Fault tolerance in distributed power systems

The demand for continuously available information processing systems increases every day. Transaction processing, communications systems, and critical processes all require nonstop, fault tolerant operation. There are four steps to achieving fault tolerance in any system. However, no system will operate without a supply of electrical power. This paper is a tutorial that reviews the four principles of fault tolerance. Special considerations for high availability and fault tolerance in distributed power systems are highlighted. The paper discusses reliability, availability and life, redundancy and partitioning, fault isolation, fault detection and annunciation, and on-line repair.