Uncertainty on transmission grids

Abstract

Following the climate goals of the Dutch government, a tremendous amount of vRES (variable Renewable Energy Sources) will be built. These energy sources will produce large amounts of power, which must travel great distances on the transmission grid. The Dutch TSO, TenneT, is responsible for the transmission grid's design, construction, and maintenance by European and Dutch law. The planning for transmission grids has become more complex since the liberalization of power markets. Because of this complexity, the planned expansions are published every two years. In energy systems, the ever-changing context of the system makes the possible future states of the system uncertain. In this research, we 1. identify the relevant uncertainties for the transmission system, 2. Identify the most vulnerable lines under these uncertainties, and 3. Look at the effect of the delays on the performance of the overall power system. This gives insights into how the transmission system must evolve to maintain the grid's performance.

Not only more power from vRES will be generated, but coal plants will also be phased out, and the demand for other energy sources (fuel, gas, etc.) will be electrified. More power flow will lead to congestion on the current grid. The Dutch Transmission System Operator (TSO) uses congestion management to alleviate acute congestion and expansion planning to prevent future congestion. To decide whether grid expansion should be used, Production Cost Models (PCMs) are used. To account for uncertainties in a system, the concept of deep uncertainty can be employed to map the uncertainty of a system. Exploratory modeling uses this concept to understand the consequences of deep uncertainty. Experimental modeling and analysis could thus be used to understand the effects of vRES on the power grid.

Direct Current Optimal Power Flow (DCOPF) is used for the formulation of the Production Cost Model (PCM) for the power network modeling. The chosen exploratory modeling framework is a simplified form of the XLRM framework, in which the policies in the model (L) is not altered. The data used as input for the model is selected from a wide range of sources, following the format of the DCOPF formulation. The tool used to build and simulate the DCOPF model is Scenario Analysis Interface for Energy Systems (SAInt) software. The EMA workbench is used to perform exploratory modeling and analysis. Two times 1,000 experiments for a week period have been run with these tools, one set in January and one set in June. To assess which lines in the transmission network might be susceptible to congestion, we analyzed the experiments' outcomes with a congestion-seeking algorithm. The adopted methodology has been verified using a 5-node model.


The results of these experiments are divided into two types of outcomes: the general system outcomes and the line outcomes; the latter has been split up based on the urgency of capacity expansion. The analysis tools by the EMA workbench were used to analyze the results of the experiments, especially the Patient Rule Induction Method (PRIM) algorithm. The experiments showed significant differences in congestion between June and January...