Finding the reason behind a machine fault is not always simple. Industrial equipment produces sensor readings throughout the working day, and a change in one part of a machine may affect other parts as well. This paper presents an Industrial Knowledge Graph Framework intended to bring machine-learning methods and connected equipment information into one diagnostic process. Isolation Forest is used to flag readings that look unusual, and XGBoost estimates the risk of equipment failure. The Knowledge Graph records how machines, components, sensors, faults, and maintenance activities relate to one another. By following these links, maintenance staff can investigate possible causes and consider how a fault might spread. The framework is also intended to suggest maintenance actions and show the findings on a dashboard. The overall aim is to give staff earlier warning of developing problems and useful context for deciding what to inspect.