公开(公告)号：US20240210932A1

公开(公告)日：2024-06-27

申请号：US18064895

申请日：2022-12-12

Applicant: Industrial Technology Research Institute

Inventor： Meng-Lin LI ,  Yu-Hung PAI ,  Hung-Tsai WU ,  Chun-Chieh WANG

IPC: G05B23/02

CPC classification number: G05B23/024 ,  G05B23/0221

Abstract: A device state evaluation method based on current signals is applied to a target device that is powered on, the device state evaluation method includes: collecting a plurality of target current signals corresponding to the target device via an acquisition module; performing a signature extraction operation and a normalization operation via a computing module to obtain a target matrix by using the plurality of target current signals; and performing a diagnosis operation on the target matrix via a diagnosis module to identify whether the target device is in a malfunction state, where an identification result of the diagnosis operation is used as target information. Therefore, whether the target device is in the malfunction state can be evaluated by analyzing the plurality of target current signals. A device state evaluation system is also provided.

公开(公告)号：US20240210911A1

公开(公告)日：2024-06-27

申请号：US18069952

申请日：2022-12-21

Applicant: Industrial Technology Research Institute

Inventor： Yu-Hung PAI ,  Hung-Tsai WU ,  He-Chen LIAO ,  Kai-Jhih YANG

IPC: G05B19/406

CPC classification number: G05B19/406 ,  G05B2219/45244

Abstract: A mold state monitoring system is provided, and the monitoring method thereof is: dividing multiple processing signals of a mold into initial state information and wear state information, so as to obtain a target model and a wear threshold based on the initial state information, and input the wear state information into the target model to obtain a wear index of the mold; inputting the latest multiple processing signals and corresponding wear indices thereof into a time series prediction model for training to obtain wear prediction values of hypothetical times, and then performing a predicting operation based on the wear prediction value, so that when the wear prediction value is greater than the wear threshold, it can be estimated as a damage time point of the mold. Therefore, via the design of the time series prediction model, the target information can be changed at any time on the production line, and the state of the mold can be monitored online in real time.