Data Mining Methods and Techniques for Fault Detection and Predictive Maintenance in Housing and Utility Infrastructure

Recent development of new approaches to data collection, storage and analysis make the data-driven condition monitoring techniques a powerful instrument in housing and utility infrastructure maintenance. Advancements in software development and sensor construction lead to spread of "Internet of Things" concept suggesting the devices to be equipped with various sensors producing large amount of data. The paper introduces an architecture of the information system for predictive maintenance in housing and utility infrastructure based on scalable distributed computing and data mining methods for fault detection and prognostics. Data mining methods featured in the information system are compared and analyzed.