arXiv subjects

Georges Kariniotakis

Publications and source records attributed to Georges Kariniotakis.

4 recordsLinked to original sources

arXiv2025

A Scenario-Spatial Decomposition Approach With a Performance Guarantee for the Combined Bidding of Cascaded Hydropower and Renewables

This study develops a scalable co-optimization strategy for the joint bidding of cascaded hydropower, wind, and solar energy units, treated as a unified entity in the day-ahead market. Although hydropower flexibility can manage the stochasticity of renewable energy, the underlying bidding problem is complex due to intricate coupling constraints and nonlinear dynamics. A decomposition in both scenario and spatial dimensions is proposed, enabling the use of distributed optimization. The proposed distributed algorithm is eventually a heuristic due to non-convexities arising from the system's physical dynamics. To ensure a performance guarantee, trustworthy upper and lower bounds on the global optimum are derived, and a mathematical proof is provided to demonstrate their existence and validity. This approach reduces the average runtime by up to 35% compared to alternative distributed methods and by 57% compared to the centralized optimization. Moreover, it consistently delivers solutions, whereas both centralized and alternative distributed approaches fail as the size of the optimization problem grows.

math.OC↗

arXiv2024

Decision-Oriented Learning for Future Power System Decision-Making under Uncertainty

Better forecasts may not lead to better decision-making. To address this challenge, decision-oriented learning (DOL) has been proposed as a new branch of machine learning that replaces traditional statistical loss with a decision loss to form an end-to-end model. Applications of DOL in power systems have been developed in recent years. For renewable-rich power systems, uncertainties propagate through sequential tasks, where traditional statistical-based approaches focus on minimizing statistical errors at intermediate stages but may fail to provide optimal decisions at the final stage. This paper first elaborates on the mismatch between more accurate forecasts and more optimal decisions in the power system caused by statistical-based learning (SBL) and explains how DOL resolves this problem. Secondly, this paper extensively reviews DOL techniques and their applications in power systems while highlighting their pros and cons in relation to SBL. Finally, this paper identifies the challenges to adopt DOL in the energy sector and presents future research directions.

eess.SY↗

arXiv2023

Uncertainty of short-term Wind Power Forecasts -- A methodology for on-line Assessment

The paper introduces a new methodology for assessing on-line the prediction risk of short-term wind power forecasts. The first part of this methodology consists in computing confidence intervals with a confidence level defined by the end-user. The resampling approach is used for this purpose since it permits to avoid a restrictive hypothesis on the distribution of the errors. It has been however appropriately adapted for the wind power prediction problem taking into account the dependency of the errors on the level of predicted power through appropriately defined fuzzy sets. The second part of the proposed methodology introduces two indices, named as MRI and PRI, that quantify the meteorological risk by measuring the spread of multi-scenario Numerical Weather Predictions and wind power predictions respectively. The multi-scenario forecasts considered here are based on the 'poor man's' ensembles approach. The two indices are used either to fine-tune the confidence intervals or to give signals to the operator on the prediction risk, i.e. the probabilities for the occurrence of high prediction errors depending on the weather stability. A relation between these indices and the level of prediction error is shown. Evaluation results over a three-year period on the case of a wind farm in Denmark and over a one-year period on the case of several farms in Ireland are given. The proposed methodology has an operational nature and can be applied to all kinds of wind power forecasting models

physics.data-an↗

arXiv2019

Short-term photovoltaic generation forecasting using multiple heterogenous sources of data

Renewable Energies (RES) penetration is progressing rapidly: in France, the installed capacity of photovoltaic (PV) power rose from 26MW in 2007 to 8GW in 2017 [1]. Power generated by PV plants being highly dependent on variable weather conditions, this ever-growing pace is raising issues regarding grid stability and revenue optimization. To overcome these obstacles, PV forecasting became an area of intense research. In this paper, we propose a low complexity forecasting model able to operate with multiple heterogenous sources of data (power measurements, satellite images and Numerical Weather Predictions (NWP)). Being non-parametric, this model can be extended to include inputs. The main strength of the proposed model lies in its ability to automatically select the optimal sources of data according to the desired forecast horizon (from 15min to 6h ahead) thanks to a feature selection procedure. To take advantage of the growing number of PV plants, a Spatio-Temporal (ST) approach is implemented. This approach considers the dependencies between spatially distributed plants. Each source has been studied incrementally so as to quantify their impact on forecast performances. This plurality of sources enhances the forecasting performances up to 40% in terms of RMSE compared to a reference model. The evaluation process is carried out on nine PV plants from the Compagnie Nationale du Rh{ô}ne (CNR).

stat.AP↗