記駆動電流の実績値を出力として、所定の機械学習アルゴリズムに従って前記駆動電流の最大値を予測するように機械学習が行われた学習済み最大駆動電流予測モデルと、前記学習済み最大駆動電流予測モデルにより予測される前記駆動電流の前記最大値と前記取得部により取得された前記駆動電流の前記実績値とに基づき、前記鋼板材料に搬送異常が発生したか否かを判定する判定部とを備える。【選択図】 図３