Dynamic Security assessment: Challenges (An European TSO perspective)

Abstract

Summary form only given. The complexity of the power systems is increasing. More and more generations based on renewable energy are installed in the system. Some are dispersed (PV in distribution system) or others far way from load centers (off shore wind). They are generally intermittent (day ahead forecasts are not very accurate). TSOs have a lot of difficulties to build overhead power lines. People don't like their impact on the landscape and they are now afraid of hypothetical effect of EMF on health. We must use more complex solutions: numerous Phase Shifters Transformers, upgrades of existing corridors by using new conductors (ACSS), underground cables, HVDC links embedded in AC systems. The Single European Electricity market is an optimizer which maximizes the use of existing assets, pushing the system to its limits. To operate such a very large and complex system, new tools are needed to help operators to make decisions. One of the challenges is the very large scale, the full European system must be taken into account, electrical phenomena don't stop at administrative borders, (10000 electrical buses, 2000 generators, 100 PSTs, 10 HVDC links, ...). More and more post-fault actions are implemented to control the system using topological actions and flexible devices (PST, HVDC link,). In iTesla, ongoing project funded by the European Commission, we propose to develop a platform to offer solutions to tackle some of these issues. The Online Security Assessment is based on “Dynamic Security Assessment”. Corrective or remedial actions are performed after the occurrence of a fault, they are post-fault actions. The actions are event-based or measurement-based. They are implemented via automatic devices (SPS) or human actions (operating rules in control rooms). Interactions between these multi-actions can't be easily understood without a time domain simulation. The possible failure of one of these corrective actions implemented through IT systems which can't be considered as hundred percent reliable, must be considered. Moreover, post-fault steady states depend on the trajectory and can't any longer be computed using a conventional power flow. We are operating the system with less margin and unstable dynamic phenomena could appear (for example, poorly damped inter area oscillation). Local dynamic problems (for example, Voltage collapse or transient stability issue) could initiate a cascade of events leading to a very large blackout. The only practical tool available today to assess these possible phenomena is time domain simulation. This time domain simulation must cover the whole Pan-European system which is a very large system (around 125.000 state variables); this is also a tough mathematical problem: Non-linear, stiff, oscillating, poorly damped, discontinuous... The first challenge is to find the appropriate tradeoff between three conflicting requirements: speed of computation, the accuracy and the flexibility. We want a computation time as small as possible; this tool is at the core of decision making process in real-time, short term look head (few hours) or it is embedded in offline MonteCarlo simulation. We want a reasonably accurate solution; we prefer to be slightly conservative, we need to avoid numerical stabilizations of physical unstable system which is a true challenge. We must be able to model easily new devices and protections/controls. Some very specific components, protections and controls are installed in the system. We need a flexible mean to describe their behaviors using a equation-based modeling and not any longer just enter parameters for hardcoded equations. The second challenge is “the accuracy”. To ensure a credible assessment, validation of models is critical but how to perform this validation far away from nominal conditions? A rigorous data management is also mandatory. We must manage more data and very technical data describing the dynamic behaviors. An incorrect value for a single parameter could lead to very different results. A unique data base largely used by all the teams (planning, maintenance, operation) within the company managed by an expert dedicated team ensuring the data quality, is a prerequisite. The third challenge is “the initialization process”. The current practice for DSA in real-time and near to real-time is to initialize the dynamic models using the results of state estimators. The estimated voltage and current are not 100% accurate and some dynamic states could be outside limits hidden in the dynamic models generally when the system is under stress. This initialization process could fail, unfortunately when DSA is required. This is even more frequent with power electronic equipments. Moreover, this process can't take into account memory effects; for example, if the pre contingency state is not a “normal” steady state: if some over voltage excitations are activated. It is impossible to estimate the remaining time before the end of this over excitation mode using only a snapshot. An extended state estimation is required. As a conclusion, to ensure a more reliable and more easy to use DSA, some challenges must be solved from organizational ones (data management) to more technical ones (extended state estimation).