コンテック Official site

[Free White Paper] Basics of Anomaly Detection through Vibration

コンテック

In recent years, the advancement of sensing technology and IoT technology utilizing that data has led to increased visibility in production sites. As the next step, it is necessary to utilize the acquired data to contribute to productivity improvement. One immediate thought is to analyze the acquired data for equipment maintenance. There are three types of maintenance based on timing: corrective maintenance, preventive maintenance, and predictive maintenance. Corrective maintenance, which involves repairing after a failure, poses a significant problem due to its impact on production planning. Additionally, preventive maintenance requires regular maintenance such as parts replacement to ensure safety margins, which can lead to increased costs. In contrast, predictive maintenance allows for the detection of signs of potential failures, minimizing the impact on production planning while also keeping maintenance costs to a minimum, thus resolving the issues associated with the first two types. Until now, it has been challenging to ensure reliable predictive accuracy, and the applicability has been limited. However, recent advancements in AI have created an environment where the realization of predictive maintenance through anomaly detection can be expected. In this article, we will analyze the common vibration issues in production sites and aim to improve productivity through equipment and machinery maintenance by building an anomaly detection AI to capture signs of failure.