Enhanced, Risk-Based System Development Process: A Case Study from the Belgian Transmission Network

Abstract

Abstract-A large share of renewable generation raises some unique challenges and is forcing change in the operation and planning of electricity transmission systems. In this paper, using real transmission network data representing a part of the Belgian network, the need for building a large set of credible future operating conditions to identify and validate system development decisions is demonstrated. The paper argues that the traditional approach of checking system development options on a few selected operating conditions is no longer adequate in a power system with high penetration of variable generation. An application of a clustering algorithm is demonstrated to build a representative set of operating conditions for system development that not only gives better representation of future operation, but also provides a system development planner with greater confidence with respect to investments that are likely to be necessary. The paper describes the way in which operability of the future system is evaluated and possible need for reinforcement is identified. A discussion is presented of the meeting of demand under initially intact network conditions and while facilitating maintenance outages securely.Index Terms-transmission planning; renewable energy sources; optimal power flow.

Identifying high-impact operating states in power system reliability analysis

“…Ref. [8] also found that for the case considered there, the critical operating states were to a greater extent characterized by high wind power generation than by high (i.e. peak) system load.…”

“…agglomerate clustering or K-means clustering) was not equally important to obtain accurate estimates for the reliability indices. More recently, several applications of clustering of operating states have been reported for transmission system planning [6]- [8]. Ref.…”

“…feature sets based on generation and load. However, there is a consensus that clustering methods generally are not well suited to handle outliers and extreme operating states [2], [8]. In a sense, clustering can be seen as counterproductive to the identification of outliers, given that the basic aim of clustering is not to search for specific operating states but rather to be able to discard as many operating states as possible.…”

“…In [9], a small number of critical operating states were identified through a process involving expert elicitation and power flow analyses. References [6], [8], on the other hand, use the term critical operating state in a system development context as states for which investment might be required due to very high costs of generation re-dispatch or load shedding. Several clustering approaches for reducing the number of operating states to consider for the system development analysis were compared, and it was found that none of them succeeded to capture the most critical of the operating states.…”

Making energy system models useful: Good practice in the modelling of multiple vectors

“…The relationship between scenarios as exogenous visions of the future and scenarios as different realisations of local variables, shown here for the example of investment in generation capacity from(Bukhsh, Bell, Vergnol, Weynants, & Sprooten, 2018).…”