• Energy Research

Maintaining a close balance between power generation and demand is essential for sustaining the quality and reliability of a power system. Currently, due to increased renewable energy generation, frequency deviations and power fluctuations of greater concern are being introduced to the grid, particularly in regions that are weakly interconnected with their surrounding areas, such as small islands. This paper addresses the problem of frequency control in isolated power systems with relevant inclusion of wind power generation. With this aim, we have analyzed the contribution of the demand side to the primary frequency control together with an auxiliary frequency control, which is carried out by variable-speed wind turbines through an additional control loop that synthesizes virtual inertia. We have evaluated both the suitability of these two additional control actions counteracting frequency deviation and their potential reserves and compatibility. The results indicate a substantial improvement in both the dynamic performance and grid frequency stability. Simulations also indicate a decrease in the steady-state frequency error, which may relieve the secondary frequency control.

Converter-connected renewable energy sources (RES) in the distribution network alter the power flow in the network. Often, grid reinforcements and additional local reactive power sources are needed to maintain steady state network operation. Converter-connected RES is an accessible local reactive power source for distribution system operators to optimize network operations. However, the non-linearity and non-convexity in the converter equations make their modelling challenging, which further results in the under-utilization of the reactive power capability from these sources. Oftentimes, the converter-connected generator is partially modelled using only apparent power constraints. This paper presents an approach based on Schur's complement and piecewise planar approximation to incorporate the converter equations in a convex power flow optimization routine. The advantages of employing complete converter capabilities for optimal operation of the distribution network are studied via simulations on a real distribution network dataset. A comparison between deploying reactive power according to the minimum grid codes requirement, partial modelling of the converter, and complete converter capabilities is correspondingly presented. Active power losses decrease by 4.1% over 10 month period with high penetration of renewable generation. Furthermore, the voltage profile in the distribution network is improved and the reactive power dependency on the transmission network is reduced.

Abstract The quantification of reactive power reserves can be essential for transmission system operators for the secure operation of large power systems with high share of renewables. Reactive power support from renewable energy sources needs to be included in reactive power reserve estimation. The reactive power capability of variable renewable energy sources like wind power plants depends on active power production. The stochastic nature of wind speed, and thereby active power production, makes reactive power reserve estimation quite complex. This article proposes a novel methodology to estimate reactive power reserve of wind power plants, taking the uncertainties such as the active power fluctuation, the technical availability of wind turbines as well as the on-load tap changer position of wind power plant transformers into account. The methodology incorporates the definition and determination of robustness index for wind power plant as a reactive power source. The proposed methodology is evaluated on a test system, by assessing the impact of the robustness index on long-term voltage stability. The results show that the choice of an appropriate time period for assessment of wind power plant as a reactive power source is critical to have a high degree of confidence in reactive power reserve estimation. Furthermore, the investigation reveals the importance of considering the power fluctuation and availability in the estimation of the reactive power reserve. The presented case study shows that ignoring fluctuation and availability can lead to overestimation of maximum power transfer, resulting in wrong assessment of the power system security. The proposed methodology can be used by the transmission system operators to quantify the reactive power reserve in real-time and allow better integration of wind power plants in voltage stability assessment studies.

AbstractThe aim of this article is to investigate the market penetration and share of different wind turbine concepts during the years 1995–2004, a period that represents the maturational era of the modern wind power industry. A detailed overview is given based on suppliers' market data and concept evaluation for each individual wind turbine type sold by the Top Ten suppliers over the selected decade. The investigation is processing information on approximately 160 wind turbine types from 22 different manufacturers that have featured in the Top Ten list of wind turbine suppliers during 1995–2004. The analysis is based on comprehensive data covering approximately 97% of the cumulative wind power installed worldwide at the end of 2004. The article also provides an overall perspective on contemporary wind turbine concepts, classified with respect to both their speed control ability and power control type. Current and future trends for wind turbine concepts are discussed. Copyright © 2006 John Wiley &Sons, Ltd.

AbstractThe emphasis in this article is on the impact of fault ride‐through requirements on wind turbines structural loads. Nowadays, this aspect is a matter of high priority as wind turbines are required more and more to act as active components in the grid, i.e. to support the grid even during grid faults.This article proposes a computer approach for the quantification of the wind turbines structural loads caused by the fault ride‐through grid requirements. This approach, exemplified for the case of a 2MW active stall wind turbine, relies on the combination of knowledge from complimentary simulation tools, which have expertise in different specialized wind turbines design areas. Two complimentary simulation tools are considered i.e. the detailed power system simulation tool PowerFactory from DIgSILENT and the advanced aeroelastic computer code HAWC2, in order to assess of the dynamic response of wind turbines to grid faults. These two tools are coupled sequently in an offline approach, in order to achieve a thorough insight both into the structural as well as the electrical wind turbine response during grid faults.The impact of grid requirements on wind turbines structural loads is quantified by performing a rainflow and a statistical analysis for fatigue and ultimate structural loads, respectively. Two cases are compared i.e. one where the turbine is immediately disconnected from the grid when a grid fault occurs and one where the turbine is equipped with a fault ride‐through controller and therefore it is able to remain connected to the grid during the grid fault. Copyright copy; 2010 John Wiley & Sons, Ltd.

Nowadays, co-locate renewable power plants and energy storage systems, forming hybrid power plants (HPPs) have raised commercial interests. One popular configuration of HPP is the hybrid wind-battery plant (HWBP). This paper proposes a data-driven energy management system (DDEMS) for enhancing the profits of HWBPs in spot markets and balancing markets. The two-level scheme is adopted, where the first level models day-ahead optimal offering of energy in spot markets and the second level models imbalance energy settlement in balancing markets. Hybrid stochastic optimization and Wasserstein metric-based data-driven robust optimization are applied to model uncertainties associated with market prices and wind power, respectively. In addition, a novel parameter selection algorithm is proposed to determine the radii of Wasserstein ambiguity sets. Then, the two-level model is reformulated as single-level mixed integral linear programming. Simulation results from two different years show that the proposed parameter selection algorithm helps the DDEMS to find the trade-off between robustness and economy. In addition, the results also demonstrate that the proposed methodology is able to enhance the profits of HWBP in comparison with deterministic optimization and pure stochastic optimization.

Emphasis in this article is on variable speed wind turbines (VSWTs) capability to provide short-term overproduction and better understanding of VSWTs' mechanical and electrical limits to deliver such support. VSWTs' short-term overproduction capability is of primary concern for the transmission system operators (TSOs) in the process of restoring critical situations during large frequency excursions in power systems with high wind power penetration.This study is conducted on a simplified generic model for VSWTs with full scale power converter (Type IV), which includes several adjustments and extensions of the Type IV standard wind turbine model proposed by the IEC Committee in IEC 61400-27-1. This modified standard model is able to account for dynamic features relevant for integrating active power ancillary services in wind power plants, such as frequency support capabilities.The performance of VSWTs during short-term overproduction is assessed and discussed by means of simulations for different wind speed levels, overproduction percentages and durations. The results show that the capability of VSWTs providing short-term overproduction to the grid strongly depends on the initial pre-overproduction conditions.

AbstractThis article compares three reduced models with a detailed model of a doubly fed induction generator system for wind turbine applications. The comparisons are based on simulations only. The main idea is to provide reduced generator models which are appropriate to simulate normal wind turbine operation in aeroelastic wind turbine models, e.g. for control system design or structural design of the wind turbine. The electrical behaviour such as grid influence will therefore not be considered. The work presented in this article shows that with an ideal, undisturbed grid the dynamics of the doubly fed induction generator system is very well represented by the dynamics due to the generator inertia and the generator control system, whereas the electromagnetic characteristics of the generator can be represented by the steady state relations. The parameters for the proposed models are derived from parameters typically available from the generator data sheet and from the controller settings. Thus the models are simple to apply in any case where the generator data sheet is available. Copyright © 2005 John Wiley & Sons, Ltd.