Empirical Regret Bounds for Control in Spatiotemporally Varying Environments: A Case Study in Airborne Wind Energy

2019 ◽  
Author(s):  
Ben Haydon ◽  
Jack Cole ◽  
Laurel Dunn ◽  
Patrick Keyantuo ◽  
Tina Chow ◽  
...  
Keyword(s):  
Wind Energy ◽  
Control Strategies ◽  
Synthetic Data ◽  
Energy System ◽  
Mobile Systems ◽  
Wind Data ◽  
Maximum Probability ◽  
Data Set ◽  
Actual Performance ◽  
Regret Bounds

Abstract This paper focuses on the empirical derivation of regret bounds for mobile systems that can vary their locations within a spatiotemporally varying environment in order to maximize performance. In particular, the paper focuses on an airborne wind energy system, where the replacement of towers with tethers and a lifting body allows the system to adjust its altitude continuously, with the goal of operating at the altitude that maximizes net power production. While prior publications have proposed control strategies for this problem, often with favorable results based on simulations that use real wind data, they lack any theoretical or statistical performance guarantees. In the present work, we make use of a very large synthetic data set, identified through parameters from real wind data, to derive probabilistic bounds on the difference between optimal and actual performance, termed regret. The results are presented for a variety of control strategies, including a maximum probability of improvement, upper confidence bound, greedy, and constant altitude approaches.

Generalized Empirical Regret Bounds for Control of Renewable Energy Systems in Spatiotemporally Varying Environments

2021 ◽  
Author(s):  
Ben Haydon ◽  
Jack Cole ◽  
Laurel Dunn ◽  
Patrick Keyantuo ◽  
Fotini Chow ◽  
...  
Keyword(s):  
Renewable Energy ◽  
Control Strategies ◽  
Energy Resource ◽  
Mobile Systems ◽  
Wind Data ◽  
Maximum Probability ◽  
Data Set ◽  
Actual Performance ◽  
Regret Bounds ◽  
Varying Environments

Abstract This paper focuses on the empirical derivation of regret bounds for mobile systems that can optimize their locations in real time within a spatiotemporally varying renewable energy resource. The case studies in this paper focus specifically on an airborne wind energy system, where the replacement of towers with tethers and a lifting body allows the system to adjust its altitude continuously, with the goal of operating at the altitude that maximizes net power production. While prior publications have proposed control strategies for this problem, often with favorable results based on simulations that use real wind data, they lack any theoretical or statistical performance guarantees. In the present work, we make use of a very large synthetic data set, identified through parameters from real wind data, to derive probabilistic bounds on the difference between optimal and actual performance, termed regret. The results are presented for a variety of control strategies, including maximum probability of improvement, upper confidence bound, greedy, and constant altitude approaches. In addition, we use dimensional analysis to generalize the aforementioned results to other spatiotemporally varying environments, making the results applicable to a wider variety of renewably powered mobile systems. Finally, to deal with more general environmental mean models, we introduce a novel approach to modify calculable regret bounds to accommodate any mean model through what we term an "effective spatial domain."

Comparison of and Discrepancies Between TMY and TMY2S Predictions for Simple Photovoltaic and Wind Energy Simulations

2000 ◽  
Vol 123 (1) ◽  
pp. 6-9 ◽  
Author(s):  
Darrell D. Massie ◽  
Jan F. Kreider
Keyword(s):  
Wind Energy ◽  
Lower Energy ◽  
Energy System ◽  
The United States ◽  
Clearness Index ◽  
Data Set ◽  
Climatic Regions ◽  
Radiation Data ◽  
Wind Energy System ◽  
Energy Simulations

A new typical meteorological year (TMY2s) data set has been derived from the 1961–1990 National Solar Radiation Data Base (NSRDB). This paper compares PV and wind energy system simulation results using new TMY2s data with results using the original TMY data set. PV and aerogenerator simulations are compared in seven different climatic regions of the United States. Results indicate that PV simulations using TMY2s data provide higher energy values for cloudy regions where the clearness index is low and lower energy values for sunnier climates. TMY2s wind simulations produce lower energy predictions in nearly all cases tested.

Study on the Power Control Method of a Stand-Alone Wind Energy System Based on DSEG

2011 ◽  
Vol 383-390 ◽  
pp. 4096-4102
Author(s):  
Jian Fei Sun ◽  
Bo Zhou ◽  
Hong Hao Guo ◽  
Guang Jie Zuo
Keyword(s):  
Wind Energy ◽  
Control Method ◽  
Control Strategies ◽  
Energy System ◽  
Storage Device ◽  
Operating Characteristics ◽  
Energy Supply System ◽  
Wind Energy System ◽  
Electro Magnetic ◽  
Magnetic Generator

The Stand-Alone wind energy supply system equipped with a doubly Salient Electro-Magnetic Generator (DSEG) is presented in this paper. First, the analysis of the operating characteristics of wind turbine and DSEG is introduced. Based on the strategy of power signal feedback (PSF), the excitation current of DSEG is regulated, as a result, Electro-Magnetic torque of the DSEG is controlled and the maximum wind energy tracking is achieved. Battery is employed as an energy storage device and interfaced to the wind energy system through bidirectional DC/DC converter. The control strategies of the bidirectional DC/DC are proposed to balance the power differences between the generator and the power the load, so that voltage of the DC bus could be constant under varying wind speed and load. The simulation results show the effectiveness of the control methods.

scholarly journals Artificial Neural Networks Applied to a Wind Energy System

2020 ◽  
Vol 9 (4) ◽  
pp. 416-420
Keyword(s):  
Neural Networks ◽  
Artificial Neural Networks ◽  
Wind Energy ◽  
Control Strategies ◽  
Direct Torque Control ◽  
Energy System ◽  
Torque Control ◽  
Wind Energy Conversion ◽  
Wind Energy System ◽  
Artificial Neural

In this context, we are taking a close interest in the optimization of wind energy production. It consists in designing simple to implement control strategies of a wind energy conversion system, connected to the network based on the Double Fed Induction Generator (DFIG) driven by the Converter Machine Side (CSM) in order to improve the performance of Direct Torque Control (DTC) and Direct Power Control (DPC). For this purpose, the artificial neural networks (ANNs) is used. Hysteresis comparators and voltage vector switching tables have been replaced by a comparator based on artificial neural networks. The same structure is adopted to build the two neural controllers, for the DTC - ANN and for the DPC - ANN. The simulation results show that the combination of classical and artificial neural network methods permit a double advantage: remarkable performances compared to the DTC-C and DPC-C and a significant reduction of the fluctuations of the output quantities of the DFIG and especially the improvement of the harmonics rate currents generated by the machine.

Modelling long term investments in wind energy – benefits of combining high resolution geo data, energy system modelling and auction design

2020 ◽  
Author(s):  
Frieder Borggrefe ◽  
Simak Sheykhha ◽  
Kai von Krbek ◽  
Yvonne Scholz
Keyword(s):  
Renewable Energy ◽  
High Resolution ◽  
Wind Energy ◽  
Wind Power ◽  
Data Model ◽  
Energy System ◽  
Wind Data ◽  
System Modelling ◽  
Electricity System

<p>This paper addresses the link between geo data models, market design of renewable energy auctions and energy system models. Renewable energy accounts for around 20% of electricity supply in Europe. In countries such as Sweden, Finland and Germany we already reached a share of more than 40%. In these countries renewables became the main energy source. The dash for building renewable energy in Europe will continue with the EU and national climate targets.</p><p>The impact of renewables on the grid and system operation will increase. Key elements to build an efficient energy infrastructure in the long term are a good understanding on (1) how renewables will penetrate the energy system (regional investments) and a good perception on (2) the effects renewables have on the energy system including (3) the additional infrastructure required, enabling a secure electricity system.</p><p>Since 2005 the DLR uses geo data model ENDAT to predict wind power feed-in and investments in the years up to 2050 based on historic weather data. In order to allow for better modelling of the potentials of wind energy high resolution of wind data and efficient clustering methods are applied to allow a more detailed representation of the long term potentials of wind energy.</p><p>In this paper we combine three modelling approaches: The geo data model ENDAT (DLR), a model of the renewable auctions based on a system dynamics model HECTOR (RWTH Aachen) and an energy system model REMix (DLR) – that allows investigating the long term impact of renewables on the electricity system for 2030, 2040 and 2050. The key questions this paper aims to answer are: How will detailed spatial and temporal modelling of renewable energy data as well as auction design influence the predictions for future distribution of wind power plants? What policy recommendations can be drawn from predictions for the years up to 2050 with regard to policy design and investments in wind energy in Germany and Europe?</p><p>The paper divides in two parts. The first part investigates different approaches to model potential for wind power investments and power generation based on historic wind data. While in the past ENDAT used to generate time series for wind on a country by country basis or on NUTS-1 level, improved models allow for more detailed representation of wind data. Key element of this part is to understand the benefits of high resolution of wind data for the results of the overall energy system modelling.</p><p>The second part of the paper describes how the detailed representation of wind potentials and wind speeds will affect future auction results - and therefore influence long term investments in renewable energy. Model results for the German electricity system will be presented. To benchmark different scenarios, each scenario will be evaluated based on the regional distribution of renewable energies and the resulting impact on the energy system (with regard to grid investments, operation costs and aspects of security of supply).</p>

Sign in / Sign up

Export Citation Format

Share Document