Risk-based approach to the Optimal Transmission Switching problem

Benoît Jeanson, Simon H. Tindemans · 2025

Read Risk-based approach to the Optimal Transmission Switching problem by Benoît Jeanson, Simon H. Tindemans (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2505-17723.txt.

Introduction

Benoît Jeanson and Simon Tindemans add risk thinking to the choice of which power lines to switch off on purpose, in work from 2025. Grid operators sometimes open lines to keep the rest safe. In weaker sub-networks, policy may accept losing part of the grid rather than breaking hard limits.

The paper first sets out that policy plainly. It then extends the standard mixed-integer model of safe switching with grid-loss checks for single line failures. Links that keep the network whole get special care. Tests on a standard fourteen-bus system match what operators do in practice. A known pattern appears: early openings that foreshadow cascades. The model helps planners price risk before acting.

Catalog record

CREATORBenoît Jeanson, Simon H. Tindemans
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE30 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2505-17723

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Risk-based approach to the Optimal Transmission Switching problem Thanks: This work is part of the CRESYM project OptGrid, funded by RTE - Réseau de Transport d’Électricité. Benoît Jeanson, Simon H. Tindemans Abstract This paper deals with the secure Optimal Transmission Switching (OTS) problem in situations where the TSO is forced to accept the risk that some contingencies may result in the de-energization of parts of the grid to avoid the violation of operational limits. This operational policy, which mainly applies to subtransmission systems, is first discussed. Then, a model of that policy is proposed that complements the classical MILP model of the N-1 secure OTS problem. It comprises a connectivity and notably a partial grid loss analysis for branch outage contingencies. Finally, its application to the IEEE 14-bus system is presented. Solutions similar to those observed in operation are reached by the algorithm, notably revealing the preventive-openings-cascade phenomenon. Nomenclature We consider a graph \mathcal{G} and denote vectors in lowercase bold \boldsymbol{u} , and matrices in uppercase bold \boldsymbol{A} . Functions and operators \odot Hadamard product: elementwise multiplication \langle.,.\rangle Inner product \delta_{c,i} Kronecker function: 1 if c=i ; 0 otherwise \text{org}(e) Origin vertex index of edge e \text{dst}(e) Destination vertex index of edge e \text{inc}(i) Incident edges of vertex i \text{opp}(e,i) For an edge e connected to a vertex v , the vertex on the other end of e , distinct from v . For example, if e\in\text{inc}(v) , then v is either in \text{org}(e) or \text{dst}(e) . Moreover, if v=\text{org}(e) , then v=\text{opp}(e,\text{dst}(e)) . Sets \mathcal{V} indices of the vertices or buses of \mathcal{G} \mathcal{E} indices of the edges or branches of \mathcal{G} \mathcal{C}^{\star} edge indices of contingencies, \mathcal{C}^{*}\subset\mathcal{E} . \mathcal{C} edge indices of contingencies including 0 for none Matrices, vectors and scalar \boldsymbol{A} The incidence matrix of \mathcal{G} . \boldsymbol{A}_{ve}=0 if the edge e does not connect the vertex v , 1 if it is oriented toward v ; -1 otherwise. \boldsymbol{0},\boldsymbol{1} vectors filled respectively with 0 and 1 \boldsymbol{g},\ \boldsymbol{d} Generations, loads on buses \boldsymbol{p} \boldsymbol{p}=\boldsymbol{g}-\boldsymbol{d} Net injections on buses \boldsymbol{b} Susceptances of branches \bar{\boldsymbol{f}} Flow limits on branches \boldsymbol{v} Binary vector indicating branch openings in the base case. v_{e}=1 if edge e is open, 0 otherwise For a contingency c\in\mathcal{C} : \hat{\boldsymbol{g}}_{c},\ \hat{\boldsymbol{d}}_{c} Post-contingency generations and loads \boldsymbol{f}_{c} Directed flows in the edges \boldsymbol{\phi}_{c} Voltage phase angles in the buses \boldsymbol{p}^{\star}_{c} Net injections in N-1 mirror graph \boldsymbol{f}^{\star}_{c} Flows in N-1 mirror graph \boldsymbol{w}_{c} Binary vector indicating that the branch is opened either in the base case ( \boldsymbol{v} ) or because of the contingency c \boldsymbol{\mathfrak{p}}_{c} Probability of occurrence of contingency c \boldsymbol{\pi}_{c} Binary vector indicating the energized status of buses. \boldsymbol{\pi}_{c,i}=1 if bus i is energized, 0 else \psi_{c,i,e} Variable indicating whether the bus connected to bus i via edge e is energized \sigma_{c} Generation scaling factor I Introduction One of the primary concerns of Transmission System Operators (TSOs) is ensuring the security of the power system. To achieve this, they must anticipate potential events that can compromise it while simultaneously striving to maintain economic efficiency. When it comes to security policies, the cornerstone is the N-1 rule. Sometimes, it is reduced to the statement that the consequences of the tripping of any single element of the grid shall have no impact on the users of the power system. This is actually a narrow view. In fact, the rule derives from a risk-based principle. As contingencies are likely to occur, what is at stake for the operator is that any resulting state of the system remains under control. In other words, the consequences of any likely tripping of a branch shall be mastered. Furthermore, to align the risks to the same level, the more likely a contingency is, the less acceptable its consequences in case of occurrence. So, the criteria the operator follows are based on the definition of the limits between an acceptable and an unacceptable consequence with respect to each contingency considered. Keeping flows or voltages within security limits is one of the key criteria. Losing a part of the grid and de-energizing users could also be considered unacceptable, but is it realistic? In fact, enforcing the latter constraint implies that the real-time operator must avoid at all costs losing customers as a result of a single tripping. This may involve operations on the grid itself, for example, by implementing switching actions or activating redispatch or curtailment of customers, which rapidly becomes expensive. EU regulation [ 1 ] does not contain such standards. Loss is only explicitly mentioned in the description of the security states where it is introduced for the black-out state, which is defined by the loss of more than 50% of the total consumption of the control area of the TSO. The fact that there is no specific standard reflects at least that there is no requirement to not lose any part of the grid, and opens to the interpretation that the regulation admits this may happen. Moreover, there may be operational scenarios where solutions that avoid all customer disconnections do not exist. This paper considers situations where no solution is available, or where the costs associated with the mitigation of potential loss of parts of the grid are prohibitive. Such cases are addressed in the context of the Optimal Transmission Switching (OTS) problem. The OTS problem consists in finding the combination of branch openings in the grid that optimizes a given objective function. Many optimization approaches to the OTS problem have been developed. [ 2 ] by Fisher et al. is the first DC-OPF (Direct Current Optimal Power Flow) based on an MILP (Mixed-Integer Linear Program) approach to it with the goal of solving congestion . This pioneering study lays the foundation for numerous subsequent efforts that employ the DC approximation to solve the OTS. In this model, the switching problem is formulated using a big-M approach and the necessary flexibility – when no switching scheme is enough to cope with the congestion – is provided by the OPF equations that aim at minimizing generation adjustment cost. Hedman et al. extended that approach in [ 3 ] to include security analysis and incorporate the N-1 rule. The algorithm does not permit the loss of load or generation, but it does allow electrical islands either in the base case or after a contingency if the resulting islands satisfy all the constraints. In practical operation, islanding should only be considered in very particular power systems that are designed to be sustainable after a grid split, which implies the fulfillment of a lot of requirements in terms of balancing, control, and stability. To avoid islanding, various recent papers therefore take into account a connectivity constraint, for which various approaches have been proposed ( [ 4 ] , [ 5 ] , [ 6 ] , [ 7 ] ). The contribution of this article is threefold. First, a discussion of the N-1 rule is presented and consequences of including it in operation are presented. Then a deterministic optimization model that reflects that risk-based operational policy is developed. At its core, the connectedness of the network in the base case and in the N-1 situations is assessed. Finally, its application to the IEEE

Read the complete text