Wind Farm Generation Research Articles

Using Deep Learning to Predict Complex Systems: A Case Study in Wind Farm Generation

Making every component of an electrical system work in unison is being made more challenging by the increasing number of renewable energies used, the electrical output of which is difficult to determine beforehand. In Spain, the daily electricity market opens with a 12-hour lead time, where the supply and demand expected for the following 24 hours are presented. When estimating the generation, energy sources like nuclear are highly stable, while peaking power plants can be run as necessary. Renewable energies, however, which should eventually replace peakers insofar as possible, are reliant on meteorological conditions. In this paper we propose using different deep-learning techniques and architectures to solve the problem of predicting wind generation in order to participate in the daily market, by making predictions 12 and 36 hours in advance. We develop and compare various estimators based on feedforward, convolutional, and recurrent neural networks. These estimators were trained and validated with data from a wind farm located on the island of Tenerife. We show that the best candidates for each type are more precise than the reference estimator and the polynomial regression currently used at the wind farm. We also conduct a sensitivity analysis to determine which estimator type is most robust to perturbations. An analysis of our findings shows that the most accurate and robust estimators are those based on feedforward neural networks with a SELU activation function and convolutional neural networks.

A Robust Decentralized Wide Area Damping Controller for Wind Generators and FACTS Controllers Considering Load Model Uncertainties

This paper presents a systematic procedure to design a decentralized wide area damping controller for the wind farms and flexible ac transmission system (FACTS) controllers in a smart grid considering uncertainties in the load parameters. The effect of uncertainties in load parameters on damping of inter-area modes of oscillations is investigated through the probability collocation method. A robust $\boldsymbol {H_{\infty }}$ output feedback controller is then designed for the uncertain fuzzy system by satisfying linear matrix inequalities, which are obtained from the original bilinear matrix inequalities derived from the Lyapunov-based stability criterion. The input and output signals for the proposed controller are selected based on a joint observability/controllability index. The input signals are wide area signals obtained from the phasor measurement units and the output signals are fed to wind farm generators and the FACTS controllers. The input signal delay compensation is addressed in this paper. The performance of the controller has been tested on New England 39 bus system developed on real time digital simulator and on a practical large Indian power system under various operating scenarios and faults. Its effectiveness is also compared with a conventional power system stabilizer, local damping controllers, and an existing robust centralized wide area controller designed based on mixed $\boldsymbol {H_{2}} / \boldsymbol {H_{\infty }}$ output feedback synthesis.

Analysis of the development of the wind power industry in China \u2014 from the perspective of the financial support

BackgroundThe wind energy industry is an important part of the renewable energy industry. Helpful financial support plays an essential role in the process of its development. This study analyzes the financial support efficiency not only from the aspect of capital raise, but also from the aspect of the allocation of up-, middle- and down-stream of the Chinese wind power industry chain (which includes the fan component manufacturing enterprises, the fan production enterprises, the wind farm generation and operation enterprises).MethodsBased on a data envelopment analysis (DEA) model, this study selected 30 representative public companies which sampled and extracted their financial data in a panel analysis from 2010 to 2015, in which the financial support efficiency for the Chinese wind power industry was investigated from the aspect of capital raise and allocation.ResultsIn terms of capital raise, the comprehensive efficiencies of these three streams all reached their peak in 2011, and then indicated a slight decline, whereas the fan component manufacturing enterprises had both the largest pure technical efficiency value in 2011 and a relatively high scale efficiency value during the sampling period. In terms of capital allocation, the fan component manufacturing enterprises and the wind farm generation and operation enterprises indicated both the highest comprehensive efficiency in 2011, which is merely the expansion period of the new energy industry. The wind farm generation and operation enterprises showed the lowest pure technical value in 2012, which is also the depression period of the new energy industry. The scale efficiency of the fan production enterprises as well as the wind energy generation and operation enterprises had a relatively high scale efficiency value from 2010 to 2015.ConclusionThe overall efficiency of financial support in the Chinese wind power industry has a close relationship to the macro-economic environment and capital raise while the allocation efficiency of up-, middle-and down-stream show different characteristics. Moreover, the lag of the core technology is the biggest barrier to the financial support efficiency of the wind power industry.

Advanced Reactive Power Reserve Management Scheme to Enhance LVRT Capability

Abstract: To increase the utilization of wind power in the power system, grid integration standards have been proposed for the stable integration of large-scale wind power plants. In particular, fault-ride-through capability, especially Low-Voltage-Ride-Through (LVRT), has been emphasized, as it is related to tripping in wind farms. Therefore, this paper proposes the Wind power plant applicable-Effective Reactive power Reserve (Wa-ERPR), which combines both wind power plants and conventional generators at the Point of Interconnection (POI). The reactive power capability of the doubly-fed induction generator wind farm was considered to compute the total Wa-ERPR at the POI with reactive power capability of existing generators. By using the Wa-ERPR management algorithm, in case of a violation of the LVRT standards, the amount of reactive power compensation is computed using the Wa-ERPR management scheme. The proposed scheme calculates the Wa-ERPR and computes the required reactive power, reflecting the change of the system topology pre- and post-contingency, to satisfy the LVRT criterion when LVRT regulation is not satisfied at the POI. The static synchronous compensator (STATCOM) with the capacity corresponding to calculated amount of reactive power through the Wa-ERPR management scheme is applied to the POI. Therefore, it is confirmed that the wind power plant satisfies the LVRT criteria by securing the appropriate reactive power at the POI, by applying of the proposed algorithm.

A Markovian wind farm generation model and its application to adequacy assessment

Wind profile, wake effect, and wind turbine outage create considerable impact on the energy production of a wind farm. This paper proposes a Markovian wind farm generation model that incorporates these factors. This model considers the wind farm as a generating unit with multiple generation states. The probability, frequency of occurrence, and transition rate of each state can be obtained using the collected wind profile data and wind turbine reliability parameters. The power output of each state is calculated using Jensen wake model and enumerated wind farm layouts. The proposed model is verified by a sequential Monte Carlo simulation approach using a test wind farm and recorded wind profile data from four different sites in North Dakota, USA. A state merging technique is developed to enable the application of the proposed model to adequacy assessment. The Roy Billiton Test System with a test wind farm is used to demonstrate the application of the proposed model and the procedure to adequacy assessment. Moreover, this paper investigates the influence of wake effect, peak load, seasonal wind pattern, wind turbine reliability parameters, and wind turbine type on system adequacy.

Probabilistic Multiobjective Transmission Expansion Planning Incorporating Demand Response Resources and Large-Scale Distant Wind Farms

Transmission expansion planning (TEP) is a critical issue, particularly coping with the new challenges of smart grids. Under the smart grid environment, demand response resources (DRRs) are contemplated as virtual power plants in energy policy decisions. DRRs affect the controllability of power systems, ranging from short-term to long-term scheduling. In this paper, a nonlinear economic model of responsive loads is presented based on price elasticity of demand and customers benefit function. In order to investigate the impacts of DRRs on TEP, a probabilistic multiobjective TEP incorporating demand response programs (DRPs) is introduced (PMO-TEP $^\mathrm{DRPs}$ ). IC, congestion costs, risk costs, and total incentive costs for participating in DRPs are considered different objectives of PMO-TEP $^\mathrm{DRPs}$ . Here, a probabilistic analysis technique, the so-called two-point estimation method, is applied to handle the uncertainty of wind-farm generation in the TEP problem. Furthermore, the maximum achievable potential and realistic achievable potential for participating customers in DRPs are derived. Due to the problem's nonconvex formulation, a nondominated sorting genetic algorithm II is utilized to evaluate PMO-TEP $^\mathrm{DRPs}$ . Several analyses are carried out on the IEEE Reliability Test System and Iran 400-kV transmission grid to confirm the capability of the proposed framework.

Analysis of the Impact of Wind on Electricity Prices Based on Selected European Countries

Abstract This paper presents an analysis of how wind farms affect electricity prices in four European countries, which although they have invested heavily in wind power differ from each other in the number and type of operational power plants. The countries selected were: Poland, the UK, France and Italy. The approach focused on short-term analysis of market and generation data, to determine if there are behaviors typical for all countries. The results showed that, despite differences in the power mix, all of the countries present similarities in price fluctuations – high power demand and cold weather in the winter period usually overlaps with low wind farm generation levels and electricity can become 2 or even 3 times more expensive. This paper clearly demonstrates the need to develop sufficient means for energy storage hand-in-hand with commissioning increased wind farm capacity.

Exploring the mean-variance portfolio optimization approach for planning wind repowering actions in Spain

The repowering of already installed wind farms is considered one of the most promising and cost-effective short-term strategies to scale-up wind capacity. In this study, we apply Markowitz's mean-variance (MV) portfolio optimization theory to explore alternative repowering actions in Spain. The efficient portfolios – a direct outcome of the MV optimization – offer optimal repowering alternatives to current wind farm generation mixes. They deliver the highest possible average power output (yield) for a given level of supply risk. Different repowering scenarios are considered in this paper that range from a full restructuring of the existing wind generation mix to restricting by certain amounts the percentage of down-/uprating of each reference region. Results show that, depending on the configuration of the MV portfolio optimization problem, hourly fluctuations in the aggregate power supply can be reduced as much as 12–31%, while retaining the current level of energy productivity. In addition, for the level of energy supply risk experienced with the existing portfolio of Spanish wind farms; we can derive more efficient mixes that boost-up productivity by 16–55%. This work aims at providing valuable insight for energy policy-making in the direction of optimally repowering future renewable generation.

Implementation of SVC based on grey theory and fuzzy logic to improve LVRT capability of wind distributed generations

Due to the great absorption of reactive power after voltage drops caused by faults in network, the low-voltage ride through (LVRT) capability of squirrel cage induction generators (SCIGs) in wind farms is a great challenge. If a static VAR compensator (SVC) is installed at the point of common coupling (PCC) of a wind farm with a main network, it can improve the wind generation's LVRT capability with reactive power compensation. In the voltage control loop of a conventional SVC, the voltage actual value of the PCC is compared with the reference voltage value. This paper presents a method for implementation of a SVC based on grey theory and fuzzy logic to improve the LVRT capability of SCIG wind turbines. In this method, instead of the voltage actual value of the PCC, the voltage of the PCC is predicted by the GM (1,1) grey model. Predicted voltage is then compared with reference voltage. After obtaining voltage error, a fuzzy controller with a PI controller in the SVC voltage control loop controls the SVC output. The simulation results are compared for a conventional SVC, fuzzy SVC and fuzzy-grey SVC. These results show the superiority of the fuzzy-grey controller for the SVC in improving the LVRT capability of wind farms with SCIGs.

Calculation of current limiting reactance of hybrid SFCL for low voltage ride-through capability enhancement in DFIG wind farms

This paper presents a new approach for the low voltage ride-through capability enhancement of doubly-fed induction generator (DFIG) wind farms using a hybrid superconducting fault current limiter (SFCL) of the first peak current limiting type, which has the advantage of a fast recovery time. A design for the hybrid SFCL, focusing on current limiting reactor (CLR) reactance calculation, is proposed in this paper to determine an appropriate value for the CLR reactance that satisfies the grid code requirements of the DFIG wind turbines. High-temperature superconductors (HTS) such as Bi-2212, YBCO, Bi-2223, Ti-2223, and Hg-1223 are investigated for use in the hybrid SFCL. The recovery time and first peak current reduction effectiveness are the criteria for selecting the HTS. In the proposed method, a during-fault voltage from the grid code requirement is used as an input parameter to compute the reactance of the CLR. Based on the results of the base-case simulations conducted on the DIgSILENT PowerFactory software, the Bi-2223 HTS with the first peak of fault current less than 4.5 kA and recovery time of 2.5 s is selected for adoption in the hybrid SFCL. The calculated results of the proposed method are compared with the simulation results of the hybrid SFCL, in terms of the DFIG wind turbine fault ride-through capability enhancement, to demonstrate that the performance of the proposed method satisfies the grid code requirements.

Lightning protection for wind turbines in Vietnam

AbstractWind energy has become increasingly important in the total electrical energy supply mix in Vietnam over the last few years. Small, kW turbines were installed in isolated areas a decade ago, while wind farms of several MW to few hundred MW are now being connected directly to national grid, with many additional projects in planning or under construction to fulfill an objective of 6% of the total installed capacity by 2030 (approximately 6200 MW of wind energy component). The increase in wind farm generation results in increased damage from lightning. In this paper, the annual frequency of lightning strikes to wind turbines in Vietnam is calculated using electrogeometric model. Reported lightning incidents to three major wind farms in Vietnam are summarized. Possible causes of failure are discussed, and an EMTP simulation for each incident was performed accordingly. The simulations suggest the failure mechanisms as well the potential of improved grounding to reduce lightning induced damage in future ...

Iterative Sequential Optimization Method for Maximizing the Wind Farm Generation Considering the Wake Effect

The wake effect due to aerodynamic interaction among wind turbines makes the wind farm control a challenging problem. The issues on the wind farm control considering the wake effect include the power generation control, reduction of structural loads, and maximization of energy capture. Among these issues, this paper focuses on the maximization of energy capture and proposes an iterative sequential optimization method (ISOM). The ISOM is not limited to a particular wake model and is based on the actual measurement of the generated power. The ISOM aims to reduce the number of iterations by exploiting the properties of the maximization problem, which include the iterative single-variable optimizations and the sequential optimization in the direction from upwind turbines to downwind ones. Further, the ISOM achieves a significant level of performance in terms of the generated power at the early stage of the iteration process. The effectiveness of the ISOM is verified by the simulations using a four-turbine wind farm and the Horns Rev wind farm. The results show that the ISOM can effectively search the optimal solution with the trade-off between the number of iterations and the precision of the solution.

Wind farms participation in electricity markets considering uncertainties

Rising global temperature and environmental pollution as well as the demand for energy consumption have made finding new and affordable clean energy resources a serious challenge for governments. A possible solution could be renewable resources such as solar, wind or geothermal energies. Restructuring and deregulation have provided a competitive environment which makes analysis of these new energy sources necessary. Wind farms have been receiving more attention from governments because of their noticeable generation capability. The stochastic nature of the wind inflicts uncertainty on the output generation of wind farms which then causes some limitations for the participation of these farms in the electricity market. Thus, in this paper the effects of uncertainty in predicting the wind farm's power on locational marginal price in the market have been studied. According to the advantages and disadvantages of wind farm's power uncertainties, a procedure to maximize the social welfare is presented. The studies have been done on an 8-bus network for 24 h in a day-ahead electricity market. To do this, the farm power is predicted using Neural Network and Wavelet Transform and its uncertainties are calculated using the asymmetric Quantile Regression method.

A novel optimal frequency control strategy for an isolated wind–diesel hybrid system with energy storage devices

This article presents the design of a new effective control strategy to enhance frequency stability of an isolated micro-grid-based wind–diesel hybrid system. The suggested control methodology involves load frequency control coordinated with battery energy storage systems. A recently developed meta-heuristic algorithm called multi-verse optimizer was applied to design an intelligent load frequency control scheme in the aim to handle the frequency fluctuation due to load changes and wind farm integration. The multi-verse optimizer algorithm was used to optimize the proportional–integral–derivative controller parameters for the load frequency control loop. The proposed controller was coordinated with two different kinds of storage system, which are redox flow batteries and electric vehicles. To demonstrate the effectiveness of the proposed control strategy, the simulation was performed under step load changes and then was extended with doubly-fed induction generator wind farm integration. Furthermore, to show the potential of multi-verse optimizer algorithm, a comparative study was done with other approaches available in the literature. In addition, robustness analysis was carried out. The obtained simulation results show that the proposed strategy is a very effective means for providing robust load frequency control controller and to avoid hybrid system instability. Furthermore, the system frequency can be improved using an optimal power management of the stored energy in both redox flow batteries and electric vehicles to compensate the load frequency control capability of the diesel groups, which allow to the possibility of integration of a large penetration of wind farms. In summary, the proposed control strategy may be helpful to identify the needed load frequency control capacity in the presence of dispersed generation’s units.

Enhanced fault location algorithm for smart grid containing wind farm using wireless communication facilities

Modern smart grid (SG) should efficiently accommodate the renewable resources such as wind energy under different operation conditions. High level of wind energy penetration into SG results in operation constraints to avoid any tripping out of wind generators from the grid during fault conditions. Therefore, following the fault occurrence, the utility has to restore power as quickly as possible. This requires accurate and efficient fault location algorithms to have appropriate action for rapid service restoration. This study presents two fault location algorithms taking into account the dynamic behaviour of wind farm generation. In the first conventional algorithm, fault location is estimated using the recorded voltage and current signals at one end of the line. While second algorithm is two-end fault location based on the synchronised voltage and current signals from the line ends. Calculations are carried out in time domain within half cycle using movable windows for recorded signals. The obtained results of two-end algorithm are superior and more accurate compared with the conventional one-end fault location algorithm. This is attributed to SG facilities with modern communication links and data synchronisation technology. The proposed algorithms are tested through grid and wind farm simulations to justify their effectiveness.

Stochastic Optimization-Based Economic Dispatch and Interruptible Load Management With Increased Wind Penetration

In this paper, stochastic optimization of economic dispatch (ED) and interruptible load management is investigated using short-term distributional forecast of wind farm generation. Specifically, using the statistical information of wind farm generation extracted from historical data, a Markov chain-based distributional forecast model for wind farm generation is developed in a rigorous optimization framework, in which the diurnal nonstationarity and the seasonality of wind generation are accounted for by constructing multiple finite-state Markov chains for each epoch of 3 h and for each individual month. Based on this distributional forecast model, the joint optimization of ED and interruptible load management is cast as a stochastic optimization problem. Additionally, a robust ED is formulated using an uncertainty set constructed based on the proposed distributional forecast, aiming to minimize the system cost for worst cases. The proposed stochastic ED is compared with four other ED schemes: 1) the robust ED; 2) deterministic ED using the persistence wind generation forecast model; 3) scenario-based stochastic ED; and 4) deterministic ED, in which perfect wind generation forecasts are used. Numerical studies, using the IEEE Reliability Test System-1996 and realistic wind measurement data from an actual wind farm, demonstrate the significant benefits obtained by leveraging the Markov chain-based distributional forecast and the interruptible load management.

Minimum number of permanent-magnet synchronous generators for coordinated low-voltage ride-through of induction generators in hybrid wind farms

Minimum number of permanent-magnet synchronous generators for coordinated low-voltage ride-through of induction generators in hybrid wind farms

Wind Energy Generation and Assessment of Resources in India

The gap between energy demand and its generation is constantly widening. People have started giving more emphasis on renewable sources of energy. This paper presents the estimation of potential for wind energy generation maps based on fixed wind turbine capacity. Although wind energy has developed substantially in recent years, we have only wind speed and wind potential density maps. Our attempt here is to generate wind energy generation potential maps. Major step in achieving this goal is modeling of wind energy conversion system using TRNSYS software. The model consists of three main components namely the weather, the turbines and energy conversion parameters. The weather data are provided from the meteorological database, namely Meteonorm. The simulated output is compared with actual wind generation of wind farms. After comparing our model results with the existing wind energy generation data, we have extended to compute the wind energy generation for all locations in India. For simulation, 4691 locations are identified considering 0.25° × 0.25° interval. The energy generation simulated data are compiled and developed into maps that are useful to all wind energy developers. The data generated and presented in the form of maps are for all the 30 states of India.

A comparative study of the implementation wind farms integration based on maximization of voltage stability and system loadability

This paper presents voltage stability in power systems which are connected to the wind farm generators. Two steady state models of wind generators, squirrel cage and doubly feed induction generators have been used. P-V curve that is used shows the maximum loading factor when the wind farm was installed. A modified IEEE 14 bus system has been selected to show the performance. For improvement of voltage stability, parallel FACTS devices are selected. Results show that STATCOM can improve the maximum loading factor of both wind generators better than SVC.

Robust active disturbance rejection controller design to improve low‐voltage ride‐through capability of doubly fed induction generator wind farms

This study presents the design of a robust active disturbance rejection (ADR) controller in order to improve low-voltage ride-through (LVRT) capability of wind farms connected with doubly fed induction generator (DFIG). The ADR controller is particularly effective in real-time estimation and mitigation of the total effect of various uncertainties against a wide range of parameter variations, model uncertainties and large disturbances. The performance evaluation of the designed controller is performed on an IEEE system under different test cases. The simulation results show that the proposed controller is robust against uncertainties in operating conditions and successfully improves the damping and voltage stability and thus the LVRT capability of DFIGs.