Dispositivo de detecção e monitoramento em tempo real dos robôs emendadores por análise de vibração

Abstract

The difficulty in identifying stopped splicing robots resulted in a monthly loss of 2,250 minutes. In order to mitigate this problem, we sought to improve the detection of operational status, distinguishing stationary robots from those in motion, sending notifications in the event of a stop. Vibration analysis was the method chosen for this optimization, enabling the development of an IoT device containing an accelerometer, and applying a logic where there is a threshold defining the threshold that divides what is a stationary robot or a moving robot was collected samples 2048 vibration points every 10s. Obtaining the magnitude value and comparing it with the threshold magnitude, identifying a change in the state of the robot, a visual signal is triggered that instantly sends messages via MQTT informing which machine, robot and state of the event. Tools were used for its development such as Jupyter Notebook (python), the Arduino IDE and the ESP8266. With the solution for the slowness in identifying the stopped robot, a reduction of 98.88% of this time was obtained, reaching 10s, which had an impact on an increase in production of around 150,000 kilos and the maintenance teams obtained a better optimization of activities.