Data & AI 2
Domain : Data and AI
Industry : Manufacturing
Challenge
In a 24*7 production environment, sudden failures of equipment lead to unplanned idle time, reduction of output and extra spending on repair and maintenance.
Approach
Leverage new and existing sensors to closely monitor health status of equipment. Identify how and where equipment stops working. Design machine learning models to select sensors that truly represent the degradation of various parts critical to the equipment. Based on historical maintenance logs, model and predict the most likely time of failure and alert in advance.
Outcome
The solution brought straightforward overview of health status of all equipment in the manufacturing plant. In addition, the estimated time of occurrence for all identified failure modes were presented in a single pane of glass, which were refreshed on a daily basis. This led to direct reduction of repair and maintenance cost, ensured actual versus planned production output, and drove an uplift in efficiency and productivity.