METHOD DEVICE AND SYSTEM FOR DETERMINING MOLD ABNORMALITIES USING ARTIFICIAL INTELLIGENCE-BASED TEMPERATURE AND VIBRATION SENSOR DATA

In accordance with one embodiment, an apparatus acquires molded product information about the shape and size of a molded product manufactured through a mold; acquires mold shot count information measured through a mold counter measuring the number of mold operations for a preset period, mold cycle time information, mold temperature information measured from a temperature sensor and mold vibration information measured from a vibration sensor; checks a temperature change pattern of the mold according to a change in the temperature information; checks a vibration change pattern of the mold according to a change in the vibration information; trains a first artificial neural network by setting the molded product information, the shot count information, the cycle time information, the temperature change pattern and the vibration change pattern to the first artificial neural network as an input signal, and setting a molded product production quality, a molded product production time and a mold lifetime as an output signal; determines a shot count, cycle time, temperature and vibration of a mold for a specific molded product based on the trained first artificial neural network; determines a production time of a molded product produced through a specific mold and the lifetime of the specific mold based on the trained first artificial neural network; when a change level of the temperature change pattern reaches a preset reference point, changes a production temperature such that the change level of the temperature change pattern can be adjusted to less than the preset reference point; and when a change level of the vibration change pattern reaches a preset reference point, changes vibration change elements including a movement of a mold, an operation of injecting resin into a mold, a packing operation and injection pressure such that the change level of the vibration change pattern can be adjusted to less than the preset reference point. Therefore, the present invention is capable of detecting a defect cause of a molded product and predicting a time when a defect occurs. 일실시예에 따른 장치는 금형을 통해 제작되는 성형품의 형상 및 크기에 대한 성형품 정보를 획득하고, 미리 설정된 기간 동안 금형 작동 횟수를 측정할 수 있는 금형 카운터를 통해 금형 작동 횟수(Shot Count) 정보, 금형 제작 주기(Cycle Time) 정보, 온도 센서로부터 측정한 금형의 온도 정보 및 진동 센서로부터 측정한 금형의 진동 정보를 획득하고, 온도 정보의 변화에 따른 금형의 온도 변화 패턴을 확인하고, 진동 정보의 변화에 따른 금형의 진동 변화 패턴을 확인하고, 성형품 정보, 작동 횟수 정보, 주기 정보, 온도 변화 패턴, 진동 변화 패턴을 제1 인공신경망에 입력 신호로 설정하고, 성형품 생산 품질, 성형품 생산 시간 및 금형 수명을 출력 신호로 설정하여 제1 인공신경망을 학습하고, 학습된