Search

The energy system needs a range of forecast types for its operation in addition to the narrow wind power forecast. Therefore, the group behind the former IEA Wind Task 36 Forecasting for Wind Energy (running 6 years, from 2016-2021) has broadened its perspective on forecasting issues in part by reaching out to other IEA Technology Collaboration Programmes such as the ones for PV, hydropower, system integration, hydrogen etc. The three existing Work Packages (WPs) on NWP Improvements (WP1), Power and Uncertainty Forecasting (WP2) and optimal use of Forecasting Solutions (WP3), are complemented by thirteen work streams in a matrix structure. The three work packages span three distinct areas of challenge in forecasting for the weather driven energy system. The first area is the continuing effort to improve the representation of physical processes in weather forecast models through both new high performance initializations and tailored parameterizations. The second area is the heterogeneity of the forecasters and end users, the full understanding of the uncertainties throughout the modelling chain and the incorporation of novel data into power forecasting algorithms. A third area is representation, communication, and use of these uncertainties to industry in forms that readily support decision-making in plant operations and electricity markets. Task 51 will focus on facilitating communication and collaborations among international research groups engaged in the improvement of the accuracy and applicability of forecast models and their utility for the stakeholders in the wind industry, in the power sector and in the energy system. The first two activities of Task 51 were (1) a workshop in Dublin on the State of the Art and Research Gaps for Forecasting. The results of the workshop will be compiled into a journal article, and (2) the publication of the IEA Recommended Practice for the Implementation of Renewable Energy Forecasting Solutions as an open access book by Elsevier. Other planned activities include further workshops on seasonal forecasting with emphasis on Dunkelflaute, storage and hydro in May 2023, a workshop on minute-scale forecasting (2024), and a workshop on extreme power system events (2025). The results of these interactive workshops will be compiled into a journal articles. Additionally, the Recommended Practice on Forecast Solution Selection will be updated to reflect the broader perspective. Reference: Corinna Möhrlen, John Zack, Gregor Giebel (eds): IEA Wind Recommended Practice for the Implementation of Renewable Energy Forecasting Solutions. Elsevier, 348 pages, Nov. 2022. ISBN: 9780443186813. Download the individual chapters from https://www.sciencedirect.com/book/9780443186813/iea-wind-recommended-practice-for-the-implementation-of-renewable-energy-forecasting-solutions.

Standard practice of decision-making in energy systems relies largely on complex modeling chains to address technical constraints and integrate numerous sources of uncertainty. The increased penetration of Renewable Energy Sources (RES) such as solar and wind plants adds complexity due to the weather dependency of their electricity production. Artificial Intelligence (AI) based tools have proven their efficiency in different applications in the energy sector ranging from forecasting to optimization and decision making. They permit to simplify modeling chains and to improve performance due to higher learning capabilities compared to state-of-the-art methods. However, decision-makers of the energy sector need to understand how decision-aid tools construct their outputs from the data. AI-based tools are often seen as black-box models and this penalizes their acceptability by end-users (traders, power system operators a.o.). The lack of interpretability of AI tools is a major challenge for the wider adoption of AI in the energy sector and a fundamental requirement to better support humans in the decision-aid process. Agents of energy systems expect very high levels of reliability for the various services they provide. As energy systems are impacted by multiple uncertainty sources (e.g. available power of RES plants, weather and meteorological conditions, market conditions), developed AI tools should not only be performant on average situations but be able to guarantee robust solutions in the case of an extreme event. Therefore, our research focuses on understandable representations of data-driven decision-aid models for human operators in the energy sector. In order to enhance the interpretability of the AI models, a technique borrowed from the computer science domain is explored and further developed. Genetic programming and more precisely Symbolic Regression is used to derive a symbolic representation for the data-driven model that can take the form of a single equation. This equation results according to a specific reward function. The optimal solutions are selected naturally mimicking the biological theory of survival of the fittest. The main outcome is the production of symbolic representations of the AI models that require minimum changes when applied to different case studies. In this presentation a real-world use case is considered, to demonstrate the added value of the proposed tools for decision-making when trading the production of wind and solar power plants to the day-ahead market. An annual period of data is considered to train and test the proposed model. The typical modeling chain involves as many as 12 models for forecasting RES production, weather and meteorological conditions, together with stochastic optimization to derive trading decisions. A single AI-based model here replaces this complex chain. Such simplification is a significant enhancement to the modeling chain interpretability and facilitates trust to the human decision-maker. This work is carried out in part in the frame of the European project Smart4RES (Grant No 864337) supported by the H2020 Framework Program and in part in the frame the Marie-Curie COFUND project Ai4theSciences (Grant No 945304)

Operational forecasting models of Renewable Energy Sources (RES) are usually developed for specific time frames, using a reduced set of data sources as a function of the forecasting horizon. This results in discontinuities in predictions between the different time frames from the next minutes to the next days, which is detrimental to decision-making in power systems. Although advanced forecasting model may lower forecasting error, an alternative and maybe more impactful option consists in expanding the set of potential sources for RES forecasting, going beyond recent power measurements, classical Numerical Weather Prediction and information from satellites. It is therefore crucial for next-generation RES forecasting to propose continuous, seamless predictions that harvest the full potential of heterogeneous data sources. This presentation highlights the different solutions developed in the Horizon2020 project Smart4RES for seamless RES forecasting based on the combination of multiple data sources, including high-resolution weather measurements and forecasts.A step towards high-resolution RES forecasting, i.e. temporal resolutions below 5 min and spatial resolutions at the scale of 100 m, is achieved by the integration of high-resolution into RES forecasting models. The use of lidar measurements for the minute-ahead power forecasting of wind turbines improves RMSE against persistence. The production of wind speed forecasts at a 100m-30s resolution thanks to Large Eddy Simulation (LES) at different wind farms of an isolated power systems enables to predict the total variability of wind power production on 10-min rolling intervals, with higher reliability than an approach based on traditional NWP that have a lower spatio-temporal resolution.The combination of multiple data sources has proven to be efficient for the improvement of RES forecasting, especially at intraday horizons. Smart4RES proposes optimal combinations for both weather and RES forecasting. The combination of high-resolution irradiance maps derived from a network of All-Sky-Imagers (ASI) with satellite images outperforms the predictions of the ASI network only. Similarly, the combination of filter-based models for Wind and PV forecasting improves the forecasting scores of the individual models.

International audience; This paper presents an optimization model for Home Energy Management Systems from an aggregator’s standpoint. The aggregator manages a set of resources such as PV, electrochemical batteries and thermal energy storage by means of electric water heaters. Resources are managed in order to participate in the day-ahead energy and local flexibility markets, also considering grid constraint support at the Point of Common Coupling. The resulting model is a Mixed-Integer Linear Programming problem in which the objective is to minimize day-ahead operation costs for the aggregator while complying with energy commitments in the day-ahead market and local flexibility requests. Three sources of uncertainty are considered: energy prices, PV production and load. Adjustable Robust Optimization is used to find a robust counterpart of the problem for including uncertainty. The results obtained show that using robust optimization allows strategic bidding to capture uncertainties while complying with obligations in the wholesale and local market. Data from a real-life energy community with 25 households is used to validate the proposed robust bidding methodology.

The last 6 years, the International Energy Agency (IEA) Wind Task 36 “Forecasting for Wind Energy” has provided forecasting stakeholders (weather institutes, forecast service providers, end users and academics) a platform to discuss challenges and benefits of forecasting for wind power. These discussions have led to a number of activities and initiatives to overcome challenges and to broadcast the benefits of forecasting. Among the major outputs are an information portal with links to free data, a collection of use cases for probabilistic forecasts, and the IEA Recommended Practice on Forecast Solution Selection, including 4 chapters dealing with the (1) solution selection process, (2) benchmarks and trials, (3) verification process and use of online measurements from wind farms for real-time forecasting applications. In the future, we will no longer “integrate” wind and solar into existing power systems, but instead are wind and solar going to be the backbone of our power systems.To address those challenges in an integrated fashion, the IEA Task for Forecasting under the IEA Wind Technology Collaboration Programme (TCP) relaunched with a new Task number (51) and a new work program. The work packages (WPs) are still structured according to stakeholder topics: WP1 deals with weather forecasting, and mainly addresses meteorologists, WP2 deals with the conversion of the weather feeds to the application specific variables such as wind power and addresses forecast vendors, and WP3 deals with the applications and how to get most value out of the forecasts, and therefore addresses the forecast users, including recent advances in data science and digitalisation. However, many of the topics the new Task takes up are cross-cutting, and are therefore now Work Streams (WS):Atmospheric physics and modelling (lead by WP1) Airborne Wind Energy Systems (WP1) Seasonal forecasting (WP1) State of the Art for energy system forecasting (WP2) Forecasting for underserved areas (WP2) Minute scale forecasting (WP2) Uncertainty / probabilistic forecasting (WP3) Decision making under uncertainty (WP3) Extreme power system events (WP3) Data science and artificial intelligence (WP3) Privacy, data markets and sharing (WP3) Value of forecasting (WP3) Forecasting in the design phase (WP3) Most of these work streams require collaboration, and therefore have dedicated partners in other IEA Wind Tasks, or in IEA Tasks outside of the Wind TCP. Task 51 will therefore collaborate with IEA Wind Tasks 32, 44, 48 and 50, IEA PVPS Task 16, IEA Hydro, the IEA Hydrogen TCP, IEA Bioenergy Task 44 and WMO. A major activity of Task 51 will be four public workshops in the next four summers, starting with a workshop on the State of the Art and Research Gaps in 2022, on seasonal forecasting with a special emphasis on hydro power and storage in 2023, on minute scale forecasting in 2024, and on extreme power system events in 2025. We keep the community updated on events, new publications and other relevant information on our website ieawindforecasting.dk and via LinkedIn and Research gate.

International audience; Dynamic Line Rating is a technology devised to modify an overhead line's current-carrying capacity based on weather observation. The benefits of this modification may include reduced congestion costs, an increased renewable energy penetration rate, and improved network reliability. DLR is already well developed, but few papers in the literature investigate DLR day-ahead forecasting. The latter is central to DLR development since many of the decisions related to grid management are taken at least on a day-ahead basis. In this paper, two problems related to DLR forecasts are dealt with: how to achieve precise, reliable calculations of day-ahead forecasts of overhead line ampacity and how to define a methodology to calculate safe rating values using these forecasts. On the first point, four machine-learning algorithms were evaluated, identifying the best approach for this problem and quantifying the potential performance. On the second point, the developed methodology was tested and compared to the current static line rating approach.

International audience; Real-time current-carrying capacity of overhead conductors is extremely variable due to its dependence on weather conditions, resulting in the use of traditionally conservative static ratings. This paper proposes a methodology for exploiting the latent current-carrying capacity of overhead transmission lines taking into account line ampacity forecasts, power flow simulations and the network operator's risk aversion. The procedure can be described as follows: Firstly, probabilistic forecasts for the current rating of transmission lines are generated, paying particular attention to the reliability of the lower part of the distribution. Secondly, a cost benefit analysis is carried out by solving a bilevel stochastic problem that takes into account the reduction in generation costs resulting from a higher power transfer capacity and the increased use of reserves caused by forecast errors. The risk appetite of the network operator is considered in order to accept or penalize high-risk situations, depending on whether the network operator can be described as risk neutral or risk averse.

The International Energy Agency (IEA) Wind Task 36 on Wind Power Forecasting organises international collaboration, among national weather centres with an interest and/or large projects on wind forecast improvements (NOAA, DWD, ...), forecast vendors and forecast users to facilitate scientific exchange to be prepared for future challenges.The talk discusses the general setup of the Task, and the latest developments. Among those are decision making under uncertainty. To this aim, a series of forecasting experiments are being developed and one initial experiment was tested by a wide audience. The forecasting experiments took the form of a game, during which the participants could experience the benefit of probabilistic information on their decisions to trade. Other results include an information portal for meteorological data, and the IEA Recommended Practice for Forecast Solution Selection which is divided into 3 parts: (1) "Forecast Solution Selection Process", (2) "Designing and Executing Forecasting Benchmarks and Trials", and (3) "Evaluation of Forecasts and Forecast Solutions". The Recommended Practice guideline encourages forecast users to establish a framework of metrics that help identify, whether the user's forecast performance criteria effectively incentivize the forecast provider to optimize towards the forecast target variable that has the most value for the user's application(s). For this year, we intend to update the guideline in the light of the experiences throughout the industry in its initial application, and after collecting this experience at 3 Open Space workshops.Collaboration is open to IEA Wind member states; 12 countries are already actively collaborating.The Task is divided in three work packages: Work Package (WP) 1 is a collaboration on the improvement of the scientific basis for the wind predictions themselves. This includes numerical weather prediction (NWP) model physics, but also widely distributed information on accessible datasets. This WP also currently organises a benchmark for NWP models, based on the Wind Forecast Improvement Project 2 (WFIP2) datasets. WP2 deals with the power conversion from the wind speed forecasts and the associated vendor issues. Amongst other things, WP2 published the IEA Recommended Practice on how to select an optimal wind power forecast solution for a specific application. The focus of WP3 is on the engagement of end users to disseminate the best practice in the use of wind power predictions, especially probabilistic forecasts and also what kind of measurements are required in real-time environmentsA major activity of the Task is the organisation of workshops and special sessions at conferences, like this one. Previous workshops on e.g. forecasting on the minute scale including lidars, a workshop on the value of forecasts, or special sessions on the Wind Energy Science Conference, the Wind Integration Workshop, the ESIG Meteorology and Market Design for Grid Services workshops are still visible online from the IEAWindForecasting YouTube channel.

International audience; Uncertainty in the upcoming production of photo-voltaic (PV) plants is a challenge for grid operations and also a source of revenue loss for PV plant operators participating in electricity markets, since they have to pay penalties for the mismatch between contracted and actual productions. Improving PV predictability is an area of intense research. In real-world applications, forecasts are often needed for different time frames (horizon, update frequency, etc.) and are derived by dedicated models for each time frame (i.e. for day ahead and for intra-day trading). This can result in both different forecasted values corresponding to the same horizon and discontinuities among time-frames. In this paper we address this problem by proposing a novel seamless probabilistic forecasting approach able to cover multiple time frames. It is based on the Analog Ensemble (AnEn) model, however it is adapted to consider the most appropriate input for each horizon from a pool of available input data. It is designed to be able to start at any time of day, for any forecast horizon, making it well-suited for applications like continuous trading. It is easy to maintain as it adapts to the latest data and does not need regular retraining. We enhance short-term predictability by considering data from satellite images and in situ measurements. The proposed model has low complexity compared to benchmark models and is trivially parallelizable. It achieves performance comparable to state-of-the-art models developed specifically for the short term (i.e. up to 6 hours) and the day ahead. The evaluation was carried out on a real-world case comprising three PV plants in France, over a period of one year.

Wind power forecasts have been used operatively for over 20 years. Despite this fact, there are still several possibilities to improve the forecasts, both from the weather prediction side and from the usage of the forecasts. The International Energy Agency (IEA) Task on Wind Power Forecasting organises international collaboration, among national weather centres, forecast vendors and forecast users. The Task looks back on the first 3 years, and just started the second three-year period. Collaboration is open to IEA Wind member states, 12 countries are already therein.The Task is divided in three work packages: Firstly, a collaboration on the improvement of the scientific basis for the wind predictions themselves. This includes numerical weather prediction model physics, but also widely distributed information on accessible datasets and a benchmark. Secondly, we try to improve the derived power forecasts and deal with forecast vendor related matters to bring the entire industry forward. Thirdly, we will be engaging end users aiming at dissemination of the best practice in the usage of wind power predictions.The main result of the first phase is the IEA Recommended Practice for Selecting Renewable Power Forecasting Solutions.This document in three parts (forecast solution selection process, designing and executing forecasting benchmarks and trials and evaluation of forecasts and forecast solutions) takes its outset from the recurrent problem of forecast users of how to choose a forecast solution. The first report describes how to tackle the general situation, while the second report specifically describes how to set up a forecasting trial so that the result is what the client intended. Many of the pitfalls we have seen over the years, are avoided. Other results of the first phase include a comprehensive review paper on the use of uncertainty forecasts in the power industry and an information portal related to forecasting.In the second phase of the Task, we will take up additional topics such as the uncertainty propagation through the modelling chain, the use of distributed measurements for forecasting, and some initial standardisation of data flows and formats.

International audience; Short-term forecasts of generation or demand are required as inputs into several power system management functions. Forecast models are in general tuned to provide optimal accuracy and reliable estimations of associated uncertainty. In a second step, these forecasts are used as input in tools that perform various functions such as scheduling, reserves estimation, and trading in electricity markets. These functions often integrate algorithms that in turn optimize a value-related criterion, like cost, reliability, income, etc. The literature has shown that in some processes, like energy trading, this value may be optimized if a specific quantile forecast is selected rather than the forecast considered as most accurate. In this paper we propose a new data-driven approach in which the two steps of forecasting and decision-making are unified into one function that is optimized considering a single criterion, i.e. the value featured in the decision-making process. This approach allows us to bypass the use of specific forecasting models and could be extended to any decision-making process for which the inputs, outputs and objective functions are well defined. An intermediate approach is evaluated where a meta-optimization is applied to tune the forecast model as a function of the value it brings. This intermediate approach can also prove efficient, but does not allow us to bypass the forecasting models. The applications considered to evaluate the concept are photovoltaic and wind power generation trading in a day-ahead market, where we simultaneously optimize a forecasting model and a trading strategy, considering the final revenue on the electricity market as the objective function. For the meta-optimization, an Analog Ensemble (AnEn) model is used to forecast PV production, coupled with a Support Vector Regression (SVR) to forecast market prices. For the data-driven approach, we use an Extreme Learning Machine (ELM). As the objective function is highly complex and non-linear, the optimization is carried out by Particle Swarm Optimization (PSO).

International audience; Reduction of energy consumption in the building sector has been identified as a major instrument to tackle global climate change and improve sustainability. In this paper, we propose a methodology to address a retrofit planning problem at a community level, with a building resolution. The resulting tool helps local decision-makers identify pertinent actions to improve the environmental behavior of their territories. Two building retrofit levers are considered, namely envelope insulation and heating systems replacement. Retrofit planning is treated here as a single-objective optimization problem aimed at reducing the total costs of retrofit actions by minimizing their net present value. A multidimensional multiple-choice knapsack problem formulation is proposed through the adoption of adequate decision variables. It suitably balances the complexity induced by the large number of potential retrofit action combinations and the number of variables in the problem and permits a linear formulation. An optimization of virtual building stocks is performed to highlight the developed model's capacity to tackle large problems (5,000 buildings) in a few minutes. Finally, three analyses finally are led on a real case-study territory, featuring both appropriate retrofit solutions and building stock information. Long-term evaluation of retrofit strategies over the short-term results in an additional 10% reduction of energy consumption and greenhouse gases emissions and encourages thermal insulation. When targeting a 40% reduction in energy demand, retrofit costs ranging from 20 to 800e/m 2 are observed. Finally, the developed method was used to draw a CO 2 abatement cost curve at territory level. A 70% reduction of emissions can be achieved with costs under 50 e/tCO 2 e.

International audience; This paper presents a methodology for estimating the optimal amount of automatic Frequency Restoration Reserve provided by an aggregation of renewable power plants. The increasing penetration of distributed weather-dependent renewable generation presents a challenge to grid operators. Wind and photovoltaic power plants are technically able to provide ancillary services, but their stochastic behavior currently hinders their integration into reserve mechanisms. In the methodology developed a Quantile Regression Forest model is used to forecast the aggregated production and a copula-based approach integrates the dependence between prices and renewable production. We then propose and compare three strategies to derive an optimal quantile of the combined production forecasts that can be used as basis to provide a reliable ancillary service to the System Operator. The methodology is evaluated using historical prices for energy and automatic Frequency Restoration Reserve along with production measurements from several renewable power plants.

International audience; This paper analyses the benefits of dynamic line rating (DLR) in the system with high penetration of wind generation. A probabilistic forecasting model for the line ratings is incorporated into a two-stage stochastic optimization model. The scheduling model, for the first time, considers the uncertainty associated with wind generation, line ratings and line outages to co-optimize the energy production and reserve holding levels in the scheduling stage as well as the re-dispatch actions in the real-time operation stage. Therefore, the benefits of higher utilization of line capacity can be explicitly balanced against the costs of increased holding and utilization of reserve services due to the forecasting error. The computational burden driven by the modelling of multiple sources of uncertainty is tackled by applying an efficient filtering approach. The case studies demonstrate the benefits of DLR in supporting costeffective integration of high penetration of wind generation into the existing network. We also highlight the importance of simultaneously considering the multiple sources of uncertainty in understanding the benefits of DLR. Furthermore, this paper analyses the impact of different operational strategies, the coordination among multiple flexible technologies and installed capacity of wind generation on the benefits of DLR.

International audience; In recent years, the penetration of photovoltaic (PV) generation in the energy mix of several countries has significantly increased thanks to policies favoring development of renewables and also to the significant cost reduction of this specific technology. The PV power production process is characterized by significant variability, as it depends on meteorological conditions, which brings new challenges to power system operators. To address these challenges it is important to be able to observe and anticipate production levels. Accurate forecasting of the power output of PV plants is recognized today as a prerequisite for large-scale PV penetration on the grid. In this paper, we propose a statistical method to address the problem of stationarity of PV production data, and develop a model to forecast PV plant power output in the very short term (0-6 hours). The proposed model uses distributed power plants as sensors and exploits their spatio-temporal dependencies to improve forecasts. The computational requirements of the method are low, making it appropriate for large-scale application and easy to use when on-line updating of the production data is possible. The improvement of the normalized root mean square error (nRMSE) can reach 20% or more in comparison with state-of-the-art forecasting techniques

International audience; Microgrid operations are challenging due to variability in loads and renewable energy generation. Advanced tools capable of taking uncertainty into account are essential to maximize microgrid benefits when operating microgrid owned DERs. This paper proposes a novel optimization model for day-ahead economic dispatch of flexible resources within a microgrid environment, considering uncertainty of PV and loads.This model is conceived to support the microgrid supervisory control layer, providing a security-constrained day-ahead strategy to operate three types of microgrid flexible resources: PV, electric storage and controllable loads. The work presented in this paper introduces a novelty in microgrid operations by presenting a stochastic version of the day ahead scheduling of microgrid DERs to deal with uncertainties associated with PV, load and temperature while considering microgrid network limits and end-user comfort as optimization constraints. An annual analysis quantifies the benefits of to the microgrid-owner of a stochastic formulation over a deterministic one both in terms of ensuring end-user comfort and decreasing operation costs.

International audience; Photovoltaic (PV) power generation is characterized by significant variability. Accurate PV forecasts are a prerequisite to securely and economically operating electricity networks, especially in the case of large-scale penetration. In this paper, we propose a probabilistic spatio-temporal model for the PV power production that exploits production information from neighboring plants. The model provides the complete future probability density function of PV production for very short-term horizons (0-6 hours). The method is based on quantile regression and a L1 penalization technique for automatic selection of the input variables. The proposed modeling chain is simple, making the model fast and scalable to direct on-line application. The performance of the proposed approach is evaluated using a real-world test case, with a high number of geographically distributed PV installations and by comparison with state-of-the-art probabilistic methods.

International audience; This chapter shows some test cases from the EU project SAFEWIND on different ways that the planning phase of wind farms can benefit from early knowledge of wind power predictability. The interest is focused on day-ahead short-term forecasting, as the most relevant range for TSOs and wind energy traders. We shall use the term predictability assessment when we quantify wind predictability during the planning phase of a wind farm. This can be associated to the prediction of the hub-height wind speed (the meteorological wind resource), the gross wind power from a specified turbine power-curve model, or the net wind power output, which includes wind farm losses due to wake effects and other environmental aspects that are site and wind farm layout specific. Within the SAFEWIND project, wind power predictability is evaluated to suggest potential benefits during the planning phase of wind farm deployment. Predictability assessment can be useful, for instance, to anticipate operational costs due to market imbalance for energy traders, costs related to prediction of offshore weather windows for turbine installation or maintenance, or to benefit from portfolio effects when aggregating wind farms that operate in uncorrelated wind climates. While the weight of predictability for a wind farm developer is low in the decision-making process compared to the capacity factor, it remains a useful indicator of the quality of wind for grid integration purposes. A series of case studies are presented in various European regions to present tools for wind power predictability assessment at different scales.

International audience; Distributed energy resources (DER) and new types of consumer equipment create many challenges for distribution system operators (DSOs). Power congestions that can potentially be created during normal or contingency situations will lead to increased investments into grid reinforcement. An alternative solution is to use the flexibility provided by the local resources in the grid. In this paper value of flexibility (VoF) is used as an indicator that can be utilized by the DSO to compare it against costs of the different active elements (AEs) providing flexibility services (FSs). The paper proposes flexibility characterization framework that allows to generalize the process of the cost estimations of any AE by using combinations of cost functions. A case study based on an actual distribution grid is provided to demonstrate the potential application of the framework. Results show that by comparing VoF and total cost of the flexibility the most cost-efficient solution could be found.

International audience; The detailed modeling of distribution grids is expected to be critical to understand the current functionality limits and necessary retrofits to satisfy integration of massive amounts of distributed generation, energy storage devices and the electric consumption demand of the future. Due to the highly dimensional non-convex characteristics of the power flow equations, convex relaxations have been used to ensure an efficient calculation time. However, these relaxations have been proven to be inexact during periods of high RES injection. In this paper additional linear constraints were introduced in the power flow formulation to guaranty an exact relaxation. This convex relaxation is then applied within a multi-temporal algorithm in order to evaluate the benefits of storage grid integration. The case study of a French medium voltage feeder is studied to evaluate the maximum capacity of the grid to host RES sources and the advantages of storage systems in reducing curtailment of RES.

International audience; Predictions of wind power production for horizons up to 48-72 h ahead comprise a highly valuable input to the methods for the daily management or trading of wind generation. Today, users of wind power predictions are not only provided with point predictions, which are estimates of the conditional expectation of the wind generation for each look-ahead time, but also with uncertainty estimates given by probabilistic forecasts. In order to avoid assumptions on the shape of predictive distributions, these probabilistic predictions are produced from non-parametric methods, and then take the form of a single or a set of quantile forecasts. The required and desirable properties of such probabilistic forecasts are defined and a framework for their evaluation is proposed. This framework is applied for evaluating the quality of two statistical methods producing full predictive distributions from point predictions of wind power. These distributions are defined by a number of quantile forecasts with nominal proportions spanning the unit interval. The relevance and interest of the introduced evaluation framework are discussed.

International audience; Wind power forecasts have been used operatively for over 20 years. Despite this fact, there are still several possibilities to improve the forecasts, both from the weather prediction side and from the usage of the forecasts. The new International Energy Agency (IEA) Task on Forecasting for Wind Energy tries to organise international collaboration, among national weather centres with an interest and/or large projects on wind forecast improvements (NOAA, DWD, …), operational forecaster and forecast users. The Task is divided in three work packages: Firstly, a collaboration on the improvement of the scientific basis for the wind predictions themselves. This includes numerical weather prediction model physics, but also widely distributed information on accessible datasets. Secondly, we will be aiming at an international pre-standard (an IEA Recommended Practice) on benchmarking and comparing wind power forecasts, including probabilistic forecasts. This WP will also organise benchmarks, in cooperation with the IEA Task WakeBench. Thirdly, we will be engaging end users aiming at dissemination of the best practice in the usage of wind power predictions.

International audience; When participating in an electricity market, wind power generation may be penalized by increased regulation costs due the stochastic nature of the wind resource. The negative impact associated to the stochastic nature of wind may be reduced by coupling the wind farm with energy storage facilities, thus constituting a virtual power plant. In this paper, focus is put on advanced methods for reducing regulation costs. A novel method is proposed for the intra-day scheduling and operation of such a plant in an electricity market environment. Such method is able to minimize the imbalance penalty risks associated to wind power forecast uncertainty through a rolling-window approach. Results based on a real-world test case are presented and discussed.

International audience; The consortium of the European Coordination Action 'Prediction of Waves, Wakes and Offshore Wind' (POW'WOW) has established a Virtual Laboratory (ViLab) for the development and operational use of wind power prediction systems. The objectives are to stimulate research in this field, to tighten the collaboration between forecasters and forecast users, and to follow and communicate the state of the art in short term prediction of wind generation. The present note clarifies the benefits for the participants to the ViLab. Also, it describes the methodology employed and some practical issues related to data and confidentiality.

Article also available at http://iopscience.iop.org/1742-6596/75/1/012046; International audience; This paper presents an investigation on alternative approaches to the providing of uncertainty estimates associated to point predictions of wind generation. Focus is given to skill forecasts in the form of prediction risk indices, aiming at giving a comprehensive signal on the expected level of forecast uncertainty. Ensemble predictions of wind generation are used as input. A proposal for the definition of prediction risk indices is given. Such skill forecasts are based on the dispersion of ensemble members for a single prediction horizon, or over a set of successive look-ahead times. It is shown on the test case of a Danish offshore wind farm how prediction risk indices may be related to several levels of forecast uncertainty (and energy imbalances). Wind power ensemble predictions are derived from the transformation of ECMWF and NCEP ensembles of meteorological variables to power, as well as by a lagged average approach alternative. The ability of risk indices calculated from the various types of ensembles forecasts to resolve among situations with different levels of uncertainty is discussed.