23 rd International Conference on Electricity Distribution Lyon, 15-18 June 2015 Paper 1337 CIRED 2015 1/5 COMPUTER-AIDED DISTRIBUTION NETWORK PLANNING USING EXPERT RULES Neel van HOESEL Marinus GROND Anton ISHCHENKO Han SLOOTWEG Phase to Phase B.V. Enexis B.V. Phase to Phase B.V. Enexis B.V. Eindhoven University Eindhoven University The Netherlands Eindhoven University of Technology of Technology a.a.ishchenko@phasetophase.nl of Technology The Netherlands The Netherlands The Netherlands n.v.hoesel@gmail.com m.o.w.grond@tue.nl j.g.slootweg@tue.nl ABSTRACT This paper presents a computer-aided planning tool for the capacity planning of medium voltage distribution (MV-D) networks. Requirements originating from network planning practice are incorporated in an expert system. The proposed approach focusses on finding the locations and type of assets that satisfy topological, geographical, and operational constraints. The main principle of the approach is to start with all possible expansion options and then to reduce the number of these expansion options significantly by using expert rules based on established network planning practice. Practically feasible expansion options are modelled automatically to evaluate if they offer a solution to bottlenecks. Finally, a performance indicator is calculated to evaluate and select the expansion options. An example case is used to illustrate the planning tool. Goal of the tool is to support planning of MV-D networks in practice. INTRODUCTION Distribution Network Operators (DNOs) require more advanced planning tools to deal with the challenges of future network planning. New technologies, are often connected to the distribution part of the electricity network (e.g. distributed generators, electric vehicles, and heat pumps) as a result of the on-going energy transition. These technologies complicate the planning process and solutions become less straight-forward. Nevertheless, generation of alternative expansion plans is still done by hand [1]. This makes the process labour-intensive to assess various bottlenecks or scenarios and open to inconsistency and subjectivity. Automated generation and evaluation of expansion options can support the planning process [1],[2]. Therefore, this paper will focus on a computer-aided planning tool for the capacity planning of medium voltage distribution (MV-D) networks. Requirements originating from network expansion planning practice are translated into if-then rules like in an expert system. An expert system is particularly useful for translating expert knowledge into a computer system. Additionally, an expert system increases the reliability of decisions, and offers an explicit explanation of how a decision was reached [3]. The proposed approach focusses on finding the locations and type of assets, in our case MV-cables, which satisfy topological, geographical, and operational constraints. The main principle of the approach is to start with all possible cable expansion options and then to reduce the number of these expansion options significantly by using expert rules based on established network planning practice. The first section will present an overview of common bottlenecks found in the Dutch MV-D network resulting from interviews with DNOs. The second section will describe the tool. The third section will offer an example case. The fourth and final section contains a summary and conclusions. COMMON BOTTLENECKS IN MEDIUM VOLTAGE DISTRIBUTION NETWORKS Dutch MV-D network have a “European” distribution layout in which meshed networks are operated radially by Normally Open Points (NOPs) [4], like the example in Fig. 2. This means that the network can be reconfigured in situations of emergency or maintenance. Moreover, nearly all Dutch MV-D networks consist of underground cables. From interviews with experts working in the field of MV- D network expansion planning, common bottlenecks and their solutions are gathered as they occur in practice. The main capacity and voltage bottlenecks in the MV-D networks are: - Overloaded cable, where cables are loaded beyond their allowed level according to DNO design criteria or manufacturer specifications. - Lack of reconfiguration possibilities, where MV-D network segments cannot be reconfigured due to capacity limitations or reconfiguration requires more than a specific number of steps. - Voltage level problems, where node voltage levels are too high or too low as specified in DNO design criteria or manufacturer specifications. - Overloaded MV/LV transformer, where transformers are loaded beyond their allowed level according to DNO design criteria or manufacturer specifications. Each of these bottlenecks can be resolved in a number of ways. The ones most commonly applied in practice are listed in Table 1 (see Fig. 2 for terminology).