Evaluating network operations strategies with fuzzy discrete-time simulation. Case study: Western and central Victoria-Australia

Network control is a key process for distribution network operations, where electricity supply is restored through a combination of remote switching and field crew activities. Although analytical methods have been widely applied to problems such as network expansion and restoration sequencing, broader operational decisions, such as workforce availability and the extent of cooperation between adjacent depots, are often based on subjective management judgement, partly because evaluating their business-wide effects requires substantial computational resources. The emergence of cloud computing and high-performance processors allow network operations, and the temporal and spatial uncertainty of faults, to be modelled across an entire distribution network within a practical timeframe and at an acceptable computational cost. In this paper, we explain how an Australian Distribution Network Service Provider (DNSP) has developed an integrated model utilising discrete-time simulation, fuzzy logic, and design of experiments (DOE) to evaluate strategic network operations decisions and optimise the trade-off between regulatory penalties and operations costs. The selected case study showed that partially relaxing the boundaries between depots can reduce average outage duration and regulatory penalties, although it may also lead to higher operations costs. Managerial Relevance Statement One of the challenges facing distribution network service providers (DNSPs) is assessing how changes in network control and operations processes will affect operations costs and regulatory penalties. The uncertainty involved in estimating how such changes affect the company's bottom line can make management reluctant to implement proposed changes, as their effects on operational costs, regulatory penalties, and employees’ work-life balance are difficult to quantify. Consequently, potentially beneficial operational changes may not be pursued. To help evaluate these decisions, we developed a decision-support tool that combines discrete-time simulation with fuzzy logic. The tool is currently being used to support decision-making by a DNSP serving more than one million customers in Australia, across Central and Western Victoria.