-
Decision-Oriented Learning for Future Power System Decision-Making under Uncertainty
Authors:
Ran Li,
Haipeng Zhang,
Mingyang Sun,
Fei Teng,
Can Wan,
Salvador Pineda,
Georges Kariniotakis
Abstract:
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 thro…
▽ More
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.
△ Less
Submitted 7 April, 2024; v1 submitted 8 January, 2024;
originally announced January 2024.
-
Uncertainty of short-term Wind Power Forecasts -- A methodology for on-line Assessment
Authors:
Georges Kariniotakis,
Pierre Pinson
Abstract:
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 a…
▽ More
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
△ Less
Submitted 4 October, 2023;
originally announced October 2023.
-
Sizing of a PV/Battery System Through Stochastic Control and Plant Aggregation
Authors:
Thomas Carriere,
Christophe Vernay,
Sebastien Pitaval,
Franccois-Pascal Neirac,
George Kariniotakis
Abstract:
The objective of this work is to reduce the storage dimensions required to operate a coupled photovoltaic (PV) and Battery Energy Storage System (BESS) in an electricity market, while keeping the same level of performance. Performance is measured either with the amount of errors between the energy sold on the market and the actual generation of the PV/BESS i.e. the imbalance, or directly with the…
▽ More
The objective of this work is to reduce the storage dimensions required to operate a coupled photovoltaic (PV) and Battery Energy Storage System (BESS) in an electricity market, while keeping the same level of performance. Performance is measured either with the amount of errors between the energy sold on the market and the actual generation of the PV/BESS i.e. the imbalance, or directly with the revenue generated on the electricity market from the PV/BESS operation. Two solutions are proposed and tested to reduce the BESS size requirement. The first solution is to participate in electricity markets with an aggregation of several plants instead of a single plant, which effectively reduces the uncertainty of the PV power generation. The second is to participate in an intra-day market to reduce the BESS usage. To evaluate the effects of these two solutions on the BESS size requirement, we simulate the control of the PV/BESS system in an electricity market.
△ Less
Submitted 15 October, 2019;
originally announced October 2019.
-
Short-term photovoltaic generation forecasting using multiple heterogenous sources of data
Authors:
Kevin Bellinguer,
Robin Girard,
Guillaume Bontron,
Georges Kariniotakis
Abstract:
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…
▽ More
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{รด}ne (CNR).
△ Less
Submitted 14 October, 2019;
originally announced October 2019.
