A decision support system for electricity distribution network refurbishment projects

Abstract

The electricity distribution network in the UK is currently undergoing major refurbishment due to the expiration of equipment installed in the 1950s and 1960s during the large-scale expansion of the network. Regional Electricity Companies are attempting to minimise and control the expenditure required to comply with the distribution licence, to maintain a sufficient level of system reliability and balance the costs of refurbishment with those incurred in operating a sub-standard network. Scottish Hydro-Electric plc. has a five year plan for refurbishment of these networks which involves assessment to identify sections requiring attention followed by proposals to implement the necessary work within the objectives of network refurbishment. The planning engineer faces many pressures in making decisions to adopt the best scheme for a particular section of network and the paper seeks to show how decision analysis methods, in the form of current computer software, may assist him or her in assessing influencing factors and their effect on the outcome of the project. The paper discusses these factors primarily for overhead rural networks, and develops a set of variable and non-variable attributes, including costs, fault clearance, security, wayleaves, visual impact, political impact, etc., to form the basis of a decision model. The application of Artificial Intelligence is introduced using a decision analysis package to produce a decision support system (DSS), codenamed PRODEX, and a computer model is developed on a PC to demonstrate the feasibility of the DSS. A spreadsheet is integrated with the decision analysis through Dynamic Data Exchange to input, store, manipulate and output data. The Visual Basic programming language is used to demonstrate the possibilities for creating a customised, user-friendly front-end to the system. The final DSS shows how the interfaces between applications can be created to enable a customised system to be developed so that no specialised knowledge of the applications is required by the user. A case study is given in the form of a 33 kV refurbishment project in the Highlands and Islands region of Scottish Hydro-Electric pic's system. The study considers the problems of a section of the network, and details four options proposed for improvement of the supply. The PRODEX system implementation is explained, along with its analysis of the options through the use of tools such as outcome distribution probability graphs and event sensitivity comparison diagrams.