Autonomic Computing Application in Power System Distributions M. Al-Zawi M. R.Ahmed-AI-Zawi@2005. Ijmu. ac. uk D. Al-Jumeily a. aljumeily@ljmu. ac. uk A. Hussain a. hussain(,ljmu. ac. uk School of Computing and Mathematical Sciences, Liverpool John Moores University, Byrom Street, Liverpool, L3 3AF, UK. T. F. Abdelsalam Ta222@bath. ac. uk Department ofElectronic & Electrical Engineering, University of Bath, BATH, BA2 7AY, UK Abstract Maintaining the reliability and stability is an important subject in power system distributions research area. Recently many IT applications have been implemented successfully in power system; such as numerical relays, digital Distribution Management Systems (DMS), and Energy Management Systems (EMS). This paper presents an implementation of an integrated external monitoring system to identify and recover the failure in numerical over current relay using neural network. This approach will then be used as a monitoring unit in Self-Healing which is one of the main characteristics of Autonomic Computing. Feed forward neural network is implemented to measure and predict the availability of substation numerical protection in order to enhance their availability. 1. Introduction This paper deals with the implementation of Autonomic Computing (AC) in protection of power system distributions. The goal of autonomic computing is to build a self-management system that has the capability (a) to increase the reliability by implementing autonomic systems that are self- protecting and self-healing and (b) to increase autonomy and performance by enabling the autonomic systems to adoption to various changing circumstances, using self-configuring and self- optimizing mechanisms [1]. In this paper we focus only on self-healing which is the most important aspect of AC, and the goal was achieved by neural network as monitoring and adaptation unit. An external model is used as a self-healing system to monitor the response of overcurrent and earth fault relays, then take the correct adaptation action in case of relay failure. A neural network was used as an external monitoring unit to examine its response with the overcurrent and earth fault, then to activate a Circuit Breaker (CB) signal. 2. Self Management Classification Table 1 summarizes the general components of autonomic systems. An autonomic system is self- management system that can be achieved by satisfying four aspects, which are self-configuring, self-healing, self-optimizing, and, self-protection referred to as (CHOP). A system that satisfies all the four self- management aspects called Autonomic Computing [2]. Self-healing concerns with ensuring effective recovery 978-1-4244-1841-1/08/$25.00 C2008 IEEE 417