Многоуровневая архитектура системы предиктивного анализа работы генерирующего оборудования

Reliability asset management is a crucial issue in the energy domain. A predictive analysis system of energy generating equipment is a key component of Reliability asset management. A predictive analysis system development requires solution for data acquisition about equipment operation as well as for decision support towards efficient equipment maintenance. The main objective of these solutions is to minimize energy generating equipment failures. A paper presents multilayer architecture of the energy generating equipment system predictive analysis. The architecture is based on three main ideas, which make it differ from previous solutions. The first finding is splitting a data acquisition component and a data-transferring component due to scale reasons. The second is the usage of a queue broker for managing data streams collected from different data sources. The broker provides unified access to data streams. The third component is an API providing access for third-part analytics predictive solutions towards "seamless" integration and deployment. The proposed architecture was applied in a prototype of a predictive analytics system for energy generating equipment risk-oriented maintenance.