We have implemented a real-time expert system as part of a supervisory control and data acquisition system for the Public Utilities Board of Singapore, controlling its 22 kV distribution network. It works as an operator support tool by diagnosing network disturbances and device malfunctions, and by presenting a switching sequence which can be executed immediately to restore supply. In this paper we describe how the expert system is integrated into a large supervisory control and data acquisition system for power distribution networks. The necessary techniques to cover on-line processing of real-time data, intelligent alarm processing and network reconfiguration/restoration are discussed. The expert system is based on a hierarchic multi-level problem-solving architecture, integrating model-based, heuristic and algorithmic techniques acting on an object-oriented data structure. Techniques such as the automatic creation and update of the knowledge base, event filtering, on-line event processing, and structural and temporal focusing have been implemented to enable the system to perform its task on-line and in real time.