MOLD ANOMALY DETECTION DEVICE, LEARNING MODEL GENERATION DEVICE AND PROGRAM FOR THE SAME

• Main Authors: HISAMITSU KAN, FURUTA KATSUMI, KATENA KENICHI, FUJIWARA KOICHI
• Format: Patent
• Language: English; Japanese
• Published: 24.06.2021
• Subjects: MEASUREMENT OF MECHANICAL VIBRATIONS OR ULTRASONIC, SONIC ORINFRASONIC WAVES; MEASURING; MECHANICAL METAL-WORKING WITHOUT ESSENTIALLY REMOVINGMATERIAL; PERFORMING OPERATIONS; PHYSICS; PUNCHING METAL; TESTING; TESTING STATIC OR DYNAMIC BALANCE OF MACHINES ORSTRUCTURES; TESTING STRUCTURES OR APPARATUS NOT OTHERWISE PROVIDED FOR; TRANSPORTING; WORKING OR PROCESSING OF SHEET METAL OR METAL TUBES, RODS ORPROFILES WITHOUT ESSENTIALLY REMOVING MATERIAL

To provide a mold anomaly detection device that can detect an anomaly highly accurately without being influenced by noise, and to provide a learning model generation device and a program for the same.SOLUTION: A mold anomaly detection device 4 includes: a detection object data input unit 45 into which a displacement amount of a mold in a press period which is measured by a strain sensor 2 is input as detection object data; and a detection unit 46 for calculating an anomaly degree of the mold from the detection object data input to the detection object data input unit 45 by referring to a learning model M in which a displacement amount of a normal mold is preliminarily learned as a waveform.SELECTED DRAWING: Figure 3 【課題】雑音の影響を受けることなく、高精度に異常検知を行うことができる金型異常検知装置、学習モデル生成装置、及び、これらのプログラムを提供する。【解決手段】金型異常検知装置４は、歪センサ２で測定したプレス期間における金型の変位量が検知対象データとして入力される検知対象データ入力部４５と、正常な金型の変位量を波形として予め学習した学習モデルＭを参照し、検知対象データ入力部４５に入力された検知対象データから金型の異常度を算出する検知部４６と、を備える。【選択図】図３